• q = the search term
  
• resultTypes = the types of items inside Clearspace to search
  
• peopleEnabled = true/false - whether to include searches on profiles
• dateRange = searching for something in a specific time period
  
• communityID = searching in a specific community
  
• rankBy = sort criteria

1. Login to your Google Analytics account and navigate to the "Analytics Settings" for your account.
2. Edit your profile settings for the profile analyzing Clearspace
3. Select "Edit" next to your "Main Website Profile Information"
4. Scroll down to the Site Search section and enter the following information. You can only add 5 categories, and Clearspace has 6 different parameters (above), so you'll have to choose 5 of them (I don't include the CommunityID as most searches are not targeted at a specific community on the sites I've seen)

Want to see the answers to previous questions?  The pre-blog newsletter archives are here.

Q: The research folks in my company are trying to convince me that measuring sessions and Page Views per Session is more effective than using Recency and Sessions, as you advocate in your book, for a retention metric.

A: For a content site, the Page Views / Session measure can be used as a measure of visitor quality and appropriate marketing to the right audience - a customer acquisition idea - not retention.  And it really needs to be broken out by Source - the average has little actionable meaning.  You want to know the Visitor Sources, and then look at this metric by Source.  This is still Frequency though  - what about visitors who don’t come back?

Q: I am having some difficulty in making a decision regarding this. They want to give me a matrix with Page Views per Session on the Y axis and Total Sessions on the X axis as the “customer retention map”.

A: I’d call that a Customer Activity or Customer Engagement map, I don’t see how it has anything to do with Retention.  This is a “Frequency Map” with no Recency, which is more of a visitor quality thing, it doesn’t tell you anything about Retention.

For example, as long as you are growing your visitor base, with this kind of map you would never realize Visitors were defecting until the Visitor totals started to go down.  By then, it would be too late
to take action, the damage is done and you’re into the death spiral.

Amazon used to boast about how many “best customers” they had based on Frequency.  When Wall Street found out 60% of these best customers had not made a purchase in over a year (Recency), the stock was cut in half.  Frequency is a measure of Current Value, it doesn’t mean anything for Retention without Recency.

Recent-Frequent visitors are best customers, now and in the future; Frequent visitors who are not Recent are former best customers.  Further, without Visitor Source, I don’t know how you take action on the reporting above.

Q: They say that if I need to send messages to only the recent users I could by only querying the users with high sessions that were only generated over the recent past, say, 1 week.  This will be enough to target the users who you say are most likely to respond.

A: Well, this is true, they are saying to pull the listby Recency and Frequency.  So if targeting an e-mail to most likely to respond is what you want to do, this would work.  But it doesn’t give you a “Visitor Retention Map”, if that is what you want.

See a detailed example of such a map here.

The method your folks are suggesting doesn’t tell you how to get and keep Best Visitors, or which Best Visitors seem to be defecting, providing a head’s up on a weak Visitor Source.

So if growing the best Visitor base as rapidly as possible is what you want to do, you need to know if you are churning through a ton of best customers each week.  I would be more concerned about mailing to best customers who are not coming back.  You can’t do that without Recency.  Sounds to me like they want to take the easy way out.  It’s easy to pull a list of people who have been to the site, harder to pull a list of people who have not been there for some period, which is what you want if you intend to act on Retention data.

Q: What are we missing here?  Why, if at all, is it more important to gauge Recency and Frequency (sessions) than Sessions and Page Views per Session?  It sounds like Page Views per Session is more
like “Intensity” than anything else, and I can see how it is also valuable.  Can you shed some light here?

A: Yes, Intensity, which implies Quality, as I said earlier - for most sites a desirable thing. It’s just another way to measure Frequency.  So if this metric is more important than Total Page Views or Total Sessions to your business, then use it.  But without Recency, you have no predictive power.  If you are going to do list pulls based on Recency but not track based on Recency, it seems to me you will be missing the data you need to achieve your retention Objective.

How do you know if it’s working?

Under this method, you could start with 10,000 Best visitors and work your way all the way down to 100 that actually still visit, never learning anything or coming up with a plan to reduce defection.  Worse, this mass defection could be covered up by Acquisition efforts without the Recency reporting, meaning you’d just be churning Visitors and not making any progress.

This kind of Frequency reporting, at best, tells you what has happened in the past, e.g. at some point you had 10,000 visitors who were considered “Best”.  It does not tell you how long it has been since any of these Best Visitors visited the site.

For example, say you are running two ad campaigns.  Ad A generates Views/Session of 6.4, Ad B Views/Session of 3.2.  Based on the tracking proposed, Ad A generates the “best customers”.  So you put all your money into Ad A, and kill Ad B.

Later, you want to do a high-response e-mail, so you pull all the 1 week Recent visitors with high Views per Session, send it out, and get high response.  So far so good.  On further analysis, you find most of these people originally visited as a result of Ad B.  Curious, you look at the Recency of Ad A visitors and find 90% never came back after the 2nd or 3rd visit.

So you have just wasted money and time, because you were not tracking Customer Retention, just using a Retention-oriented metric to pull lists.  You’re only hitting people who decide to keep visiting, and as a result, never learn why people stop visiting.  Whatever you are doing to create this scenario goes unchecked, and all of a sudden the business is in freefall.  The Recency metric can predict this is going to happen for you, give you the head’s up that something is not working for the Visitor segment and allow you to take action.

See the difference?  The value of this kind of modeling is to be able to predict the future, not react to the past.  I don’t know what you are trying to accomplish in the end, what is the Objective of all this?  But assuming high Page Views / Session is good, then you have a way to track Frequency, which is one vector on the Customer Retention Map, the Current Value. But what good is knowing “initial quality” if this quality doesn’t stick over time?

That’s the other half of the Customer Retention map, the Potential Value.  And that’s where Recency comes into the picture.

If you intend to increase Retention, try to keep your customers longer, you want to:

1. Understand which Visitor sources create Visitors who don’t come back, and if paying for any of these sources, reallocate the budget to those sources creating Visitors who do come back again and again.

2. Take action with the folks that were high value Visitors who have stopped Visiting.  Getting a high response rate from Current Visitors doesn’t imply anything about Retention; it just tells you that engaged visitors are more likely to respond.

What you want to know is what Visitor dis-engagement looks like, how many high value Visitors are you losing?  What were their Sources?  What kinds of content did they Engage with / Products did they buy?  Were they frequent contributors?  Did they have Service encounters?  Then find out why they are leaving and take action.

For detailed examples of Visitor & Customer Retention Mapping, see the Measuring Engagement Series.  For examples of how to measure Visitor Recency in Google Analytics, see this post by Avinash.

Jim

Have a question on Customer Valuation, Retention, Loyalty, or Defection? Go ahead and send it to me here. If on the topic above, please leave a comment.

Visitor Retention Mapping

The future is now. Sometimes, we don’t look outside our little angle of it, and that means we miss some possibilities. Other times, we realize something’s out there and we have part of the puzzle, but we’ll catch a different view that gives us even more. I’ve compiled a list of 100 possibilities and things happening out there on the Internet that might be of interest to you. You may have even more to add. Please do.

57 Internet Possibilities to Investigate

Streams of Information

1. Think about FriendFeed as your own personal newsroom or communications center.
2. Revisit Reddit as a news source, and also as a roll-your-own tool for information clustering.
3. Use Alltop as an idea-starter, or as competitive analysis of content spaces.
4. Cook up six powerful searches on Twitter Search and add those RSS outputs to your feed reader.
5. Visit Slideshare weekly and search for new presentations to learn from.
6. Use YouTube as a source for lectures and learning. Use Magnify as a curation tool.
7. Rediscover what’s interesting in Delicious/Popular. Go back and find tags that matter more to you.
8. Reconsider your current blog reading list. Do some quick math. If it is more than 60% related to your industry, you’re doing yourself a disservice. Find the outliers.
9. Pick at least one news source from outside your country to monitor information from a different perspective. One of my sources is The Guardian.
10. Subscribe to at least one blog with a strong opposing view. Learn from it. Learn where your position is different. Learn how they say what they say successfully. Discover what has impact and what doesn’t.

Web-Based Work

Quickly, what I mean by this is that you might consider looking for web-based equivalents to most of your applications, or web-desktop hybrids. Here’s a list of applications for your consideration, and how they might fit into your work flow. In almost all cases, the apps mentioned work well on both Mac and PC.

11. Notes and Storage - Evernote
12. Project Management - Pelotonics or BaseCamp, maybe Staction. (Any others you recommend?)
13. Sales CRM - Pipeline Deals (evaluating it now). Contact management - Batchblue.
14. Invoicing - Freshbooks
15. Scheduling - Google Calendar
16. Brief Communications - Tweetdeck for Twitter. If you haven’t tried it, do it. (Requires Adobe Air.)
17. Online documents, spreadsheets, etc - Zoho or Google Docs.
18. Online presentation management - SlideRocket (or if that scares you, Zoho Show).
19. Info storage (wiki) - PBWiki.
20. Instant messaging - Skype (video, chat, audio), Trillian for PC, or Adium for Mac.

The Video Thing

21. Hulu is good. So is Fancast. The new stuff on Fancast is cool. The OLD stuff is amazing. (**Update: Embeds are there. D’oh!)
22. If you’re making and uploading videos and not using TubeMogul, why not?
23. Most responsive to their community - Blip.tv.
24. If you’re not rocking the Flip MinoHD to make portable video easy, why not?
25. Devices like Panasonic’s new line of Viera TVs (not out yet) and the Samsung LN52A650 are not just TVs. They do Internet stuff, and Panasonic’s will offer direct Video On Demand service to Amazon Video.
26. If you’re looking for Netflix instead, check out the Roku box, or check out Boxee.
27. Video as marketing tool is hot again. Check out PermissionTV, for example. Look past the entertainment angle a moment. See?
28. Video meets mobile phone? News says Sling has a SlingPlayer Mobile expansion coming for other networks shortly.
29. I believe (prediction alert) that the 2009 US switch to all digital signals will also shepherd in a huge side-effect shift of high speed Internet adoption. Some of this will be for video.
30. Need to prove something about all this to the boss? Your research paper.

Small and Local

31. Think about EveryBlock. Is there a model there for you? Should you consider your local web presence?
32. Have you looked at Google results lately, if you ask about a location? There’s room for reviews and other information. There are easy tools to submit more search results. What does that mean to you?
33. On a different tack, Google can’t own local, because it doesn’t scale well. What can you do in helping local information and resources flourish? Is there business in that for you?
34. Is your organization growing too large? Have you considered the pirate ship strategy?
35. Have you thought about small communications? What about tools like SocialCast?
36. What does location mean on platforms like Facebook? It serves local ads well. What else?
37. Have you built a small, powerful network for your needs?
38. Have you built a group around your interests like the So Cal Action Sports Network did?
39. What could Flickr be doing for you? Is there value in coming up with projects that draw attention to your efforts?
40. Are you adding your local events to Upcoming and Eventful?

Mobile Applications

41. Explore iPhone apps, even if you don’t have an iPhone. There are ideas there.
42. Keep your eye on BrightKite. It’s a key to the annotated world.
43. Mobile audio is a useful communications tool. Have you tried Utterli? (disclosure: I’m an advisor.)
44. Microcontent is really here. Seen Jesus and Vishnu on Christmas Eve? (YouTube video. Might be offensive on religious grounds.) That’s microcontent. Verizon’s V-Cast is the old thinking, but not far off.
45. If you’re not thinking about games, do. See things like Gameloft and think about the value of the mobile casual games market.
46. Have you seen and thought about Modu? It’s hot. The concept is hot. Can you leapfrog anything with that concept?
47. You think computer viruses are nasty? What about phone viruses?
48. If snaptell doesn’t blow your mind, why not?
49. There is more than beer money to be had in developing applications for mobile devices these days. And there is more exposure to be had with that market.
50. When I say mobile, you think “phone.” Stop. Nike+ is a mobile device. So’s your GPS. So, technically, is a Chumby.

Money

51. Banking on the net is here. Have you tried Mint? Also, if you haven’t broken down and done it, get a PayPal account. Tie it to a small checkings account with limited funds if you want.
52. If you’re hoping to help a charity, try a ChipIn widget to help raise funds. It shows progress nicely.
53. Timing is funny. Christopher S. Penn wrote about money as trust as I was compiling this list. Worth checking out.
54. You should be eagerly awaiting Chris Anderson’s Free: The Past and Future of a Radical Price
55. Explore the Freemium model. See if it makes sense.
56. Affiliate marketing will be a $13 Billion dollar market in 2009. Check it out. (I’m keynoting Affiliate Summit West this fall in New York.)
57. Have you seen StockTwits? Real time stock talk.

That’s all I’ve got. Here’s hoping we get tons of interesting add-on links and ideas on the comments. Oh, don’t forget that comments with URLs often get flagged as spam. Give me a little while to fish you out.

And hey, if you like this post, please subscribe for free to get more information daily (if not more).

Site owners try to account for misspellings

Site owners have tried putting the misspellings in the meta keywords tag (doesn’t really work; search engines typically don’t use the keywords tag) or occasionally using the misspelled version in the word in place of the correct spelling (can hurt conversion as it can make the site look less professional).

Search engines say they account for misspellings

Search engines generally tell site owners not to worry about it. They have spell checkers that work pretty well so when searchers misspell queries, the results will still match the right sites. Searchers know this and tend not to worry about spelling things right, knowing they’ll get the “did you mean” prompt that provides the right spelling. And search engines see so many queries that they know better than anyone how people are likely to misspell things.

What really happens in search results

For the most part, the search engine process works well. Someone misspells a query, the search engine asks them if they really meant the correct spelling, the searcher clicks and all is well. But I’ve recently come across two very different behaviors. (Note: neither of these behaviors are new, some recent searches just caused me to think about them a bit more.)

When a search engine is REALLY sure the query is misspelled
I assume this can happen if the search volume for the typed query is really low. If the search engine is pretty sure the searcher meant something else, they may not only show the “did you mean” prompt, but they’ll treat the misspelling the same way they treat synonyms and just show results for what they think is the correctly spelled query.

You can tell this is happening because a different word than the search term is highlighted. Normally, this is a great search experience, but it can be frustrating when the searcher really meant the original. For instance, this morning, I was looking for my doctor’s number. This was the result:

Google autocorrect search results

As you can see here, Google autocorrected the query from [robon] to [robin] and not only showed me the “did you mean” prompt, but is treating “robin” as a synonym for “robon”. You can tell this is happening because “robin” is bolded in the results. Normally, this behavior is helpful. For instance, “washington” is highlighted in the results because Google is matching it to the “wa” in my query, and that can help me as a searcher see results for my state, whether the site used the abbreviation or the spelled out word. But in the case of finding my doctor, he doesn’t show up until the very last result of the local one box.

So in this case, letting Google sort out the misspelling isn’t helpful for either the searcher or the site owner. (And in this case, the query wasn’t misspelled; Google’s misspelling algorithm just thought it was.) If you see this behavior as a searcher, you can force Google to only return exact matches by adding a “+” to your query. For instance, in my search for my doctor, this was the new result:

Use + to force an exact match in Google search results

Google still showed the “did you mean” prompt, but only showed exact match results and for this particular query, the relevance of the results were much better. But how many searchers know to use the + sign? Maybe 3.

When A LOT of searchers mispell a word
Search engines might see so few queries spelled a certain way, that they are sure it’s misspelled and show the “corrected” result automatically, but the opposite can happen as well. A search engine might see so many misspelled queries, that they think searchers are really looking for that misspelling. (Spelling correction is primarily, if not entirely based on query data.)

Take for instance my post on my Sony Vaio. Apparently, I don’t know how to spell it. But apparently, neither does anyone else (including Sony, as you’ll see in a second). So when you search for the incorrectly spelled [sony viao], Google does suggest the correct spelling, but it doesn’t automatically show sites with the correct spelling in the results. How do I know this? Because my post is the first result. (And a sony.com page with the incorrect spelling is in the top five as well.)

Sony Vaio Mispelled Query

[sony viao] is currently the second highest volume query to my site. (Of course, it’s not bringing my target audience, but that’s a different issue.)

Query Volume: Analytics for Sony

So what’s a site owner to do?

In most instances, spelling corrections from the search engines work great and generally, I don’t think site owners need to worry. In the second case of the misspelling have enough query volume to warrant a full display of its own results, one answer may be user-contributed content. If that many people misspell something in a query, you’re likely to get people misspelling it when they add comments (reviews, ratings, discussions) on your site. I wouldn’t suggest misspelling it on purpose to get the mispelled queries. If you optimize for the misspelling (as I accidentally did), you’ll miss the far greater number of correctly spelled queries. And if you optimize for them both, you’re likely to look unprofessional and may lose visitors. Credibility at a glance can be really important online.

The far greater concern is when you don’t show up for correctly spelled queries because the search engine thinks the searcher meant something else, as in my search for my doctor. I don’t necessarily know the answer here, as this is based on query volume and if very few people are searching for you compared to a similarly spelled word, you could run into problems like this. Certainly the first step is making sure that your site is well-optimized to be found in search. As you can see from my + query, the doctor’s site still is nowhere to be found, even though the search results are all sites that reference him. (And why that might be is an entirely different post!)

Update: Thanks to Glenn Murray for pointing out in the comments that I misspelled “misspell”. In my defense, it’s a common mistake. On the other hand, I do have an English degree, so that’s just embarrassing.

This post, the first part of three, is a How-to guide for integrating your voice of customer data with Google Analytics. I have chosen two popular VoC tools: , Clicktools and Kampyle, though it is not my intention to review the merits of the respective tools themselves. However I do use them both, which makes me a fan of them both.

This is Part I - Integrating Clicktools with Google Analytics. Part II - coming soon.

My standard word of caution for all “GA Hacks” posts - This is a tech tip and requires you to have a knowledge of html and JavaScript to implement and use it


Integrating third party data with Google Analytics

Why should you do this? Well, knowing where visitors come from that take the trouble to give you their valuable time and feedback, is a key component of understanding why things may, or may not, be working on your web site. For example, perhaps visitors that come via a banner advertisement have a poor user experience compared to those that click through on a email link or find you via a search engine. In order to know these when viewing your survey results, you need to obtain that information from your Google Analytics data - and specifically at the point when the visitor clicks through to start your survey.

Essentially, there are two approaches to any kind of data integration - either importing or exporting the data of interest. By this I mean importing your third party data into Google Analytics, or alternatively exporting from Google Analytics into your third party system - a third option is to export both data sets into another system.

Google’s approach is to facilitate the latter. That is, make the exporting of your Google Analytics data as easy as possible. An often used phrase at Google is “data democratisation”, which means making the data accessible to everyone (with suitable access restrictions of course). So the general rule for integration, is to always think in terms of how can I get data out of Google Analytics to integrate it elsewhere. This is the approach I use here.

Of course, that may change in the future. The recent announcement of the Analytics API in theory allows for both the import and export of data into Google Analytics. For now though I am considering only the export of Google Analytics data and without any API programming skills!

Clicktools background information

Clicktools positions itself as an enterprise version of the very popular SurveyMonkey product. And although the name isn’t as intuitive, it is a powerful visitor survey system. Founded in 2001, Clicktools is a UK company based in the beautiful coastal town of Poole with offices in San Francisco (another beautiful place, but sorry Poole beats it!). The product already integrates with SugarCRM and Oracle CRM, yet surprisingly not yet with any web analytics vendors. Prices start at $3000 for an annual license, though there is a 30 day free demo.

For me, there are two strengths to the Clicktools approach:

• The power of the standard built-in reporting capabilities that help you understand the results. This includes a function called “Cross Tab”, that can show how responses to one question influenced another. For example, how people who answered “yes” to Q1 responded to Q5 - see screenshot
• The ability to generate custom reports using the Analytics add-on i.e. build your own report dashboard with custom filters

Integrating Google Analytics with Clicktools

By integrating with Google Analytics data, you can understand how survey respondents found you, which search engine, keyword, referral, affiliate etc. - without having to laboriously ask for this in the survey. The result is Google Analytics visitor data incorporated into your Clicktools survey reports.

There are 4 parts to this integration:

1. Setup your survey in Clicktools
2. Add hidden questions within Clicktools to collect the referral information - upto 5, though these can also be combined
3. Add JavaScript to your page(s) where you invite survey participants - this captures the referral information
4. Populate the hidden questions with Google Analytics information by modifying the call to your survey

1. Setup your survey

I have assumed you have opened your Clicktools account and built your survey - a relatively straight forward process of writing your questions and deciding how you wish visitors to respond to them i.e. using a free form text box, radio button, drop down selector etc., and that your survey is ready to be deployed. That is, there is a link on your web site soliciting feedback - no pop-ups please!

Important for the purposes of this example, I have assumed you have a total of 7 survey questions. Always try to keep this to single figures or risk loosing your valuable respondent.

2. Add hidden questions to your survey

At the end of your Clicktools survey setup, add hidden questions to capture the referral data from Google Analytics. These questions are not visible to the respondent when they complete the survey, but can be populated with values passed through the URL link that calls the survey.

In this example I use two hidden questions, corresponding to the referral source (e.g. search engine) and search term (if used). However, you can add all five referral variables from Google Analytics if you wish - for source, medium, campaign, term and content. As my survey example contains 7 visible questions, my hidden questions are numbered q8 and q9 for search engine and source respectively.

3. Add JavaScript to your pages to grab Google Analytics data

Google analytics stores visitor information, anonymously, in first party cookies. That is, cookies owned by the web site in question. Therefore, if you can add JavaScript code to your web pages, you can also access the Google Analytics cookies. Below is a function to do this for you:

function _uGC(l,n,s) {// originally from urchin.js      if (!l || l=="" || !n || n=="" || !s || s=="") return "-";     var i,i2,i3,c="-";     i=l.indexOf(n);     i3=n.indexOf("=")+1;     if (i > -1) {        i2=l.indexOf(s,i); if (i2 < 0) { i2=l.length; }        c=l.substring((i+i3),i2);     }     return c;}var z = _uGC(document.cookie, "__utmz=", ";");ga_source = _uGC(z,"utmcsr=", "|");          //alert(ga_source)ga_medium = _uGC(z,"utmcmd=", "|");          //alert(ga_medium)ga_term = _uGC(z,"utmctr=", "|");            //alert(ga_term)ga_content = _uGC(z,"utmcct=", "|");         //alert(ga_content)ga_campaign = _uGC(z,"utmccn=", "|");        //alert(ga_campaign)// Replace the "iv" value with the one from your Clicktools accountvar survey = "http://www.clicktools.com/survey?iv=123qwerty&q7=4&q8=";survey_url =  survey +ga_source +"&q9="+ ga_term;document.write('feedback survey')// modify "feedback survey" to what ever you wish to be displayed to the visitor

Save this code to your web server root in a file called survey-import-ct.js.

This code grabs the referrer information that Google Analytics has (from the cookie named __utmz) and sets 5 variables - all preceded wtih ga_. These are then available for passing over to Clicktools when the visitor clicks your survey link.

4. Populate the hidden questions with Google Analytics information

Wherever on your page you wish to invite a visitor to participate in your survey, simply call the JavaScript function as follows:

For example:

Please visit the

This generates the “feedback survey” link on your page. When rendered by a browser, the survey link will be as follows:

http://www.clicktools.com/survey?iv=123qwerty&q7=4&q8=ga_source&q9=ga_term

In this example, I am passing over two of the referral variables; ga_source and ga_term as the answers to the two hidden survey questions, q8 and q9. These are the referral source (e.g. google.com, yahoo.co.uk etc.) and the search term used (if any) by the visitor to the site, respectively.

The results look this


This report (q8 responses) shows where respondents came from in order to find the site prior to completing the survey. Each respondent is listed as a separate row, though aggregate reports can be generated using the Clicktools Analytics add-on. As with all Google Analytics reports, the entry “(direct)” corresponds to visitors that type in your website URL directly, though it can also be the result of a bookmarked link, or a non-tagged email campaign.

Important: Note the numbering used is this example. I have assumed a 7 question survey - that is, 7 visible questions to the respondent. Then I have added two hidden questions (q8 and q9) within Clicktools and passed over the answers to these via the survey URL. You will need adjust numbering accordingly to the length of your own survey - both with Clicktools and the JavaScript code survey-import-ct.js

The next post in this series will discuss integrating Google Analytics with Kampyl.

Are you integrating your data sources? I would be interest in your techniques and experiences - please add your comments below.

No related posts.

We spent about 5 years and $100 million dollars trying to prove offline media would drive new customer acquisition and sales.  We tried everything.  Billboards.  TV.  Radio.  Newspapers.  TV Guides - local, national, and cable.  Flyers,  Shoppers, FSI’s.  Spot cable.  All of it, in just about every combination you can think of.

Each time we did these tests, we set up control markets and looked for Incremental sales in the media markets versus those with no media, based on revenue per household.  We found incremental sales in just about every case. 

The problem was this: even though the media created incremental sales, these sales were never enough to pay back the media on a net basis, meaning (roughly) (Gross Margin - Campaign Cost) - Variable Overhead was negative - even when you took into account the LifeTime Value of a new customer.  Even when you looked at the test markets versus control 3 months, 6 months, and 12 months later, for those who might be thinking about “Brand” or “Awareness”.

If you’re thinking perhaps the campaigns were weak or light on exposure, I offer you this: when the campaigns included coupons, the redemptions were absolutely huge.  That’s good, right?

Well, in a word, No.  Think about it.  There was barely any lift in sales at all, yet huge numbers of coupons were redeemed.  Meaning?

This means that virtually all the coupons were redeemed by current customers, and the coupon / response did not change their behavior.  They bought at the same rate as they would have without the coupon.  It means we gave a ton of margin away in addition to the cost of the Campaign, and generated no increase in Sales.  We literally would have been better off (financially) by doing absolutely nothing.

So, we’re talking through all this and thinking, we have to be missing something.  I mean, you just can’t possibly buy a 100 showing in Boards with saturation TRP bombing on radio and 1/2 page ads in the Weds shopper section of the daily newspaper, along with FSI’s on Sunday and get no effect on sales.  That breaks all the rules of Marketing.  You know, Reach and Frequency and all that.  Awareness and so forth.  Branding.

Unless, of course, the rules have changed, for us, for this Interactive thing.  Maybe it’s, you know, different.

OK, so let’s reverse engineer the question.  Let’s find the markets where we have the best revenue per household, and compare them with the markets where we have weak revenue per household, and see if we can find any differences.  You know, demographics and such.  Household income.  Presence of children.  All that stuff.

Nothing.  No correlations.

Then we’re analyzing the results of some spot cable testing, where we were running heavy rotations on the local cable systems carrying the HSN signal, again with test and control systems.  Sometimes, it seemed to make a huge difference.  Other times, nothing.

We boosted investment in the cable testing to try and replicate the positive outcomes.  Nothing.  In fact, the cable systems with above average revenue per household seemed to stay that way, with or without ads.  The poor performers stayed that way, with or without advertising.  You would normally attribute a result like this to demographics.  But we already knew demographics didn’t matter.

OK, what do the high performing systems have in common?  What do the weak performing systems have in common?

Then it hits us.  Channel position.  The high performers tend to carry our signal on a low channel number, the low performers tend to carry us on a high channel number.  But that doesn’t make any sense, what could â€channel number” do to affect sales?

Then we get the 2nd smack to the upside.  The major broadcast networks are almost always carried by the cable system on low channel numbers, typically the Frequency (local TV 8 on channel 8).

Oh, man.  $100 million in advertising spend to find this out?

It can’t be.

So, we go down to the cable affiliate area and find some cable systems that might want to play ball with us on a test.  We’ll pay all the costs of them moving our signal from up in the 40’s down to the single digits, where the major networks are carried.  We’ll pay for customer notification letters, new channel cards, all that, if they’ll just try it.

It worked.  Almost every time.  And it made sense.  Think about it.

The HSN TV program itself (the “user interface” for you web analytics folks), the interactive nature of the real time presentation, the games, the bargains, pulls people in, gets them hooked.  But they have to see the program to get hooked, they have to experience the Interactivity of it to really “get it”.  â€Advertising” can’t do that, can’t provide or describe what the experience is like.  So we can’t Push them to the signal.  We have to Pull them, and to do that we have to go to where the eyeballs already are - on the major networks.

If we move our channel next to where all the eyeballs are, then we dramatically increase the number of times the average person passes over our channel as they search for entertainment using the remote, which increases the likelihood they will see a product they are interested in, which causes them to watch the show and experience the Interactivity of it.

In other words, we get a higher ranking in the “TV Search Engine” by being close to the broadcast networks.

This play was so consistent, we built a predictive model off the results, and could estimate the increase in profits from doing this for each cable system.  This means we can calculate how much we could spend to make it happen.  The ROI was enormous, just huge.

So we took the Mass Media budget and spent it on cable channel realignments, which involved no Advertising at all.  Because it was the right thing to do - there was no more efficient way to optimize that piece of the Marketing budget for the Company.

Sometimes, Marketing isn’t about Advertising.  It’s about understanding Behavior.  When you add the word “Interactive” to the mix, this Behavior issue becomes much, much more urgent to address.

Will you do the right thing?  Perhaps invest in analyzing customer behavior, so you can allocate your media budget more efficiently?  Maybe try some control groups in your e-mail program?

Or will you buy Advertising just because you have an Ad budget?

Have a question on Customer Valuation, Retention, Loyalty, or Defection? Go ahead and send it to me here. If on the topic above, please leave a comment.

SEO for Cable TV

Understanding the needs of small businesses and home users, MvixUSA has come up with a very attractive solution. The MvixBox offers many of the features consumers have been asking for in a network media player.  Here are some of the features that the new system offers:

Media server

The MvixBox provides gigabit network speed and large storage capacity, which makes it an ideal system for storing, serving, and transferring media files from one location to another. You can store media files, such as AVI, MPEG, and DIVX movies or MP3 files. From a different room in your home or a meeting room, you can access the content any time needed.

Web Disk Capability

The MvixBox is a Linux-based system, and it comes with what everyone expects Linux servers to include – Apache, MySQL, SQLite, and PHP. To make the server user friendly, a web-disk capability is also provided.  This means the user can drag and drop files between the PC and NAS server. User can launch applications like Windows Explorer, WebDisk Explorer, blog software, RSS reader, Torrent client, and more.  The setup ensures that anyone who knows how to operate a PC in the Windows environment will know how to handle a MvixBox.

Web Server

The web server is perhaps the most interesting feature of all. Since the MvixBox comes with support for Apache, MYSQL, and PHP, you can host web sites directly from the server. There are many different types of sites that you can setup – blogs, wikis, personal home pages, etc. If you don’t want to host web sites, you can even develop scalable intranet solutions using the database support.  This capability makes it one of the most affordable web server solution that the market has to offer for small businesses and home users.

DDNS Server

In order for a web server to recognize Internet domain names, there needs to be a DNS server (or in this case DDNS so that dynamic IPs can be supported). The domain name is mapped to the server IP address, and the DDNS service enables the server to recognize the domain name. Traditional servers supported mostly static IPs. For this reason, you can’t usually operate a web server from home, as most home Internet connections operate on dynamic IPs. Dynamic IPs change all the time, as they are assigned by the ISP. Fortunately, the MvixBox offers free support for dynamic IPs. There is no extra charge for the use of this service.

NAS, Torrent Sharing, and FTP Management

MvixBox offers a simple file sharing feature that provides file encryption when transferring files using HTTP or FTP protocols. You can use the built-in Explorer to transfer files, without having to install additional software.
The system boasts simple and secure administration.

As this is an NAS server, one would expect it to come with simple backup features. The MvixBox does come with easy-to-use data storage scheduling and synchronization. The software is included with the system.

For all the above features, an enthusiast would expect the price tag to be hefty. However, this is not true. The MvixBox retails for around $249.00.  Each unit includes two built-in hard drive bays that are capable of holding up a 1.5 terabyte (TB) hard drive. The hard drives are not included. Therefore, you can buy any size that best meets your needs.  Even with the most expensive hard drive bundled, the entire system still costs just about $600.
 

Este año he decidido regalarle a mi padre uno de de esos regalos de los que podrĂ­amos considerar clĂĄsicos: una cĂĄmara de fotos. Mi padre es un apasionado del tema y ya hace tiempo que tenĂ­a en la mente comprarle un bichillo de esos
 que si reflex digital de mil millones de pixels con rangos de ISO de 100 a 1500, slowmotion, multidisparo y de paso mĂĄquina de palomitas.

Si, no tengo mucha idea del mundillo este, por lo que me he puesto a buscar información por internet. Me he mirado todas las webs oficiales en busca de especificaciones, e-commerce comparando precios y, como no, he buscado opiniones y reviews de otros usuarios a ver que opinan del tema, no sea que compre una cåmara que sea una auténtica birria.

DespuĂ©s de unas cuantas horas y de un baile de opiniones y nĂșmeros, me he dado cuenta que estoy pero que al principio. No me entero de nada, estoy perdido. ÂżPorquĂ© no podrĂ­a existir un lugar con todo ese curro que me he pegado de forma centralizada?

Entonces aparece la luz. Mejor dicho me cae un perro del cielo.

Cae a mis manos Swotti. Una idea web/negocio/herramienta que ya me habĂ­a pasado por la cabeza alguna vez. Y me sorprende con que exactitud han plasmado todo lo que tenĂ­a en mi mente.

Swotti es ni mĂĄs ni menos un analizador de las opiniones que andan sueltas por la red. Genera resultados y te los muestra de forma muy ordenada. Los propios creadores de la criatura la definen como:

“Con Swotti lo que intentamos es acercar todas las opiniones que otros internautas han escrito, interpretarlas y organizarlas para que sean comparables. De tal forma que se pueda encontrar el mejor producto o servicio que se ajuste a las necesidades. Intentamos y creemos que hemos logrado, en gran medida, solucionar uno de los problemas de Internet.”

Personalmente Swotti me tiene enamorado y creo que han realizado un trabajo excelente. Falta ver exactamente cual es el algoritmo de interpretaciĂłn de resultados, pero la organizaciĂłn de la informaciĂłn y la estructura me parecen geniales. En el caso de mi cĂĄmara, se pueden conseguir rankings generales, por precio, prestaciones, pantalla, diseño, imagen
 ÂĄFelicidades!

Swotti aĂșn se encuentra en fase beta, pero os invito a que le hechĂ©is un vistazo. SegĂșn su anĂĄlisis de opiniones este es el pĂłdium de cĂĄmaras de fotos

ÂżEvaluarĂ  personas? ÂżEres famosete? Prueba con tu nombre. Yo he probado con un par:

P.D: Lo siento por mis amigos merengones per hay que empezar a caldear el ambiente

What we’re finding, more than anything else, is that customer relationship management is not about technology at all. It’s an attitude. It’s a mentality. This isn’t a new concept, but I do think that we take it for granted. We expect the tool to just do everything inherently. That’s not how it works - GIGO is so true and happens all too often.

To have good customer relationship management, an organization doesn’t have to have any technology *tool* - it simply has to have the commitment and attitude to manage relationships better in order to make the customer’s life better. Period. Tools may make that easier to do, but if the mentality isn’t there, the tool does nothing.

The same can be said for web analytics. All this talk about Google Analytics, Omniture, WebTrends - what they can and can’t do - none if it matters if we don’t have a true desire to use the data that is pulled from these tools to improve the usability of the web site.

Before the tool, it starts with goals. No, not website goals 
 before that.

• What are your business goals?
• Based on your business (school) goals, what are your unit goals?
• Based on your business and unit goals, what are your website goals? What do you want your users to be able to do?

Seriously, really think about those questions. The corporate world is good at answering those questions, but higher education? Not as good. Sure, the school has a mission, but do our units? Do our websites?

Remember that all websites, even higher  education websites, are customer service sites. Yes, I know we don’t like to talk about our students as *customers* but they really are. They pay us to provide them with a service. They are customers, whether we like that term or not.

KPIs! KPIs! How I love my KPIs!! What are our key performance indicators? Do we have any baseline data to measure against? Here are some quick examples of website KPIs (remember they will be different depending upon the type of website):

• Landing page bounce rate (remember a landing page is *any* entry page!)
• Form completion rate (how many times did they view the form vs. actually fill it out)
• Percent of visitors who use the course catalog
• Percent of visitors who use the knowledge base


 and on and on and on 
 again 
 all different depending upon type of site. Also remember that analytics doesn’t mean just *online* data. Are you a support desk or a call center with a corresponding website? Track call volume. When a knowledge base is implemented, does the call volume go down? It should, especially for simple questions.

Finally, *after* we answer these questions, then we can talk about analytics implementation.

Now let’s talk implementation. Just because we slap some javascript above the tag of our pages doesn’t mean we have an analytics mentality. Repeat after me - I/we have a true desire to deliver to the customer the best, most usabie website in our power. We need to become user advocates! Now we’re getting it 
 now we have the attitude!

Ok, now that we have the attitude and the mentality of analytics, let’s get started on the technical stuff.

First, take a look at this quick list of implementation tips from our good friend Avinash. That post is almost 2 years old, but I think it still holds true today.

If you’re implementing a proprietary tool like Omniture or WebTrends, make sure to talk to your vendor about your business and web site goals before implementation.

Since most of higher ed uses Google Analytics, here is a quick list of implementation and other tips and tricks:

• Google Analytics 101: Some Basic Setup Tips
• The Ultimate Google Analytics Plugins, Hacks & Tricks Collection
• 8 Awesome Google Analytics Hacks, Tips and Tricks

There are so many more as well.

Now that we have the user/customer/student advocate mentality *and* we have the tool to help us, now we can truly say we have an analytics attitude!

We at FutureNow are big proponents of setting up a business culture of testing, optimization, and continuous improvement.  There are lots of advantages like improved conversion rates, bigger bottom lines, surviving tough economic times, getting past draining internal debates, and so much more.

Another advantage that we haven’t talked about as much is that an optimization culture can get you past the dangers of “best practices.” Many of our clients have implemented the “best practices” of other sites (or even other industries) and then wonder why they aren’t seeing big gains.  The main reason is that they’ve merely copied other sites’ ideas, and not tested to validate that those practices will work for their own unique site and business model.

I saw an interesting deviation from an industry “best practice” the other day, and whether it’s true or not, I imagine that testing allowed the site designers to go their own way instead of following the herd.

See the screenshot from Amazon.com’s “submit order” page.  Take a look at how they’ve handled the common issue of getting the purchaser to agree to terms and conditions before purchasing.  Many sites and site designers struggle with their Legal departments on how to secure agreement to terms while not reducing the quality of user experience (and conversion rate).

If you’ve read Always Be Testing, you’ve seen how we reference Amazon.com’s many evolutions of their site designs over time which come out of a very disciplined testing and optimization culture.

The common “best practice” in this scenario is to either:
a) have a checkbox near the call to action button that says something like “You must check the box to agree to terms before you submit your order.”
b) have copy near the call to action button that says something like  “By clicking the button below, you agree to our terms and conditions.”
In either case, the common design pattern is to hyperlink to the terms and conditions (near the call to action) so concerned visitors can read them in full.

Now note the subtle deviation from “best practice” on Amazon.com’s current page:

• They have no checkbox.  One less click between the visitor and Getting the Cash.
• They have the standard legal language at the top of the page in a much less prominent position.  It’s outside of the ‘box’ that contains the order information.
• The hyperlinks to the terms that the visitor is accepting are in the footer of the page; nowhere near the call to action.  You’d really have to be looking for them to find them.

So how did Amazon.com get the gumption to leave “best practice” design in the dust?  Again, I can’t be sure of this, but we wager that testing results and analytics data played a major part in the design.  I highly doubt that their Legal experts are playing fast and loose with legal and financial risk!

So there you have one of the less-thought-about benefits of a testing and optimization culture.  Only through testing can you gather the data you need to safely and comfortably make the decisions that work for your business, not your industry.  Best practices are pretty dangerous when you haven’t tested to validate and re-validate them.

As we’ve said many times, don’t copy Amazon.com’s design just because they’re an industry leader.  Do your own testing to gather insight, then make the design decisions that work for you.  And in the spirit of shameless plugs, we’re here to help if you need us.

• Site Preview: lets business users set up time-based versions of their website and preview what their site will look like on different future dates, including side by side preview of date-based versions. With this new tool, business users and marketers can create seasonal and time-based online campaigns with ease.
• Customizable Business User Interface: provides the ability to configure the content management user interface by role to support customers’ business processes entirely. Now, users can have a dashboard tailored specifically to their content management process for dramatically improved efficiencies.
• Real Time Publishing: provides a transparent look into the publishing process via a dashboard view. New capabilities give full control over content publishing at a granular level, as well as error prevention and process recovery for maximum impact in this critical step of the WCM process.

Marketers are often measured by the number of leads they generate. Sales people are measured by sales. Marketers don’t want to be held accountable for sales because they aren’t actually selling. Sales people criticize “poorly qualified” web leads. This all leads to a lot of tension.

The Consequences of “Low Quality” Leads

In fact, in a survey conducted by Omniture and InsideSales.com they set up aliases, such as John@xyzcompany.com, and completed the lead or request information form of 700 different companies, several different times. Then kept track of their lead response and nurturing strategies and found:

• Average email response time: 19 hours, 31 minutes
  *Optimum response time should be within the first hour

• Average phone response time: 36 hours, 57 minutes
  *Optimum phone response time should be within the first five minutes

• How many companies even responded?
  *Only 47.3 percent responded via email, and just 7.5 percent responded via phone!

Web-generated leads decrease effectiveness by over 6x in the first hour according to InsideSales.com.

Obviously, there is a huge disaster in the making. Marketers have potential customers who indicated some level of qualification to buy from your company and sales people who practically refuse to respond. In the end everyone loses out.

Five Steps to Solving the “Lead Qualification” Problem

1. Identify which sources of traffic generation are creating improved qualification rates and ideal close rates. You need to have the analytics and a CRM / sales workflow system that helps you close the loop from marketing all the way through the close of the sale.

2. Identify which offer types improved qualification rates and close rates. Understand your personas and what actually matters to them. Spend time testing and refining offers and generating additional content that you can prove matters to your prospects.

3. Improve your method of qualifying and capturing leads. Test your lead forms to find the right balance of questions that keep the quality and lead count up. Use a platform that enables you to capture web activity (pages/content viewed, tool/calculator interactions) and include that information in the customer profile for sales. This usually involves tagging content to identify its value in the sales and buying process. Content tagging is so simple when you use Persuasion Architecture.

4. Improve your method of distributing leads. Often times the delay in getting form submissions responded to is your internal process of routing leads to the appropriate sales person. This should never be a manual process considering you lose a leads effectiveness with in the first few minutes. Think about it, the last time you submitted a form on a site, when did you want the response to your inquiry. Now! So do your prospects. Use a platform that will automatically distribute leads based on the profile of the customer you have collected through their visit(s). Distribution is often based on geographic region, company size, product/service they are interested in, etc. Either you can have the prospect fill this out in a form or most of this information can be collected and gleaned by web activity.

5. Improve your lead response time. When marketing aligns with sales using effective content planning, integrating the customer buying process with the company’s sales process, distributing leads that have not been turned off by your processes (and horrendous forms), providing sales people with details that matter to them about the prospect’s interests and motivations and then distribute those leads effectively, their isn’t a salesperson who wouldn’t want to respond to that kind of “qualified” prospect right away.

Do you need help generating more qualified leads? We are here for you.

We have been doing some research into how to rank in Google mobile web search. Google is the dominant player in mobile search in the US (source: Nielsen Mobile) - our test site hasn't yet been indexed in the other search engines so I'll have to report back later on how it does there:

Many of you will be familiar with the ranking factors report that SEOmoz put together last year. I found it very useful to see others' views even though I contributed to it. I am looking forward to the update (perhaps next year?) when I expect we will see more people expressing views on usage data being incorporated into the algorithm along with the impact of reviews on local search, etc.

I have been interested in mobile search for a long time. Before starting Distilled, I worked in management consultancy for Internet and mobile companies and the mobile stuff has stuck with me as a continuing interest. I now personally carry out an average of a few mobile searches a day and I know that the total search volume has been exploding. Driven by the iPhone and similar devices finally finding critical mass with normal users, we are finally seeing a true channel emerging for many businesses where there used to be only marketing hype. I'm not declaring 2008 or 2009 the 'year of the mobile web', nor am I going to get into a discussion here about mobile web vs. mobile interpretations of the main web because I think there is more value to be had in thinking about the value of the channel.

Whatever you believe about mobile web, it appears to be here to stay and is growing quickly - Nielsen recently released data on growth of the UK mobile internet market (source: Nielsen mobile) - the y-axis is "unique audience in millions":

Note that if you look at the number of mobile devices in the world and the number of searches performed per user who is using the mobile web, as more mobile users turn to the mobile web, we will eventually see mobile searches outstrip PC-based searches...

If you have carried out mobile searches (using, for example, Google's mobile site - if you use a PC to search there, you'll also need to click 'mobile' in the search results) you will know that the results often look nothing like the main search results. This is for the two good reasons that:

• When you are mobile you are looking for a different kind of resource
• Many mobile devices cannot cope with the same breadth of content as desktops and laptops

It really struck home when, pre-iPhone, I searched for [train times] from my old Nokia (using the trusty Opera mobile browser). With my clumsy query I was hoping to find a mobile-friendly version of either the TFL or National Rail sites (neither gets a link because they are the most annoying sites in the world - if you live in London / the UK you will have experienced them). Instead, I found train-times.mobi which isn't actually a real site - it is basically just a holding page (it's worth noting that the mobile [train times] SERPS appear a bit cleaner when I look again now).

This got me thinking - I thought that 'train times' would be one of the more competitive searches from a mobile, so how was this site ranking? I dug in a little more and it wasn't propped up with a bunch of exact anchor-text links or anything else like that. I was actually quite surprised it had gotten itself indexed. So I started to wonder what factors were influencing the mobile search rankings.

There are two main reasons that led me to hypothesise that PageRank might not be a particularly powerful factor in the mobile ranking factors (I later found a similar post from Patrick at blogstorm - and I've added his reason as a third):

1. People browsing on their mobile are less likely to follow links around the web (they browse less in general) and hence the link graph is less effective as a model for searcher behaviour
2. Mobile users don't link to mobile content so the link graph isn't created by people in the same context as the people browsing the link graph (exacerbating the issue with the regular web of the linkerati being only a subset of users - but in the mobile case, people who can link are those at their PCs and therefore an entirely distinct group from browsers - and when they are at their PCs, why would they link to mobile-specific sites?)
3. Patrick's reason: Time and location are so much more important for mobile searches and links tell you nothing about relevance relative to either of these factors

At the moment, the mobile web is relatively young and there is not a huge amount of competition compared with the regular web, so it seems plausible that you could rank for even reasonably competitive phrases without many links if other factors are relatively more powerful.

We started out by researching the sites ranking for a range of competitive mobile keyphrases (including maps, take away, taxi, train times, restaurant, poker) and formed some ideas on how we should build our test site. Following this, we built a small mobile site in a competitive niche following best development practices but without building links to it. It has two or three nofollow links which resulted in it being indexed in Google mobile search but which shouldn't be passing any ranking benefit. Once indexed, it quickly ranked for our target phrases and has gradually improved to the extent that we are thinking about building it out further (hence why I'm not disclosing it here yet).

Based on the success of this, we have put together our current recipe for mobile rankings (I have no doubt that the algorithms will get more sophisticated as the mobile web matures). We haven't done extensive testing to determine which of these factors is the most important, but this methodology works and all the steps are useful for mobile users, so it seems reasonable to include them all.

How to build a site that does well in mobile search

• Small, lightweight and fast-loading site (< 20kb / page)
  
• XHTML Mobile 1.0 Doctype
• UTF-8 character encoding
• JPEG / GIF images
• Content including "mobile"
• On-site keyphrase optimisation as usual (with a focus on short titles, and small amounts of body copy)
• Regular technical SEO principles

It is worth comparing this list with Google's mobile website advice. Note, however, that keyphrases in the domain name are not required (which I thought could be the sole reason for the 'train times' example above) - we got a site ranking for competitive London-oriented phrases with a domain that included no London-based keywords. We also haven't yet added a mobile sitemap.

Seriously? What about links?

You will also have noticed something missing from that list: links. I can't say that there is no link-based element of mobile ranking at the moment (and certainly can't say that there isn't going to be an increased element of it in the future). But right now, partly because of the relative lack of competition and partly because as I explained above, the link graph makes less sense for mobile sites and users, links are not needed for (at least some) mobile rankings.

I expect a new spate of mobile sites now. Knock yourselves out ;)

Mobile tools

In the course of this research and some other work we have been doing, it has become abundantly clear that there is a lack of tools to help mobile marketers at the moment. There are some useful tools for the technical side of things - we find the .mobi emulator useful for testing, etc. - but the marketing side of things is very under-serviced.

One area that I am particularly surprised not to have seen yet is keyword research. Traditional keyword research tools don't shed much light on how people carry out mobile searches. There are a variety of factors that lead people to search differently when mobile, including (but probably not limited to):

• A need for different kinds of products / services when mobile
• A corresponding lack of some tasks being carried out when mobile (ever bought groceries on your phone? Know anyone who has?)
• Constraints of the device - leading, I would imagine, to shorter searches, more mis-spellings, etc., even when searching for the same thing
  
• Different refinement / re-searching behaviour resulting from time constraints and device constraints
• Desire for results local to current location
• Plausible effects in either direction on branded search - on one hand, people don't have access to all their bookmarks, might not want to type .co.uk, etc. - but on the other hand, performing a search requires two page loads to get to the end location and that can be slow on some devices

For all these reasons, and also to satisfy my general marketing curiosity, I would like better information on mobile searching habits. I think it is pretty strange that you can buy mobile search ads but currently have no data on volumes ahead of time. When you think about finding out how people search, you realise that there are two kinds of people who have this data:

1. the search engines
2. ISPs

Now, the search engines haven't always been particularly keen to release keyword data. AOL recently released data on the top 5 mobile searches, which is quite sweet, but not particularly helpful. Mobile search is still in its infancy - and I suspect that the data is sparse for the smaller search engines, and they might be a little embarrassed to release their data (do Microsoft and Yahoo! admit that "google" is still their most popular search term these days?). I have spoken to representatives at all the major engines and none of them have any road map that they would tell me about for releasing proper keyword data.

In the world of the regular internet, there are so many ISPs that sampling is needed, and the data is often poor (or needing a lot of correction / effort to make it usable). In the mobile world there are far fewer ISPs - in the UK, for example, if you had access to O2's data (segmented by iPhone / non-iPhone) you would have over 20% of the market. Any of the mobile operators could do a decent job of extrapolating their data to the marketplace. Of course, there is the additional benefit that they could gather segmented data across search engines, and they have huge amounts of demographic data to go alongside the keywords.

So, if anyone from a mobile operator is reading, let me know who I need to bug to put together a co-branded tool that uses your data and our knowledge / skills!

We're also finding that regular analytics data doesn't cut it - many mobile devices don't execute javascript. Although you can fall back on log file data, there are reasons we all love the ease of use of Google Analytics, etc. If anyone has a good solution to this problem, let me know! Also - if anyone from the Google Webmaster Central team happens to be listening, we are seeing data for search impressions, but no search traffic (which we know isn't true from our log data).

Behaviour away from search

This morning I went to a presentation by Ciaran Norris about social media marketing (given to an interesting audience of people who mainly hadn't seen will it blend) where he reminded me about the great social media planning tool from Forrester that they call Groundswell. As he showed off the data, I was thinking (probably because I was writing this post in his presentation, sorry Ciaran) that it would be great if you could segment according to mobile use as well as geographic area - e.g., 18-24 year old females in the UK on their mobile - how does that change the proportion who are "joiners" vs. "Creators" vs. "Critics," for example.

Further reading and a funny data point

Brand Republic have a story about Yahoo! striking a deal with Virgin Mobile (I'm tempted to go all el reg on that and put exclamation marks all over the place) which claims that this gives Yahoo! an 80% market share in UK mobile search:

The deal takes Yahoo!'s market share of the UK mobile search audience to 80% and follows similar deals with Orange, O2, 3 and T-Mobile.

To which I say, in the UK vernacular, bollocks. It appears to be an extrapolation from the fact that they have "deals" with 4 of the 5 UK mobile operators (= 80% - see what they did there?). This is despite the fact that I have an iPhone with O2 and the default search when I got it was Google (great "deal" there, Yahoo!). I don't have accurate UK mobile search market share data - has anyone else seen any? - but I'd put money on Yahoo! not having an 80% market share.

If you liked my thoughts on mobile search, you can read more of my mobile musings on the Distilled blog:

• The differences between mobile and cell terminology between the UK and US influence how we think about the mobile web
• How to choose a URL for your mobile website
• Why Google could be the next Google

And a bonus link to great websites for tiny browsers.

Technorati Tags

Do you like this post? Yes No

Here’s an example from the world of politics:

When California’s conservatives wanted to define the word “marriage” by law, Proposition 22 was added to the California ballot.  It was officially titled the “Defense of Marriage Act.”  As the date for voting drew near, it became apparent that the proposition was going to lose by a wide margin.  Finally, a wizard said, “The meaning of a word is always bigger than its definition; words carry associations.  The word ‘defense’ is a violent word, conjuring associative memories of ‘national defense’ and ‘defense budget.’ It makes us think of Vietnam and bloodshed.  And what is the ‘marriage act’?  Sex.  Juxtapose the word ‘Defense’ with the ‘Marriage Act’ and you get a very uncomfortable feeling.  The subconscious image is that of a battered wife, defending herself in a marriage, or of a woman defending herself from sexual assault.  No one wants to vote for a thing called the Defense of Marriage Act.”

With just a few weeks to go, the new ads began talking about “Proposition 22, the Protection of Marriage Act.”

It won by a landslide.  “Protect” and “Defend” may mean the same thing in a dictionary, but they’re radically different in the human mind.*

And of course, there’s Newt Gingrich’s infamous memo, “Language: A Key Mechanism of Control,” detailing entire lists of words designed to frame and color issues to favor GOP policies and platforms.

Indeed, when it comes to using strategic word choices to frame and re-frame the way an audience perceives an issue, the academic authority on the matter is George Lakoff, author of such noted books as Metaphors We Live By, Women, Fire, and Dangerous Things, and Don’t Think of an Elephant: Know Your Values and Frame the Debate.  And, as one can guess from the title of that last book, George is not shy about applying his Linguistics theories to political debate.  In fact, he recently authored an entire post advising the Obama Campaign on how to most advantageously frame the issues.

Even more topically, there’s plenty of debate and analysis about whether “bailout” is an appropriate term for the proposed legislation to address our current financial crisis.

So clearly at least some politicians get this at a very deep level.  But what about marketers?  Here’s an example I’ll steal from Holly Buchanan: why do spas continue to talk about their treatments in terms of pampering and indulging?  Are these words really activating the right mental frames to best position a spa’s services?  Wouldn’t most women prefer to think about renewing or rejuvenating or healing rather than the more self-centered or selfish frames of indulging and pampering?

How this fits in with Web copy

When asked if and why political consultants are better marketers than most mainstream marketers, one of the very best campaign strategists in the business, Brett Feinstein, wrote back with the following:

“It’s not so much that we are better marketers
we aren’t. Most of the industry is filled with utterly incompetent marketers. This is a backwater for advertising. We make less than our equals in the commercial ad world and work a lot harder. It’s that we focus on different things


We often see things in real time. Because of how most serious campaigns deliver message (few non-political advertisers buy 1,200+ GRP a week in a given market) we generally do not have to wait a few weeks to see the effect of what we are doing. In the biggest of campaigns, we are running tracking polls nightly. At the Congressional level, we are doing it weekly. We literally see if the ad moves the needle almost instantaneously and can tweak (or change) message much more nimbly than in the commercial advertising world. We also just move faster too. If Coke’s sales were plummeting, it would take them weeks if not months to change an ad campaign’s strategy and content. Just shooting and producing one new TV spot would take weeks and huge expense. I can shoot and produce a top-notch political spot in a day and have it on the air with a new message or look or whatever. And once I put it on the air, I can see if it is working within a day or two
” [Emphasis mine]

In a word, Brett’s answer as to why the best political strategists often create better messaging strategies than their marketing counterparts is not skill or deep theory but:

1. Measurement – they invest time and money to see the effects of their efforts so they can know what is and isn’t working.
2. Testing – they tweak and fine-tune copy and messaging during a campaign
3. Agility – coming up with a great new strategy is worthless if you can’t implements it in time to win the election.

Would it surprise you to learn that Web Optimization requires the same three traits?

• Without the proper analytics and measurement, you can’t really know how your visitors are reacting to your copy/content.
• Without tools to conduct A/B and multivariate testing, you can’t effectively drive continued improvement.  An understanding of language nuance is important, but you’ll still want to take your copy changes to the “court of last resort” with testing.
• And without the ability to rapidly implement important website changes, you’ll incur enormous opportunity costs and fall behind your more aggressive competitors.

So, in my opinion, internet marketers really could learn a lot from (the very best) political consultants – both of the importance of messaging AND the importance of ongoing optimization.

* Quote taken from pg.113 of Magical Worlds of The Wizard of Ads by Roy H. Williams

** Of course, most politicians could learn quite a bit from business on the importance of creating customer loyalty by actually delivering on marketing promises

When comparing two time series, one typically wants to discuss the size of the gap as it changes over time.  This Business Week chart, for example, depicted for readers the expanding gap between intra-day high and low prices of the S&P 500 for 2008.


This chart construct is effective at pointing out large changes but lacks precision in conveying smaller differences, or trends.  It is always a good idea to plot the gap directly, as we will show below.

More importantly, a better choice of scale can help a lot.  By focusing exclusively on variability (extreme values), this chart hides the relevant information of the closing prices of the S&P.  A point spread of a 100 points means more when the index is at 800 than at 1200.  In order to capture this, we can divide the point spread by the opening price of that day so we say the gap is one-eighth or one-twelfth of the opening price. 

The junkart version makes both changes.  The top chart fixes the scale, plotting the point spread as a percentage of daily opening prices.  Relative to the original chart, the variability in the front part of 2008 was muted because the index was at higher levels back then. 

The bottom chart plots the gap sizes (lengths of the high-low lines).  It is without doubt that directly plotting the gaps showcases the key message.  The current level of volatility is more than double what occurred at the beginning of the year.

If one wants to illuminate the trend as opposed to daily fluctuations, a further improvement will be using moving averages.

For those interested, shown below is a scatter plot that compares the original point spread and the derived point spread, which shows that the change is not trivial.

Reference: "The Market: A Daily Roller Coaster", Business Week, Oct 27 2008.

Pues bien, ya no tenemos ninguno dudas de lo que tenemos encima, aunque nadie de nosotros sabe a ciencia cierta como evolucionarĂĄ la situaciĂłn.

Es el momento de descubrir quien estĂĄ en pelotas, parafraseando  a Warren Buffet en una frase que desconocĂ­a y he leĂ­do vĂ­a Sixto Arias (a quien no conocĂ­a personalmente y la semana pasada en el First Tuesday Madrid me causĂł una gratĂ­sima impresiĂłn) y Carlos Blanco : “cuando baja la marea se ve quien estĂĄ en pelotas“. Pues bien, la marea es posible que baje tanto que veamos quien estĂĄ o no incluso con los zapatos puestos.

Yo, que no sĂ© si por mi carĂĄcter conservador o porque me lo veĂ­a venir desde Diciembre, he sido muy negativo, y me preocupĂ© mucho en los Ășltimos meses por nuestra empresa mĂĄs joven, la Red de Blogs de Ocio Networks.

Me preocupó especialmente esta sociedad por varios motivos, principalmente su juventud, la debilidad del mercado publicitario basado en CPM con los blogs y por el alto y deficitario coste operativo que manteníamos con el objetivo de crecer muy de prisa de tamaño por encima de crecer de forma económicamente sostenida.

Por ese motivo desde Diciembre 2007 empecé a concienciar a nuestro staff directivo con Diego, Mónica y Jaume a la cabeza de la que nos podría venir encima. A Diego y Mónica les insistí especialmente por su función de Coordinador de la Red y Redactora Jefe, y durante todos estos meses hemos estado recortando gastos, optimizando rendimientos y estructura, priorizando inversiones y desechando algunas previstas, parando gastos superfluos, creciendo en tråfico e ingresos a un ritmo sostenido y sobre todo equilibrando balances que eran demasiado negativos para el futuro. Sé que ellos dos, que ahora leerån estos pensamientos, lo hacían un poco sorprendidos y a regañadientes y no veían la situación global tan negra como yo les pintaba. Hoy, tengo la convicción personal de que si no hubiéramos tomado todas esas medidas desde hace 9 meses, posiblemente Ocio Networks no podría sobrevivir en el próximo año y no me cabe ninguna duda que ellos lo ven ahora mucho mås claro que hace 9 meses.

En varios puntos de este trayecto la empresa ha necesitado financiación, y por ese punto conservador que tengo por virtud-defecto he ido cubriéndola personalmente con objeto de no depender de financiación bancaria ni dar aun entrada a nuevos socios manteniendo el 100% de la compañía. No sólo parece haber sido un acierto sino que hace la empresa hoy sostenible aunque yo a nivel personal tenga un pasivo por recuperar en los próximos años. Haber dependido de un banco podría haber sido en estos meses que vienen un punto crítico e irrecuperable.

En paralelo me interesĂ© por el tema comercial y quise invertir y apoyar una pequeña estructura interna que nos permitiera no sĂłlo depender de nuestra agencia, Addoor sino poder cerrar acuerdos de acciones especiales, patrocinios y partnership a largo plazo con empresas que tuvieran sinergias con nuestras publicaciones. Esta estrategia ha dado su fruto muy claramente y hemos firmado en Septiembre y Octubre 3 acuerdos anuales y 3 acuerdos semestrales que nos garantizan una fuente de ingresos añadida a nuestros habituales.

La otra pata necesaria era el crecimiento y optimizaciĂłn, con menos estructura, con menos presupuesto mensual, habĂ­a que agudizar el ingenio para seguir creciendo en trĂĄfico y por consiguiente en ingresos. Desde Enero lo hemos conseguido, siendo precisamente Septiembre nuestro mes de mayor crecimiento absoluto y porcentual.

El miedo y la crisis pueden tener un efecto negativo; la aprensiĂłn y el pĂĄnico ya que el dinero es cobarde. Pero creo que tambiĂ©n puede tener un efecto tremendamente positivo que nosotros internamente hemos vivido. Mi miedo personal ha hecho que hayamos trabajado mĂĄs que en otros periodos, tomado precauciones, agudizado el ingenio, hayamos sido mĂĄs agresivos, mĂĄs exigentes y hayamos podido desarrollar una red comercial. El miedo tambiĂ©n puede ser un estĂ­mulo, sin el que posiblemente  hubiĂ©ramos afrontado mucho peor los tiempos que vienen. Podemos incluir el miedo a la crisis en este caso como un factor mĂĄs de la pirĂĄmide de motivaciĂłn personal aplicada a la empresa.

En nĂșmeros de Ocio Networks, todo esto se traduce en que desde Diciembre hemos crecido mĂĄs de un 100% en audiencia de 1,3 a  2,73 millones de usuarios Ășnicos, casi un 100% en pĂĄginas vistas de 4,5 a 8,18 millones/mes,  mĂĄs de un mĂĄs de un 120% en ingresos publicitarios, y en patrocinios e ingresos atĂ­picos hemos quintuplicado los ingresos que tenĂ­amos entonces.

En paralelo hemos bajado nuestros gastos y costes de explotaciĂłn en un 12% desde Enero. El mes pasado fue el primer mes en el que Ocio Networks diĂł amplios beneficios y ya sabemos que con nuestro crecimiento, costo de explotaciĂłn actual, controlando los gastos y con los acuerdos firmados en los prĂłximos meses duplicaremos el beneficio de Septiembre de forma sostenida y consolidada.

Esperamos haber hecho los deberes y cuando baje la marea llevar todos el bañador puesto.

Sin el miedo a la crisis y al escenario que parecĂ­a venirse encima no hubiera sido posible; es el estĂ­mulo de la adversidad.

It was an excellent question, and today I’m happy to give readers a preview of what we believe to be an excellent answer. Today we’re announcing a measure of Visitor Engagement that, thanks to comScore, can be used to compare levels of engagement across multiple properties in a similar category.

Brand Marketing’s New Measure: Audience Engagement

Audience Engagement is a simple modification of Web Analytics Demystified’s Visitor Engagement calculation that focuses on the core site behavioral attributes, measured through the comScore panel. If you remember, the Visitor Engagement calculation is:

S(Ci + Di + Ri + Li + Bi + Fi + Ii)

The components of the Visitor Engagement calculation are:

• Click Depth Index: Captures the contribution of page and event views
• Duration Index: Captures the contribution of time spent on site
• Recency Index: Captures the visitor’s “visit velocity”—the rate at which visitors return to the web site over time
• Brand Index: Captures the apparent awareness of the visitor of the brand, site, or product(s)
• Feedback Index: Captures qualitative information including propensity to solicit additional information or supply direct feedback
• Interaction Index: Captures visitor interaction with content or functionality designed to increase level of Attention the visitor is paying to the brand, site, or product(s)
• Loyalty Index: Captures the level of long-term interaction the visitor has with the brand, site, or product(s)

(More information about the measure of Visitor Engagement, including the details behind the calculation and several example use cases, can be obtained by reading the white paper that Joseph Carrabis and I recently published, Measuring the Immeasurable: Visitor Engagement which is freely available on this web site.)

The Audience Engagement simplifies Visitor Engagement by applying a “zero weighting” to the Brand, Feedback, and Interaction indices. By removing these values from the core calculation we are left with Click-Depth, Duration, Recency, and Loyalty:

S(Ci + Di + Ri + Li)

In English:

“Audience Engagement is a function of the number of clicks a visitor generates at a site, the amount of time they spent at the site, the frequency at which they return to the site, and their loyalty to the site as a member of the category for all of the sessions to that site during the reporting period.”

We’ve selected these four indices for one very simple reason: When scored using category-level thresholds (with the exception being the Loyalty Index, see below) comScore is able to automatically generate Audience Engagement values and engagement distributions across all of the sites they track.

The result is unique view into the relationship visitors have with the thousands of web sites comScore tracks around the globe. Now, for the first time ever, marketers and advertisers are able to gain insights into the level of engagement using a much more robust measure than session duration, page views, or recency alone.

Using Audience Engagement we can say with a high level of certainty that a greater percentage of Internet users find CNN more engaging than MSNBC and Yahoo! News:

More importantly we can also say that CNN has a larger population of “highly engaged” visitors to their site (22.5% of visitors at CNN versus 15% at MSNBC and less than 10% at Yahoo! News.) We believe that assessment of the audience distribution will provide advertisers an entirely new way to evaluate sites, focusing on audience quality over more simplistic measures of quantity.

This same type of analysis applied to popular network sports sites yields similarly interesting insights:

Here we can see that ESPN, while trailing Yahoo! Sports across all traditional measures (page views, sessions, minutes spent, active days) dominates Yahoo! from an Audience Engagement perspective. A closer examination of these two sites shows that ESPN’s dominance is driven largely by the frequency at which their audience members return to the site (Recency Index of 47.2% versus Yahoo! Sports at 27.0%) — an insight that has clear value to advertisers looking to create brand awareness and drive brand impressions across a sports-minded audience.

While comScore and Web Analytics Demystified are still working on how this data will be packaged and presented, another way of visualizing the relationship between two sites or a site and the category average is using a spider chart:

This chart visually tells the same story as the table above — ESPN has a higher level of Audience Engagement (bigger footprint) that is largely driven by Loyalty and Recency.

We believe that brand advertisers, advertising planners, and marketing managers will be able to use this data to make better decisions during the ad planning and media buying process. The whole debate over the definition of engagement manifest largely from advertisers desire to find more engaged audiences juxtaposed against a lack of faith in the simple measures being proposed as proxies for engagement. Thanks to comScore, these simple measures are about to become a thing of the past, giving way to a significantly more robust measure of the level of Attention audiences are paying at advertising powered sites around the world.

Interpreting Individual Data Points

In case you don’t want to spend the time reading the 50 page white paper I wrote recently on the subject with the mathematician and cultural anthropologist Joseph Carrabis, I’ll provide a brief summary of how the data comScore is reporting can be used.

Here is a sample of sites from comScore’s automotive category:

The first line in this table says that 42.8% of the audience to KBB.com is appreciably engaged with the web site. Engagement at KBB.com is largely driven by visitors clicking deeply into the site and spending an appreciable amount of time doing so, with nearly 85% of audience members exceeding the category Click Depth threshold and over 60% exceeding the duration threshold. Finally, using the distribution data, we can also see that 63% of the audience is highly engaged versus less than 3% who are only poorly engaged.

Audience Engagement data provided by comScore can also be used in a comparative context. Looking at the most and least engaging sites in this group, the data suggests that the audience going to KBB.com is over 400% more engaged than the audience going to About.com Autos (42.8% versus 8.5%.)  This is not to say that advertising at About.com Autos is a bad idea — over 90 percent of the site’s audience appears to be moderately engaged and in some instances a moderate level of engagement may be exactly what the campaign is looking for.

A Technical Note about Audience Engagement’s Loyalty Index

In the Audience Engagement calculation, the Loyalty Index is calculated differently than in the Visitor Engagement calculation because of an advantage conferred by the comScore system. Instead of simply counting the number of times a visitor has returned to the site as we’re forced to do using a site-centric data model, comScore allow us to better approximate loyalty as more commonly used: a measure of your likelihood to prefer a single site or brand over all others in the category. This model is essentially a “share of requirements” model used traditionally in the brand advertising industry and is calculated as:

Li(AE) = Visits to Site / Visits to All Sites in the Category

So, for example, if a comScore panelist is going only to eBay in comScore’s “Auctions” category, their Loyalty Index for eBay would be 100%:

Li(AE) = 10 visits to eBay / 10 visits in the “Auctions” Category

Conversely, if another visitor goes to eBay half the time and Bidz.com half the time, their Loyalty Index for eBay would be 50%:

Li(AE) = 5 visits to eBay / 10 visits in the “Auctions” Category

The result is a distribution of Loyalty Index scores for auction sites tracked by comScore in September that looks like this:

As you can see, eBay’s Audience Engagement component indices are higher than those of their competitors, but their Loyalty Index is much higher and tells us that nearly visitors in this category strongly prefer eBay to their competitors.

One of the challenges comScore and Web Analytics Demystified face regarding the Loyalty Index is the refinement of categories. Some categories like “Auctions” are well defined and represent logical competitors in a sector; others, like “News/Information” include diverse sites like Weather.com, Discovery.com, and Court TV Online. Over time we hope to refine these categories in partnership with comScore clients to provide the most accurate view of category loyalty possible. If you’re interested in participating in this work, please contact me directly.

Next Steps for comScore and Web Analytics Demystified

This is the first time we’ve been able to apply the Web Analytics Demystified Engagement construct to a syndicated audience data base.  We’re just announcing this work today, but we can already see possibilities for the measure’s evolution. Potential next-generation enhancements could include:

• Allowing comScore clients to provide a set of branded search terms to support the inclusion of Visitor Engagement’s Brand Index (Bi)
• Allowing comScore clients to provide a set of key site interactions designed to promote visitor Attention, supporting the inclusion of Visitor Engagement’s Interaction Index (Ii)
• Incorporating third-party data sources measuring more qualitative aspects of the audience relationship with the site, supporting the inclusion of Visitor Engagement’s Feedback Index (Fi)
• Allowing comScore clients to define their own competitive set in order to drill down into a more specific engagement profile in support of the advertising sales process
• Providing comScore clients access to the details behind the Audience Engagement calculation for their site and category
• Providing comScore clients custom access to Audience Engagement data, to provide a measure of Visitor Engagement in situations where the web analytic technology deployed does not support direct measurement

These are just a handful of examples of where this data offering can go. We’re presenting this model and starting the conversation because we want to hear from you. Regardless of whether you’re a current comScore or Web Analytics Demystified client, we would love your feedback regarding the calculation, the data, and the type of insights Audience Engagement is likely to provide to your organization.

Want to Know More about Audience Engagement?

Any reader of this blog knows that I have a passion for talking about the new measures of success on the Internet. I’m tremendously excited about this announcement and happy to talk if you’re interested in how you might be able to leverage Audience Engagement data.

Also, don’t hesitate to contact us if you have concerns about how we measure Audience Engagement or, in the extreme case, don’t think engagement can be measured at all. I firmly believe that the measures of Visitor and Audience Engagement I have proposed and the work I’ve done with Mr. Carrabis and now with comScore are only the beginning of the search for more useful measures of success on the Internet. Because these measures attempt to approximate something we agree is difficult to quantify, we believe that these measures will evolve over time; nothing is set in stone.

But we also believe that Visitor and Audience Engagement are better measures than “page views” and “average time spent” and far more useful to the measurement industry as a whole than simply sticking our head’s in the sand and exclaiming “engagement is an excuse” or worse, taking a Luddite’s view and declaring that complex measures are destined to fail.

For the time being, comScore is previewing additional details on the measure of Audience Engagement with their clients selectively.  If you’d like more information about how to be added to comScore’s list, or would like to discuss the measure of Audience Engagement with me, please email me directly and we can arrange a time to chat.

This is part 7 of a 12 part series on keyword research. This series will guide you through four distinct phase of the keyword research process, providing you step by step guidelines to help you gather, sort and organize your keywords into an effective marketing campaign.

How to find search phrases

I'll continue where I left off last week, looking at the different factors involved in finding relevant search phrases. You'll remember that we discussed the first three; Time, research and deletion. The fourth factor is a bit more involved and it has to do with analyzing the phrases in a bit more detail.

The split / combo factor

Depending on your keywords, a search for any particular core term may produce results ranging from 0 to 1000 different keyword search phrases. Obviously if it produces zero then that core term can likely be scrubbed, or maybe set aside for a day if/when it gains in search popularity. If you've broken down your core terms correctly you will usually get a list between 10-300 keyword phrases returned, but again, this varies by industry and term. In the real world, things don't always work out like we hope.

As you research out your core terms you'll notice that there are three types of results produced. The first will be a core term that returns a very small list of search phrases. Almost too small to do anything with. The second is one that produces a healthy list of phrases that you can easily organize for optimization. The third produces an extremely long list of phrases that can be broken up into several other sub-core terms and groups.

Combining core terms

The ultimate goal of this keyword research process is to end up with very tightly focused groups of keywords that will be optimized most effectively into your website. Many of these groups will come together quite nicely, but others won't. And this is when you have to really start thinking about how to combine core-terms and groups of keywords so that they can still be optimized effectively. We're working towards maximizing our keyword resources instead of wasting them.

When you plug your core term into your research tools and you get very limited results, this is a good sign that you might want to look at combining this core term group with another. If you take a few core terms that are searched well enough, but neither produces a great list of search phrases, you can benefit by finding another like-minded core term to pair with.

When to combine: There are two good rules of thumb when it comes to deciding if a core term is a candidate to be combined with another.

The first is if the research simply didn't produce many phrases associated with that core term. If you use enough research tools you'll no doubt be able to produce a decent size list in almost any circumstances, but we have to weigh here the number of viable, traffic producing phrases that will be valuable to you. If there are less than a handful, say 10 phrases, then you could throw this core term keyword group in with another core term group to optimize together.

There are always exceptions, of course, most notably if you are in a very niche industry. If the number of core terms you have is limited and the phrases produced for each term is also limited, then you may not need to worry about combining and it could be to your advantage not to.

The second is if most of the search phrases produced have very low search volume. Optimization for each keyword group requires a certain amount of effort. No sense wasting all that effort on one or two good-volume phrases and a bunch of low-volume phrases when you might be able to combine with with another core term with similar results. By doing this you can double the benefit for the same amount of effort.

How to combine:There are a few things to keep in mind when combining different core terms.

1) Group most similar. Some core terms will be very similar in meaning and searcher intent. These make a great pair. For example "school supplies" and "student supplies" are likely used by two people looking for the same thing: supplies their child needs in the classroom. Look for core terms that would be a great natural fit in terms of what the searcher expects to find.

2) Avoid inconsistent spellings. Despite the rule above, don't combine different spellings of what is essentially the same core term. If you're dealing with words that can be spelled differently and still mean the same thing such as "duffelbag" and "dufflebag" it's generally not a good idea to try and target both spellings on the same page. If you mix the spellings the inconsistency can look unprofessional to your users and dilute the focus of the page being optimized.

3) Verify on-page unity. Once you've combined based on the rules above, the next thing you look for is to make sure the terms all work together on the same page. This is simply a matter of determining if the page content will be able to support both core terms and their phrases seamlessly. You don't want to force keywords together that won't allow you to write compelling content that maintains a singular focus.

Splitting core terms

Just as there are times to combine core terms, there will also be times to split them into separate sub-groups. In your research you'll find that some core terms produce so many results that it needs to be broken down into more manageable chunks. Buy splitting these core terms you'll not only make the optimization process easier, but you'll create a much more tightly focused optimization campaign.

When to Split: A good determination of when a core term needs to be split is if it produces more than 150 search phrases. In a search for the core term "travel bags" we found that over 50 of the 250+ results contained the word "golf". This knowledge can tell us that "golf travel bag" can be a core term in and of itself.

Now if you remember back when we discussed researching core terms I said not to write down terms with qualifiers. Here now, you are creating new core terms with the qualifiers that were ignored. I had you ignore this in Phase I to keep the process as simple as possible. It's not until you get to this point that you get a sense of which qualifiers matter more than the others.

The number of qualifiers that can/should be split into new sub-core terms depends on the results produced from the original core term. If you see any specific qualifier over 10 times (all with healthy search volume) then splitting that qualifier out is probably a good idea. If you get several hundred results you may be able to split numerous new core terms.

You may even find that a sub-core term split out can be split again. For example, let's say that the new core term we created, "golf travel bags", produced 15 phrases with the word "leather." If that happens then "leather golf travel bags" can be a new sub-sub-core term, giving you an even more tightly focused group of keywords. This isn't as likely, but a possibility you should be aware of.

By splitting and combining core terms at this stage in the process you'll ultimately save yourself a lot of work later on. The more core term groups you have that can produce decent search traffic, the more tightly focused you'll be able to keep each optimized page, resulting in a more focused and rewarding optimization effort.

Tomorrow we'll dive into Phase III of the research process to discuss analyzing and eliminating keywords based on their overall value to the marketing campaign.

Missed one of the steps in this series? Click here to go back to the introduction and follow the links at the bottom.

Check out our small business news site.