Adotas

Where media buyers start online

Featured Author

Gary Angel brings over twenty years of experience in decision support, CRM, and software development, to SEMphonic. Gary co-founded SEMphonic and is president and chief technology officer. He's responsible for leading SEMphonic's development of Web analytics and SEM decision making tools for web marketing professionals. In addition, he helps companies like WebMD, Intuit, American Express and Charles Schwab maximize their web channel marketing through intelligent use of Enterprise Web Analytics

More articles by Gary Angel





Features

Scouring for Sales: Why Better Measurement Yields Better Performance in the Search Strata

Written on
Nov 13, 2006 
Author
Gary Angel  |
Share
Scouring for Sales: Why Better Measurement Yields Better Performance in the Search Strata

First, there are so many paths from Search that understanding them via complete paths doesn’t work — there is too much information that can’t be easily rolled up in useful buckets. On the other hand, trimmed paths like top 100 also fail – the tail is too long — so much behavior goes on in the 1000+ pages that looking at any number of top X pages is very misleading.

The second problem with measuring Search routing is that, ideally, you want to understand how well Search is routing for a given class of search terms. If you want to compare how Search routes to your product pages, you need to check performance against the set of “product” related Search Terms — not against career, company, or support related terms. But just as Search routes to thousands of pages, so too is it driven by tens of thousands of search terms — all completely uncategorized.

Categorizing search terms is a bear, but it doesn’t really matter because most web analytic tools won’t let you do it anyway. So you’re stuck trying to understand Search Performance with no really good way to measure the key input or the key output!

There are, however, some ways to think about this “black-box” that can help an analyst better understand search’s real performance. Here is a diagram of search outcomes common to most sites:

One of the tricks the analyst can use is to focus on “bad” routes — of which there are many fewer — to derive a number for good routes. In the diagram above, many of the outcome types are actually fairly easy to measure and all can be measured with minor changes to your measurement infrastructure (your tagging).

We’ve already discussed “No Result” cases. Navigation and Browsing can often be derived from “Active” Viewing or Link reporting. Where they can’t be derived in this fashion, the analyst can often reasonably assume that any main navigation route should be classified as “Browsing Navigation.” Back-Button is trickier — even if the back-page is actually recorded. Home Page routes are a reasonable proxy on some search sites for Back-outs. Site Exits are always directly available. Refined Searches (by type or keyword) and Search pagination are often available and can always be retrievable with minor changes to tagging. Clicking on a result can also be easily captured with a small change to the tag — and, it can also be derived as the difference between Total Routes and the sum of all the other outcomes.

The key, for all of the measurements dependent on tagging is the addition of an OnClick handler to your Search Results page. The OnClick handler is a javascript function that intercepts each click and records a measurement event before executing the click action. By adding OnClick handlers to each Search Result, you can track whenever someone actually clicks through on a result. You can also track things like the position of the result and, if your search engine can spit it out, its category or result type.

If you are customizing your Search Results for key search terms (and this is probably the single most important type of Search tuning), you can also track customizations — both as a class and as individual campaigns — by using this OnClick capture method and adding appropriate coding to your customizations.

Once you can measure each of these route categories, you’re in a much better position to understand and optimize your Search. By measuring the outputs from Search, you’re measuring how well Search is performing its key function — moving visitors into the site. You’re also getting extremely valuable insight into what happens when Search fails and how much behavior happens inside search. Finally, with good tagging you’ll also be in a position to understand how effective your search customizations are — both individually and as a class.

Unless you can spit out site categories in your search results and pass them back in the tag, this approach does not solve the problem of tracking outcomes BY inputs — categories of search terms. With tools that allow for segmentation by one or more search terms, the best fallback solution for the input problem is to build segments based on a small set of “core” search terms around a concept. It isn’t unusual for a few search terms (less than 10) to account for 80-90% of the volume to an area. By creating segments based on visitors using these terms, you’re finally in a position to measure outcome classes by search type!

Measuring Search is a big topic and a lot of ground has been covered. So here’s a synopsis of the key points:

• search has some special attributes that are very worth measuring including failed searches and most common searchs;
• the high cardinality of search terms (the sheer number of different terms used) can make mapping search performance by site area tricky; as a tool, it is most fruitfully thought of and compared to pages whose job is to move visitors into the site;
• classifying outcome types can provide a much clearer picture of how and how well search is performing;
• the addition of an OnClick handler to your search results page can help you capture the critical information necessary to really understand Search performance;
• and the biggest payoffs to search optimization are often in the customization of templates for key results.

So if you’re in the hunt for better site performance and search is a significant part of your site, don’t be satisfied with simple measurements like “search performs x% better than non-search.” Understanding the routing effectiveness of Search by Search Type will almost always open up dramatic new vistas for optimization. And the more important search is to your site, the bigger the payoff good measurement will provide.





Reader Comments.

No comments yet

Leave a Comment

Add a comment

Pages: « previous page 1 2

Tags: and
Article Sponsor

More Features


  • Once Facebook goes public, what's the most important thing it'll need to do in order to live up the expectations of its real value?

    View Results

    Loading ... Loading ...

Latest News

News Archive

Spotlight

Sponsormob Leads the Way Into RTB for MobileADOTAS – For more than half a decade, Berlin-based tech firm Sponsormob has remained relevant in an industry characterized by [...] more...

Adotas Partnership