« SEO: The New Establishment in Online Marketing | Main | Link Building in Film »
Site Changes and the Casualty of Causality
April 19, 2006
Suppose you make a slight change to your site - for example, you change the sitewide anchor text leading to your services page. A month later, you've dropped below the fold on page 1 for your "money phrase," when you used to be in the top three.
Trouble is, neither the old anchor text nor the new anchor text used that exact form of the money phrase. But it's the only thing that you've done in the last 30 days to the site, so that has to be it, right?
Unfortunately, about 95% of site owners would immediately deduce that the anchor text change was the culprit, and immediately, they'd change it back. But if that didn't work, what would you do next?
One of the downfalls of a highly metricized industry is that we expect answers. We don't necessarily mind problems, because typically, given enough variables, we can identify the problem and overcome it: A/B testing. Tweaking meta descriptions to increase clickthrough. Changing nav structure to increase index counts. And so on.
But many site owners insist on finding answers where no answers - no easy answers, anyway - exist. They insist on assigning causality to the nearest suspicious variable, and in doing so, end up cutting themselves on Occam's razor. Take the example above, where the "only" thing that has changed is the anchor text. To find other suspects beyond the "anchor text" theory and to find other possible causes of the rankings drop, let's look more closely at what else changed (or may have changed) in the preceding 30 days:
- The sites you link to have linked to new sites.
- The links pointing to your site (and all other sites) are one month older.
- Your site has added no new content.
- The site that used to sit below the fold got tired of it and did some optimization of its own.
- The domain of every site in the SERP has aged a month.
- Premium ads increased from one to two, thus pushing another organic site below the fold and altering the organic clickthough distribution.
That's just a small sample. Sometimes, the SEO who knows the most is the one who truly knows how little s/he knows, and is willing to test multiple scenarios to solve the client's problem.
All posts by Erik Dafforn
posted by Erik Dafforn at April 19, 2006 11:59 PM
Intrapromote: [ Case studies | SEO services | Bios ]
Trackback Pings
To TrackBack this entry, use the following URL:
http://seoblog.intrapromote.com/mt-tb.cgi/205