Trending Upward

Let’s get back to talking analytics with a real-world example.

Last week a colleague of mine came to me with a problem. He was optimizing his program website for a few keyword phrases (via organic and paid SEO) that pertain to a specific program he was promoting to prospective students. Although those keyword phrases were getting a *ton* of traffic to his site, he was getting a lot of click-throughs, but nothing else. Why?

Before looking at any analytics, our immediate thought was that the keywords he was targeting might be misleading. Our assumption was that users were typing those keywords into a search engine, seeing our website link, clicking through, but not liking what they saw once they got to our site. Now, the question was, is this segment of our users not finding what they’re looking for on our site or is this segment of our users *not* the right segment we want to target? Are these users getting to our website basically by mistake?

How can we try to get to the truth without actually asking those users? Well, let’s roll our sleeves up.

Digging in. What we found was interesting. We were targeting 4 *suspect* keyword phrases and what we found with one, was basically true for all 4.

For a two month period, we had 3,107 visits from the keyword phrase.

• 87% of those visits ended their session immediately.
• 94% of those visits left after viewing only 1 or 2 pages.
• Not one visitor took an action on the website (filled out an application, requested more information).
• 98% of those visitors did not come back to the website.

Hmmm … not very good, huh?

So, what can we infer from these stats? The website might be broken or down or buggy. That wasn’t the case, though. The website could just be terrible and not be user-friendly at all. That’s probably not the case, though, since this specific program is one of our most popular and users are *taking action* on it every hour, every day.

This is where our research got exciting! Our next step was to look to see if any of those visitors searched for something on *our* site. Remember that onsite search can be like a mini usability test. Users type what they are looking for in their own words! Perfect! Well, that was it! Every single internal keyword search where the visit was referred from those *suspect* keywords was for something that was completely unrelated to our website … and *every* one of those searches brought back zero results (obviously).

*Our* context of the keyword was obviously not the *users* context of the keyword.

Trackback URI | Comments RSS