Common Website Personalization Mistakes to Avoid

Personalization platforms have gotten a lot easier to set up over the past couple of years, which has also made it easier to set them up badly. Here are the mistakes showing up most often in 2014, why they happen and what fixes them.

Mistake one: personalizing a segment too small to matter

It’s tempting to build a highly specific segment, visitors from a particular city who viewed a particular product category on a Tuesday, because it feels precise. The problem is that a segment that narrow often gets so little traffic that you can’t tell whether a change in its behavior means anything at all. This happens because narrow segments feel more “personal” and therefore more effective, even when there’s no data to back that up. The fix is to start with segments large enough to produce a trustworthy result, new versus returning visitors, or traffic source, and only get more granular once you’ve proven the basic approach works.

Mistake two: rolling out changes without a control group

A lot of teams turn personalization on for everyone in a segment and then look at whether conversion went up afterward. Without holding back a control group seeing the standard experience, there’s no way to separate the effect of the personalization from ordinary week-to-week swings in traffic quality or seasonality. This happens because a control group feels like leaving money on the table, half your best segment is “missing out” on the improvement. In practice it’s the only way to know if there’s an improvement to miss out on at all.

Mistake three: leaning on browsing history in a way that feels invasive

Referencing a specific product someone looked at three weeks ago, especially across sessions or devices, can come across as surveillance rather than helpful, and it’s a fast way to make visitors uneasy about the site. This mistake happens because the technology makes it easy to be this specific, and specific feels sophisticated. The fix is to personalize around broader signals, category interest, general intent, rather than replaying someone’s exact browsing history back at them in a way that feels like it’s watching too closely.

Mistake four: ignoring what personalization scripts do to page speed

Personalization tools typically load extra JavaScript to identify a visitor and decide what to show them, and that extra step can add real, measurable delay before the page renders, especially on lower-end devices or slower connections. This gets missed because the personalization itself gets all the attention during setup, while performance testing happens rarely if at all. Test page load times with the personalization script running before and after launch, and if the slowdown is meaningful, it can quietly cancel out any conversion gain the personalization delivers.

Mistake five: changing too much of the page at once

Once a team gets access to a personalization platform, there’s a pull toward personalizing the headline, the images, the offer and the layout all together for a given segment. When that combined version performs differently than the standard page, there’s no way to know which element drove the change. This happens because it feels inefficient to test one thing at a time when the tool can change everything at once. Change one meaningful element per test instead, and you’ll walk away understanding what worked, which is what makes the next test better than the last.

The pattern underneath all five

Every one of these mistakes comes from treating personalization as a feature to switch on rather than a discipline to run carefully, with the same rigor you’d apply to any other test. The technology has gotten easy enough that skipping that discipline is tempting; the businesses getting real value out of personalization this year are the ones who haven’t.

---

Keep exploring: browse more Marketing Analytics guides, see everything we published in 2014 or check out the Digital Business Marketing Awards.