I’m doing some last minute prep before I head off to NY for WebSideStory’s ActiveInsights conference, and as I think about the days to come, I keep going back to how many large scale projects use metrics in only the most passive of ways.
Most developers, when confronted with the question, can explain their data collection methodology in terms of vendor, libraries, Javascript, pixels whatever. Ask the same group how their sites are reacting to traffic patterns and you are likely to get many a blank stare in return.
From the first leaps into stateful sites years ago until now, most sites remain purely reactive. Tools like Amazon’s suggestion engine are still the exception, not the norm. What (aside from educating our clients on the possibilities) is keeping us from developing sites that could, for example, have multiple links going to the same location (A/B test style) and based off of which links were most effective at garnering our attention, alter the content on the site to better match our personality?
Imagine if every link on a site could be keyed with personality data, browsing a site could be much like filling in a Myers-Briggs survey. Color pallete could be shifted, copy could be tuned from emotional pitches to technical specs, you could get 2 larger feature items instead of 4 smaller ones, etc…
