Intuition is out, data is in

The Omniture presentation this morning reinforced in my mind my general sense that analytics is reaching maturity and that we’re beginning to parse smaller and smaller data sets until… there’s nothing more to parse (see WAMU, that metrics is an art, too). Metrics is now assumed, and now we’re moving to better application of metrics. OK. I get it. It’s important.

But there’s a really interesting question for small FIs like us when it comes to analytics. At times I don’t even think we have enough data to reach critical data mass. How do you parse visits that aren’t measured in 10000 unit increments? What does a control group even look like for us?

So how much time should I spend on public website data (banking system data is something else…)? I just don’t have enough time to spend in serious analysis (even though I’d like to). So it’s a straight prioritization/pay-off issue.

I’m starting small: using analytics to track obvious optimization (page errors, etc.) and conversions. And reporting. I’ll do reporting. When I have time, I’ll do more. But – alas – not until I have time. The improvements are marginal, given the time I’d have to spend – away from other responsibilities – to do this properly.

Aggregated data across multiple CUs using our basic layout and architecture and features would be useful for optimization and navigation. Definitely.

I’m an analytics believer. I am. I’ve even taken the courses. But it has to make sense in light of (a) site visit volume, and (b) other priorities. One day I hope to work on a site where analytics should matter more. Hopefully that site will be ours.

Other interesting things from today:

• Don’t let offline campaigns necessarily dictate online content. They can be different.
• 15% of “spend” should be on optimization
• Try different messages at different days (and times of day). During the week people want to bank; on the weekends people want to shop. This is one big-bank/big-traffic thing that we actually can do. I’m eager to try this out.