Thursday, November 29, 2007

KFI’s: Key Forecast Indicators

As I said in my presentation at the eMetrics / Marketing Optimization Summit, if you want to get C-Level people to start paying attention to web analytics, you have to get into the business of predicting / forecasting. Let’s face it, KPI’s are about the past, right? You don’t know “Performance” until it has already happened.

But C-folks don’t really care much about what has already happened, because they can’t do anything about it. What they really want to know is what you think will happen. For example, ideas like “sales pipeline” - a forecast. If you start forecasting - and you are right - you will get attention from the C-folks pronto. The web is a great forecasting tool because it’s so frictionless; it tends to provide tangible signals before many other parts of the business.

So: Do you have any KFI’s - Key Forecast Indicators?

I have one for the Lab Store, and it tripped about 2 months ago. It’s the Unwanted Exotic Index (UEI).

As part of the Lab Store, we run a moderated board where people who want to give up exotic pets can post the availability, and people looking for exotic pets can post requests. Typically, the ratio of people giving them up to wanting them is about .25 - for every post looking to give an exotic up, there are 4 posts looking to adopt.

A couple of months ago, this ratio starts popping higher. A couple of weeks ago it hit 1.25 - for every 5 posts looking to give up an exotic there were 4 posts looking to adopt. The last time something like this happened was prior to the mini-recession of 2004, when the Unwanted Exotic Index tagged 1.0 for a short time. After this happened, our sales got soft about 2 - 3 months later.

Why is the UEI predictive? Let’s go through the logic - my logic, anyway!

Keeping certain types of exotic animals can be a strain on a family, both from a time and money perspective. They can be high maintenance. On the margin, as the economy gets tougher and people look to manage household budgets, these pets can get some scrutiny - particularly if kids have lost interest or gone off to college. So more go up for adoption. At the same time, requests to adopt fall, as families who might have considered an exotic pet put the “owning decision” on hold. Taken together, these decisions cause the UEI to spike higher. Both giving up and deciding not to own exotic pets affects Lab Store revenues “expected” in the future. So the UEI ends up being predictive of future demand.

Makes sense to me.

Now, I’m a pretty good student of macroeconomics and pay attention to many economic indicators, especially predictive ones like the ECRI’s US Weekly Leading Index. If you’re an analyst, you should too; economic indicators provide context for any analysis you might have to do, and clients often want to understand the impact of these external issues on their business.

As far as the Lab Store specifically, I don’t usually pay much attention to the macroeconomic cycles. The pet business tends to be insensitive to the economic cycle; people don’t stop caring for pets as the economy wobbles up and down. That’s why it’s such a good business - if you can find a niche. So I don’t get too concerned when I see these predictive macroeconomic indexes forecasting a slowing economy.

However, what we have here with our Unwanted Exotic Index is a confirmation of the broader economic forecasting tools that is specific to our exotic pet business. That makes me sit up and take notice! Looks like our business is setting up for a repeat of the 2004 slowdown - the last time the UEI spiked like this. Why is this important? Because I can do something with this knowledge. I can re-allocate and re-prioritize based on this knowledge. For example, I can move from a “grow bigger” to a “grow smarter” mode.

And please note: this KFI has nothing to do with traffic or sales on the web site; traffic and sales are “rear view”. By the time you see the sales slow down it will be too late to do anything about it. And that’s why the C-folks don’t care much about web analytics reports.

You could track an index like the UEI with a web analytics tool, but you’d have to come up with the idea first. My point is you will probably have to look outside the usual “rear view” metrics to find one with forecasting ability. I caution you not to substitute a “survey” for a predictive model; people’s opinions are a notoriously lagging indicator. You’ll be up to your ears in the slowdown before people start turning bearish.

So: Do you have any KFI’s - Key Forecast Indicators? Tell us about them.

If you don’t have any KFI’s, now is the time to start looking for them. What can you see now that predicts what will happen in the future? Think about the business, think about the data sources, and put together a bunch of different ideas. Track them back a couple of years and post them monthly going forward. You’re bound to find something predictive. Perhaps something about posting, like the UEI. Recommendations / comments as a percent of visitors or something like that.

If you’re stuck, start with a simple “engagement” idea - percent visitors / members / customers who visited / logged in / bought in the past 90 days. If this percentage is falling, so will your business in the next 3 - 6 months. If your business has a lot of seasonality in it, look to year-over-year comps of the same metric.

If you’ve never played this game before, you won’t have proof your KFI’s work until after the business is in the soup, but you’ll be ready with accurate and actionable KFI’s the next time around!

No comments: