Feedback loops are everywhere. Our linear metrics make it nearly impossible to detect them.
- We offer discounts and free shipping in e-mail campaigns. Performance improves by 50%, so we offer more discounts and free shipping, attracting a larger e-mail subscription base, a base that craves discounts and free shipping. Then we realize that our full-price merchandise and marketing isn't working, but the inexpensive stuff sure seems to work, so we offer more of the inexpensive stuff, which the discount/free-shipping crowd crave.
- Catalog customer acquisition is struggling, so we invest more energy getting catalog-productive names from co-ops. These names are often from rural areas, and are unlikely to buy online unless they are mailed a catalog. Then, our matchback algorithms (sometimes executed by the very organization giving us the names) show us that we must mail catalogs to get online orders, forcing us to acquire more names via the co-ops.
- Our websites fail to convert customers at an acceptable rate, so we "Google-ize" them. Then the site converts customers better, so we dive deeper into "Google-ization".
Our Web Analytics systems are not set up to detect feedback loops, nor are the drill-down based Business Intelligence tools that work off of static, non-dynamic data. These days, it is really important to have a smart business leaders directing a SAS or SPSS or SQL or R programmer, converting feedback loop hypotheses into actionable computer code.
We hear a lot about "globalization", about how everything is now connected, how things spread like a virus. Our businesses, our customers, interact in this ecosystem. And within our own business, we have a unique ecosystem, one where customers freely migrate between products, brands, and channels.
We have an opportunity to spend 2009 creating the reporting and analytic tools that allow us to see the feedback loops, allowing us to act before we're consumed by the negative results of many feedback loops.