Avoiding the search spanking

I read a great article out of Avenue A recently about the spanking clients often give their search marketing firms by asking them for a bunch of useless, meaningless data. Advertising online generates a million metrics - average CPC, conversion rates, effective CPM profits, margins, CTR, impressions, CPA, etc. Having accurate data around these to drive decisions is great, but inaccurate data, or data that doesn't lead anywhere, is worse than having no data at all. I'd rather guess quickly at something than quasi-analyze meaningless data and end up guessing after several days anyway. Fast and wrong strikes me as better than slow and just as wrong. At least I'd have time left over to do something else.

Data for data's sake is bad data. To get good (useful) data, I encourage clients to:

1. Keep the high level objective of your business (or department) in mind. Hopefully that goal is to maximize profits, but sometimes it's more focused - drive usage, acquire new customers, etc.

2. Make the decision tree before you get the data. That is, ask yourself what you'll do in order to better meet your high level objective if data = A and what you'll change if data = B. If you can't decide what you'll do next when you get the data, then getting it is a complete waste of your time.

3. Understand the magnitude of the changes the data might cause. If you get the perfect data and use it perfectly, but you only increase your profits by 2 pennies a year, you probably have better things to do than get that data.

This is all pretty simple. Spend your time doing stuff that matters. Most people do that naturally, but for some reason, when it comes to looking for and analyzing data, the simple rule goes completely out the window. Strange.

Related Tags: sem, search engine marketing