Tuesday, October 21, 2008

"How to Choose the Right Marketing Analytics Partner?" Buy a Solution to a Business Problem, Not Bodies.

BtoB has a thing in the October 13 issue about "how to choose the right analytics partner." The summary factors for decision making are:

1. Industry expertise
2. Skills and resources
3. Process
4. Cultural fit
5. Intellectual property
6. Data security

These are all important factors, and they're all on my list... but having been on the vendor side for a while, I think this article misses a critical point. First of all, analytics is kind of a nonsense word. You heard it here first--analytics as a term has lost all its meaning for most business people today... Kind of like CRM did in the 1990s. So, just going out and looking for an "analytics firm" is not going to work. You're looking for a business problem / solution firm. There are a ton of firms that can do analytics--literally any consulting firm with economics or marketing or statistics PhDs--but they'll do everything custom and it'll be expensive.

Case example: Mix optimization literally shares nothing in common with data mining. These are apples and oranges from a marketing perspective, and the only thing they have in common is some kind of statistics is used to solve them. Furthermore, you might say "mix optimization... OK I'll go get an econometrics firm." Wrong again... while econometrics is a tool used for mix optimization, what you really should be buying is a solution. You're buying the increase in marketing performance, not the analysis. Truth be told, econometrics isn't playing much of a role at all in B2B mix optimization... at least not in the studies I've seen. Trite, been there done that, bla bla bla but it's true. You are buying a methodology that's been done 50 times with an 80% common approach. You are not buying analytics. It's why I think a BPO model is going to overtake marketing analytics completely in the next five years. I'm sure many will disagree here... If so please leave comments.

Another example--analytics around business partner channels. Yes, this is what MarketBridge is good at, it's what we do really well, so I'm biased, but it's true. We've optimized partner coverage or done cooperative analytics with partners 100s of times. We know the business objectives--there are 4 or 5 that are repeated across the tech industry. We know the data structure. We do it more efficiently and effectively. Thus, when you buy a "channel analytics" project from MarketBridge, you're buying ten years of experience, a database structure that is 90% complete, and most imporantly, 20% more partner capacity or 30% better partner loyalty or 50% better through-partner marketing performance.

Or yet another example--B2C segmentation. You go to a firm that's done it 100s of times. And you're not buying the best predictive segmentation latent class analysis, you're buying a firm that can deliver a segmentation to meet 2 or 3 or 4 business objectives. So the advise is--don't buy a segmentation project. Buy a business objective--the segmentation will come.

The underlying trend driving this is actually outsourcing. Just having a bunch of PhDs is now not a differentiator. PhDs are cheap and all of them can do regression analysis. These overseas body shops will work for between $50 and $100 an hour. They require a lot of handholding, but they're good for certain grunt tasks. This cost pressure has forced premium firms to own business problems, and the core:
  • Business models
  • Data assets
  • Analytic models
  • "Killer variables"

associated with them. In the long run, this trend is only going to improve B2B marketing in a big way as the space starts to mature. In fact, I think it's only through adopting this BPO approach that B2B marketing analytics reaches scale.

No comments: