"Statistical formulas don’t know whether they are being used properly, and they don’t warn you when your results are incorrect." – Deborah J. Rumsey

Attribution is tricky business. While analyzing sales lead generation programs I get to play detective and figure out where business is coming from, where prospects are congregating, and the best way to get in front of them. The lead generation world is full of people selling the ability to use AI or machine learning to predict future leads, but I find my self reminding clients it's never a clean path. At best it's a rhyme. It's a law of big numbers tool versus a specific instance tool. We say, "In general we can expect [X] outcome most of the time," a lot, but come up with exceptions every day.
I think grand unifying theories are fun, but when it comes to finding more new business they aren't helpful for directing ad spend. I sit through a presentation where the sales training company digs into the marketing budget and comes to the conclusion the 60% of budget being spend on direct mail was best re-directed to webinars. Their method includes a lot of great looking calculations like:
- Average Cost Effectiveness for a Lead within a Leadsource = Revenue Contribution Leadsource / Total Cost Leadsource
- Average Lead Turnover for a Lead within a Leadsource
= Business Days / Leadsource Closure Speed
- Average Overall Effectiveness of a Lead within a Leadsource
= Cost Effectiveness Leadsource * Ave Lead Turnover
Which are all useful, but not if it's built on shaky data. I told them my story of the car dealership where we did an in depth analysis of their lead sourcing from data in the CRM, but it didn't match what their customers were telling us about their decision process. The ownership, the brand managers, and the sales managers were all in the room when we described the mismatched story the data was telling us. Five minutes in, one of the sales managers pipes up and let me know the reason "referral" was their #1 lead source is because it is first on the dropdown list in the customer processing software. He thinks most salespeople pick it without ever asking the customer the sourcing question to move on to the rest of their voluminous paperwork.
The owner is visibly upset at hearing this. How do you think you can help us, he asks, but I he knows the answer. I suggest before changing anything with their ad budgets and campaigns, clean up the tracking. If he want to hire me to call all their customers and have them tell me the story of how they ended up at their store, I will. (but it's their job)
Lead attribution is a tricky thing. It's best to manage the process at a high level (total spend/total number of new customers = rough cost per acquisition) and do a lot of experiments until you find a lever. It's not fast, but hey, there's a reason the tenure on most VP Sales and VP Marketing positions is less than 24 months. It's easy to jump to the wrong conclusion.
In the end, we're taking the sales training company's advice under consideration, but testing more webinars and digging into the mailing process to find what works before turning anything off or shifting budget for good.
Sometimes project #1 is figuring out where you are and what works right now, today.
|