Sample · Mainspring Data
Marketing measurement
Where new clients come from, read three ways: total ad spend against all new clients, what tracking credits each channel, and what new clients say when they sign up and are asked "How did you hear about us?" Each shows part of the picture. Read together, with tests where they differ, they show the whole.
Made-up data. Because we built it, we know the true answer. It is at the bottom of the page.
The top line
Channel by channel
People can tick several answers, then pick the one that decided it. Each answer shares one client's worth of credit: the channel that decided gets half and the others split the rest. A channel ticked alone gets the whole client. Answers are scaled up to all new clients.
The test plan
What a test can't settle yet
We ran the tests on the made-up chain
Answer key
Only made-up data has one. Real data never shows the truth, which is why the tests matter. The truth here is each channel's true share of credit, split the same way as the survey, so it won't match a test's answer exactly: a test measures the new clients who would not have come at all without the channel.
| Cost per new client | Tracking said | Survey said | Truth |
|---|