Measurement before marketing
Why fixing your attribution layer before running another campaign is always the right call — even when it costs you a quarter.
Every growth engagement we run begins the same way: we ask to see the attribution data before we look at the campaigns. The response is almost always the same too — a spreadsheet assembled from three different platforms, a GA4 dashboard that stopped being trusted eight months ago, and an explanation of why the numbers do not add up that involves the words 'probably' and 'approximately' more often than they should.
The temptation, when you are behind on pipeline, is to fix the measurement later and run more campaigns now. This is the wrong decision almost every time. Campaigns running on bad data optimise toward the wrong signals, which means they do not improve — they just become more efficient at doing the wrong thing.
The instrumentation tax
We call the upfront cost of fixing measurement the instrumentation tax, and we tell every client to pay it before committing budget to channels. The tax is real — it costs time, engineering resource, and often a quarter of performance data while the new system stabilises. But the payoff is that every subsequent decision is made on information you can trust.
Campaigns running on bad data optimise toward the wrong signals. They do not improve — they become more efficient at doing the wrong thing.
The CS-04 engagement in our case studies is the clearest example. Six weeks of instrumentation before a campaign ran. The attribution rebuild exposed that forty percent of the existing paid budget was generating pipeline that the CRM was attributing to organic. Reallocating that budget — based on the corrected data — drove the majority of the 184% pipeline increase in year one. The measurement investment paid back in month three.
More notes
The brand stack in the age of AI
Why identity, voice, and visual systems matter more, not less, when generation is free and undifferentiated.
ReadOperating equity, in practice
How Famous structures venture engagements: the cap table, the carve-outs, and why we charge less than market upfront.
ReadEval-driven retrieval, in practice
A pragmatic framework for shipping retrieval systems that hold up in production after the launch.
Read