Case studies are the only marketing writing most people actually finish. Here is what four of ours changed, including the one that did not work.
We publish the numbers we agreed to measure at the start, including the ones that did not move. A campaign that raises traffic 60% while conversions stay flat has told you something important, and burying it helps nobody.
What the four campaigns had in common
The pattern that repeats across most of our work is unglamorous: fix the measurement first, then the technical debt, then spend money on acquisition. Teams that reverse that order tend to spend a lot before finding out what was broken.
In three of these four accounts, the single largest gain came before any new spend — from discovering that a chunk of existing traffic was being lost to a form that failed silently on mobile, a checkout step that timed out, or analytics that double-counted a third of sessions.
The one that did not work
A B2B client asked us to scale paid search against a 90-day sales cycle. We agreed targets on a 30-day attribution window, which meant we were optimising toward a signal that arrived long after the decision to bid. Six weeks in, the numbers looked fine and the pipeline did not move.
We stopped, rebuilt reporting around the actual cycle length, and the picture reversed: two of the three campaigns we had been scaling were buying clicks that never became opportunities. The lesson was not about paid search. It was that an attribution window shorter than the sales cycle will confidently tell you the wrong thing.
“The best marketing decision we made was agreeing what success meant before anyone spent anything.”
