Every year a platform tells me the old way is finished.
Google rewrites bidding. Meta rewrites targeting. Tracking breaks, an attribution model is retired, and now AI is rewriting parts of the job itself. In 14 years of performance marketing I haven’t had a single year where the rules stayed still.
So what actually stays?
Not the tools
Tools make work faster. Some of them are genuinely good. But a tool doesn’t decide what is worth doing, and it can’t tell you whether the number on the screen means what it appears to mean.
A dashboard will report return on ad spend to two decimal places. It won’t tell you whether the sale would have happened anyway. An agent will draft the weekly report in minutes. It won’t tell you which line in that report should change a budget.
That part is still ours.
The questions do
Three questions have outlasted every platform change I’ve worked through:
- What is the business actually trying to achieve?
- What do we really know, and what are we assuming?
- What would change our mind?
They sound obvious. Most accounts I’ve been asked to look at had skipped at least one of them.
Why this matters more now
When the platforms automate more, the work that’s left is judgement. Which conversion do we tell the algorithm to chase? Which result do we trust? Which test is worth a month of budget?
Get those wrong and automation just makes the mistake faster.
Limits
This is a belief, not a finding. There’s no data in it. Read it as the lens the rest of this journal uses.