Early in my agency days I noticed something uncomfortable. We'd run a campaign, it would do well or badly, we'd write a report, and then six months later someone would run almost the same campaign and learn the same lesson again. Reports recorded what happened. Nobody recorded why, or what we'd do differently.

Years later, leading growth with a small team, I couldn't afford to relearn anything. So we built what we call a strategy tracker. It's the simplest system in our stack and probably the most valuable.

What a strategy tracker is

It's a single place where every campaign, experiment or strategic bet is logged in the same shape:

  • Hypothesis: what we believe and why. For example: "Plant heads respond better to downtime-cost messaging than to feature lists."
  • Action: what we actually ran: channel, audience, creative, budget and dates.
  • Result: the numbers that matter, such as leads, meetings, pipeline and cost, not just clicks.
  • Lesson: one or two sentences on what we now believe.
  • Next move: scale it, change it or kill it.

That's it. The discipline is in filling it in every time, which is exactly where AI helps.

Why most teams don't do this

Because it feels like admin. At the end of a campaign everyone wants to move on, and writing up lessons loses to the next launch. Our fix was to remove almost all the effort: an n8n workflow pulls the results automatically, Claude drafts the summary and the likely lesson, and the campaign owner only has to confirm or correct it. Writing a lesson went from a dreaded task to a two-minute review.

Reports tell you what happened. A tracker tells you what you now know.

What it changed for us

Faster, braver experiments

When every test is recorded, a "failed" campaign isn't wasted; it's a documented lesson. The team became far more willing to try things, because nothing was lost.

Better budget decisions

Instead of debating opinions, we could look back at what similar bets had produced. Spending moved steadily towards what converted, which is a big part of how we brought cost per lead down by around 30%.

Onboarding in days, not months

A new team member can read the tracker and understand, in an afternoon, what the company has learned about its market over years.

Better conversations with leadership

"We tried that in March, here's what happened and what we changed" is a much stronger answer than "I think it didn't work."

A simple template to start with

Create one table with these columns and nothing else: Date, Owner, Hypothesis, Channel and audience, Budget, Result, Lesson, Next move. Fill it in for every campaign for one quarter. Only then think about tools or automation.

How AI fits in, and where it doesn't

AI is excellent at the mechanical half: gathering numbers, comparing a result against similar past campaigns, drafting the summary and spotting patterns across dozens of entries. Ask it "what do our best-performing campaigns have in common?" and it will find threads a busy team would miss.

But the lesson itself has to be owned by a person. Context matters too much: a campaign might have underperformed because of a holiday, a competitor launch or a sales team that was stretched that month. The tracker works because a human signs off on what the team now believes.

Start this week

Pick the last five campaigns you ran. For each one, write the hypothesis you had, the result, and one lesson. You'll probably find at least one lesson you were about to forget. That's the whole case for a tracker.