Method
A loop, not a launch.
Campaigns end. Programs turn. Our method is a four-stage cycle that keeps repeating, each pass a little better informed than the last. This page is the long version of how it works.
- 01 OrientAgree the outcome and the constraints.
- 02 InstrumentMake sure the outcome can be seen.
- 03 RunLaunch work as tests with stated hypotheses.
- 04 RebalanceMove budget toward measured lift.
01 · Orient
Decide what winning means before spending to get it.
Most disappointing media programs were aimed at the wrong target. A cost-per-acquisition goal that ignores returns. A lead target that ignores whether sales can close them. A return-on-ad-spend figure that counts existing customers as new.
- The numberOne business outcome with an owner on your side: contribution margin, new-customer revenue, qualified pipeline.
- The guardrailsWhat we may not trade away to hit it: brand standards, discount depth, customer quality, frequency limits.
- The economicsMargin, repeat rate and payback period, so we know what a customer is worth before deciding what to pay for one.
- The unknownsA written list of what nobody currently knows. This becomes the test roadmap.
02 · Instrument
If you cannot see it, you cannot steer it.
Tracking work is nobody's favourite. It is also where the largest, cheapest gains tend to hide, because ad platforms optimise toward whatever signal they are given. Feed them duplicates or junk and they will find you more of it.
- Event auditEvery conversion event checked for duplication, missing values and consent behaviour.
- TaxonomyOne naming convention for campaigns, audiences and creative across all sectors.
- Source of truthReporting anchored to your order system or CRM, with platform data layered on top.
- BaselinesWhat happens with no change, so later results have something to be compared with.
03 · Run
Everything ships with a way to be wrong.
A campaign without a hypothesis can only produce an anecdote. We write down what we expect and what result would change our mind, before launch, where you can see it.
- Test cardsOne page per experiment: the belief, the metric, the minimum run time and the decision it informs.
- Protected baselineThe bulk of spend stays on what is proven while a defined share funds learning.
- Creative as variableAssets are tagged by hook, angle and format, so results can be read at the level of the idea.
- Cross-sector sequencingAudiences are suppressed and handed between channels, not hit by all of them at once.
04 · Rebalance
The plan is a draft until the data argues back.
On a fixed cadence we sit down with the results and move money. Sectors that are saturating give budget up. Sectors showing lift receive it. Then the cycle starts again with a shorter list of unknowns.
- Read-outThe business number first, then what moved it, then what we tried that did not work.
- ReallocationBudget shifts are proposed with the evidence beside them and approved by you.
- Stop listTactics retired this cycle, and why. A program is defined as much by what it no longer does.
- Next questionsThe updated test roadmap for the coming cycle.
Rhythm
What you will see, and when.
Cadence is agreed at the start and adjusted to the pace of your business.
Live dashboard
The business number, spend and pacing by sector, available whenever you want to look.
Short note
What changed, what we did about it, and anything we need from you. Written to be read in two minutes.
Read-out
Test results, budget proposals and the stop list, discussed live with the people who own the number.
Step back
Is the goal still the right goal? Markets move. So should the brief.
Bring us the number
you actually answer for.
Tell us what you sell, who buys it and what the business needs from media this year. We reply with a point of view, not a deck.
Start a conversation