A methodology we have run many times, in one order.
Every engagement moves through the same four phases, because the failure modes of paid acquisition are remarkably consistent: spend scaled on broken measurement, structures that fight the platform's learning systems, and reporting nobody trusts. The order below exists to prevent each one — deliberately, and in sequence.
Diagnose
We start by establishing what is actually true. That means a full account audit across every active platform — structure, bid strategies, conversion actions, audience layers, budget distribution — and a tracking audit that traces each conversion from click to database record, looking for duplication, gaps, and consent-related signal loss.
In parallel, we build your unit-economics baseline: LTV by cohort, CAC by channel and geography, payback windows, and contribution margins. Where you are entering new markets, market research sizes demand and auction economics per candidate geography.
The phase ends with a written diagnosis and a sequenced plan. If the diagnosis is "you don't need us", we say so — it has happened, and those companies refer us to others.
Instrument
Measurement is fixed before budgets scale — never after. Depending on the diagnosis, this phase covers server-side GTM deployment, conversion action architecture, Enhanced Conversions, CAPI and Events API integrations, webhook-based purchase tracking with order-ID deduplication, and Consent Mode v2 configuration. The full scope is described on the measurement & analytics page.
Nothing here is optional ceremony: every later decision — bid targets, budget shifts, creative verdicts — rests on this data. We would rather delay a launch two weeks than scale on a signal we don't trust.
Launch & structure
Campaign architecture is designed around how each platform's bidding actually learns: consolidated where the algorithm needs volume, separated where economics differ — by product line, margin tier, or market. Budget allocation follows the payback model from Phase 01, and geographic rollouts are tiered so proven markets fund experimental ones.
Where creative matters — Meta and TikTok especially — the creative pipeline is set up now: briefs, testing matrix, production cadence, and the ledger that records what has been tried and what won.
Scale & iterate
Steady state is a weekly optimisation cadence: bid strategy maturation as conversion volume grows, budget rebalancing against marginal CAC, creative refresh on evidence, and script-based monitoring that catches anomalies within hours. Periodic incrementality checks — holdouts and geo tests — keep platform-reported numbers honest.
Monthly, we review the economics with your team: not "impressions were up", but where payback stands, which markets earned more budget, and what we are changing next month and why.
Weekly summaries
A short written update every week: what changed, what we saw, what we are doing about it. Written to be read in three minutes, with the numbers attached for those who want them.
Monthly deep-dives
A working session on the economics: payback by channel and market, creative learnings, and the next month's plan. Your finance lead is welcome — the reporting is built for them too.
Shared dashboards
Live dashboards in accounts you own, showing the same numbers we look at. No proprietary reporting portal, no data hostage-taking, nothing that disappears if we part ways.