The 30-day first engine
What actually ships in the first 30 days when Loiale starts with a brand. The artefacts, the decisions, the measurable output, and the trap of confusing discovery with delivery.
Most engagements between a DTC brand and an outside team lose the first ninety days to discovery. Slides, workshops, roadmaps, more slides. The first thing shipped to production is usually a report about what should be shipped later. We refuse to run engagements that way, because the pattern is not an accident. It is what happens when the team on the other side is not accountable for the number.
The first 30 days at Loiale are the first engine, live in production, producing measurable lift against a holdout, before the second month begins. This is the plan we run, the artefacts we produce, and the tests we apply to know it worked.
Days 1-7. Audit that fits on one page
The audit is not a fifty-slide document. It is a one-page map of the data model, the identity graph, the event coverage, and the stack. If we cannot fit it on one page, we do not understand it well enough yet.
- Every data source inventoried with owner, refresh cadence, and trust score.
- Every event mapped from source to destination, flagged where names differ.
- Identity graph: how email, phone, device, and order relate today.
- Consent model: where consent is stored, who reads it, who ignores it.
- Top three risks flagged, with the cost of leaving each one open.
Days 7-20. Build the layer, ship one flow
Two things happen in parallel. A minimum viable capture layer is stood up, and one activation flow is shipped end to end. Not three. One. The one that unlocks measurable revenue with the least new infrastructure.
| Brand shape | First flow | Why |
|---|---|---|
| Replenishable product, mature CDP | Modeled replenishment | Highest lift, lowest new infra |
| Fashion / apparel, weak identity | Post-purchase winback | Forces identity fix, contained scope |
| Luxury / low frequency | VIP tier + concierge | Margin over volume, defensible |
| Subscription hybrid | Churn recovery pre-cancel | Fastest visible P&L impact |
Every flow ships with a 5-10% holdout, an event log the analytics team can query, and a written hypothesis. If any of the three is missing, it is not live.
Days 20-30. Instrument, measure, hand the tiller
The last ten days are about the interface between the engine and the operator on the client side. One dashboard, three metrics, weekly review. Not a war room. A twenty-minute cadence the CEO can hold themselves.
- Dashboard live, refreshed daily, three top-line metrics on the front page.
- Holdout comparison automated, no manual recalculation.
- Runbook: how the flow is paused, edited, and retired.
- Named owner on the client side, with time booked on their calendar.
What we are deliberately not doing
We are not launching six flows. We are not migrating the ESP. We are not redesigning the loyalty program. We are not building a data lake. The temptation in the first 30 days is to fix everything visible. The discipline is to fix one thing well enough that the second one becomes obvious.
"The first engine is not the biggest one. It is the one that changes what the second engine can be."— Loiale internal playbook
The three tests we apply at day 30
- Is there a measurable lift against the holdout, or a clear reason there is not yet.
- Can the client-side owner explain the flow, the metric, and the next test without us in the room.
- Is the data layer in a shape where the second flow is faster than the first.
If all three answers are yes, month two begins. If not, we do not compound the mistake by adding more scope. We fix the failing test first, then move.