[ Phase 02 — Plan ]

Strategy without an experiment register isn't strategy.

Generic agency plans are 60-slide decks built to justify a retainer. Ours are working documents — falsifiable hypotheses, kill dates, channel-mix decisions tied to your P&L. The plan is the operating manual for the next 90 days, not a one-time artefact.

5–7 working days 2 working sessions + async review Strategy doc + experiment register + channel mix model Owner · Senior strategist + lead operator

What planning actually means

The plan converts the audit's prioritised findings into a sequenced execution map. Every initiative has a hypothesis, success criteria, and a kill date. Every channel has a budget tied to payback math, not platform recommendations. The plan is reviewed quarterly and revised when data demands it — not when the calendar does.

[ Artefacts ]

What this step produces.

P01 · 90-day strategy doc .pdf · 12–20 pages

Channel-by-channel playbook for the next 90 days. North-star metrics, leading indicators, decision rules, hand-offs between channels. Written for the founder + leadership team to read in 15 minutes.

P02 · Experiment register Notion · live document

Every test currently in flight, with: hypothesis, primary metric, expected lift, sample size, start date, kill date, status, learnings. Visible to your team. Updated weekly. Outcomes archived for cross-engagement learnings.

P03 · Channel mix model .xlsx · live model

Spend allocation across channels modelled against payback period, blended ROAS target, marginal incrementality. MMM-supplemented at $1M+ annual spend. Updated quarterly with actual performance feeding back.

P04 · Creative production roadmap .xlsx + Figma board

Hypothesis-driven creative briefs, production pipeline, message-market hypotheses, kill dates per asset. Standardised brief template enforced across paid social, paid search, lifecycle.

P05 · Hiring + tooling recommendations .pdf

What to hire, when to hire, what to consolidate in tooling. Headcount mapped to revenue thresholds. Tooling consolidated against actual usage data we pulled in audit.

[ Timeline ]

How the days run.

Day 1

Strategy intake

Working session with founder + leadership: business goals, P&L constraints, board commitments, growth thesis. The plan is built backward from these — not from platform best practices.

Days 2–3

Hypothesis generation

Audit findings translated into testable hypotheses. Each: 'we believe X will lift Y by Z%, measured by W, killed if not by date D.' 25–40 hypotheses generated, sequenced by impact × effort × confidence.

Days 4–5

Channel mix modelling

Spend allocation modelled against payback math. Marginal channel efficiency curves drawn. MMM-supplemented where applicable. Mix proposed with explicit confidence ranges.

Day 6

Strategy doc draft

Plan written. 12–20 pages. Reviewed by senior reviewer for cross-channel coherence. Drafts shared with founder for feedback.

Day 7

Strategy review + sign-off

90-min strategy review with leadership. Refinements. Sign-off on Q1 hypotheses + channel mix. Experiment register populated and live.

[ Who does what ]

Named owners, not "the team".

RoleLeadsContributesConsumes
Senior strategistStrategy doc + hypothesesChannel mix, creative roadmapAudit findings, founder context
Lead operatorChannel mix + experiment registerHypothesis prioritisationStrategy doc
Founder / leadershipBusiness context, sign-offConstraints, goals, must-havesStrategy doc, channel mix model
In-house teamExecution capacity inputDomain knowledge, dependenciesExperiment register, creative roadmap
[ Rules ]

Non-negotiables for this step.

  • Every hypothesis is falsifiable.
  • Every test ships with a kill date.
  • Channel mix decisions tied to a P&L line.
  • No more than 8 simultaneous tests on one channel.
  • If the plan changes, the change is logged with a reason.
  • Plans are living documents. Quarterly revisions. Daily-decision-friendly.
B2B SaaS · Series-B · $7M ARR
The plan was the first marketing doc we'd seen with kill dates. Every prior agency gave us a 60-slide quarterly deck and asked for the same retainer. RoiSpends gave us a working register we still update weekly.

31 hypotheses tested in Q1. 19 wins, 11 kills, 1 inconclusive. Cumulative paid efficiency +43%.