Organize the month around one consequential decision
Early-stage execution becomes noisy when every idea becomes a project. A useful founder execution plan begins with one decision the company must earn: whether to serve a segment, continue a product direction, invest in a channel, change pricing, or build the next workflow. The month should produce enough evidence to make that decision, not just a longer activity log.
Give each week one job
| Week | Primary job | Evidence produced |
|---|---|---|
| 1 — Focus | Choose the customer, problem, baseline, and riskiest assumption | Written hypothesis and qualified target list |
| 2 — Test | Put the smallest credible offer or workflow in front of customers | Observed behavior, objections, and commitments |
| 3 — Deliver | Help a narrow cohort reach the promised outcome | Activation, quality, support load, and repeat use |
| 4 — Decide | Compare results with thresholds and document the next bet | Go, pivot, hold, or stop decision |
Do not turn weekly jobs into departments. One person can research, sell, deliver, and analyze at this stage. The separation exists to protect sequence: choose the risk before launching work, observe before automating, and decide before expanding scope.
Use a weekly plan that fits on one screen
Protect maker time with a light operating cadence
- Monday, 30 minutes: set the outcome, evidence target, three moves, and stop-doing list.
- Tuesday through Thursday, 10 minutes: check the leading measure and remove one blocker; do not re-plan the whole week.
- Daily, one protected block: complete the hardest customer-facing or product-learning task before internal cleanup.
- Friday, 45 minutes: review evidence against the prewritten threshold and record the decision or open question.
Batch communication and administration after the protected block when possible. If an urgent task repeatedly displaces the evidence-producing work, treat that pattern as an operating constraint to solve rather than a personal productivity failure.
Track one outcome metric, one leading measure, and two guardrails
| Metric type | Purpose | Example |
|---|---|---|
| Outcome | Shows whether the customer or business state changed | Qualified users who complete the core job and return |
| Leading | Shows whether the team is creating enough chances to learn | Observed onboarding sessions or offers sent |
| Quality guardrail | Prevents a headline number from hiding a broken result | Successful result rate or support incidents |
| Cost guardrail | Keeps the test inside an acceptable bet | Delivery hours per customer or acquisition spend |
Vanity metrics become dangerous when they substitute for the decision metric. Traffic can support a channel test but does not prove activation. Signups can support message clarity but do not prove repeat value. Always connect the number to the behavior it is meant to predict.
Keep a decision log so the company can learn
The log is not bureaucracy. It stops the team from reopening settled questions without new evidence and makes prediction errors visible. Over time, it also shows whether the company tends to overestimate demand, underestimate delivery cost, or hold weak experiments too long.
End the month with a decision, not a retrospective alone
- Go: the evidence clears the threshold; name the next constraint and investment.
- Pivot: the problem is credible but the segment, promise, workflow, price, or channel needs a new test.
- Hold: an external dependency prevents a fair test; define the trigger and do not keep spending by default.
- Stop: evidence remains below the threshold; preserve the learning and release the capacity.
A stop decision is productive when it arrives before a large irreversible investment. A go decision is only useful when it names the next evidence requirement. Either way, the operating system should make the following month narrower and more informed than the last.
Frequently asked questions
What should a founder focus on in the first 30 days?
Focus on one consequential decision and the evidence needed to make it. That usually means a specific customer, painful problem, narrow offer or workflow, qualified cohort, and a prewritten success threshold.
How many goals should a startup have each week?
Use one weekly outcome supported by no more than three high-leverage moves. Other work can still happen, but it should not compete with the evidence-producing outcome.
What belongs in a founder decision log?
Record the decision, context, evidence, remaining assumptions, options considered, rationale, expected result, and the date or trigger that would justify revisiting it.