Why evidence must begin before launch

Evidence is a design question

Teams often turn to monitoring when a donor requests a report or a launch date approaches. They ask which indicators should be tracked, then build forms around the activities already decided.

By then, the most valuable questions may have been missed. What problem are we actually solving? For whom? What change do we expect, through what mechanism? Which assumptions carry the most risk? What decision will we make if the evidence moves in one direction or another?

These are programme-design questions. That is why evidence must begin before launch.

An indicator without a decision is data waiting to become a burden.

Move from activity logic to change logic

Activities matter because they describe what a programme does. They do not explain why those actions should change a person, institution or system.

A training programme may deliver ten workshops to five hundred people. The numbers are useful, but they cannot tell us whether participants gained skill, used it, received the support to sustain it or produced a better result.

A clear theory of change connects the sequence: resources, activity, immediate output, capability or behaviour, outcome and longer-term contribution. It also reveals where the programme is relying on hope.

• Who is expected to change?
• What must they know, believe or do differently?
• What conditions enable that change?
• What could interrupt it?
• What evidence would give us confidence along the way?

Design the minimum useful evidence system

Rigour does not mean maximal data. A small programme with a complex dashboard can produce less insight than a disciplined team using five well-chosen measures and a monthly reflection.

Begin with the decisions the team must make: continue, adapt, target differently, provide more coaching, address risk, change a partner or prepare to scale. Then choose evidence that can arrive in time, with acceptable quality, at a cost the programme can sustain.

• Access: who is and is not participating?
• Quality: was the core model delivered as intended?
• Capability: can people perform without continuous external support?
• Outcome: what changed for the intended group?
• Equity: who benefits least, and why?
• Cost: what resources create the result?
• Risk: what requires immediate attention?

Give evidence a rhythm and an owner

A dashboard does not create learning. People create learning when they meet around trustworthy information, ask honest questions and have authority to act.

Decide who owns each measure, who checks quality, who interprets it and who can approve a change. Build a cadence that fits the speed of the work. Frontline teams may need weekly operational signals; leadership may need monthly performance and risk; boards may need quarterly outcome, stewardship and strategic choices.

Make space for evidence that challenges the preferred story. Record uncertainty and different interpretations instead of blending them into false consensus.

Good evidence does not merely prove value. It helps a programme become more valuable.

Launch with a learning agenda

Before launch, name the questions the first phase must answer. These should be questions with consequences—not topics for a future report.

If teacher confidence is lower than expected, will coaching increase? If some schools lack storage, will the equipment model change? If girls participate but rarely lead team roles, will facilitation be redesigned? If delivery cost exceeds the planning assumption, what options will be tested?

A learning agenda makes curiosity operational. It turns uncertainty into a managed part of implementation and creates a stronger basis for the next investment decision.

• Define the decision questions.
• Establish a credible baseline.
• Make assumptions visible.
• Collect only usable evidence.
• Create a review and adaptation cadence.
• Tell the truth about what remains uncertain.