Gerald Ajam

“Successful pilots should be replicated faithfully”

Scale the load-bearing parts faithfully. Let the rest adapt, on purpose.

Status Position

Published September 2026

All the claims I’m unconvinced byImplementation

Usually said

“Successful pilots should be replicated faithfully”

The shift

From: Copy everything

To: Protect the core, adapt the rest

More precisely

A successful pilot scales when its load-bearing components are kept intact and the rest are adapted deliberately to each new setting.

In short

A pilot that worked is evidence about a particular design, so it is right to be careful about changing it. But pilots run in unusual conditions, and people adapt things whether or not they are allowed to. The precise version separates the parts that carry the effect from the parts that can safely change.

The argument

From the claim to a better one

  1. The strongest formulation

    A pilot that worked is evidence about a specific design. Changing it as it spreads throws away that evidence, and uncontrolled adaptation is how effective programmes are diluted to nothing.

  2. Why it is attractive

    It is simple to specify, fund and audit. Fidelity can be checked with a list.

  3. Where it holds

    It holds for the components that carry the effect. If the mechanism depends on a dose, a sequence or a feedback cycle, those should not drift.

    Where it breaks

    It breaks because pilots run in unusual conditions: volunteer staff, extra attention, novelty and support that will not scale. Practitioners adapt anyway, and a design that forbids adaptation gets adapted badly and invisibly. The disagreement is mostly one of evidence: which parts carried the effect.

  4. What evidence matters

    Berman and McLaughlin’s RAND Change Agent study (1978) on mutual adaptation. Elmore (1996) on getting to scale, and Coburn (2003) on scale as depth, sustainability, spread and ownership rather than numbers. What matters is which components carried the effect, and whether it survives outside the pilot’s conditions.

  5. So, more precisely

    A successful pilot scales when its load-bearing components are kept intact and the rest are adapted deliberately to each new setting.

Reading it

The precise version, word by word

A successful pilot scales3 when its load-bearing components1 are kept intact and the rest are adapted deliberately2 to each new setting.

  1. Load-bearing components

    The parts the effect depends on, such as a dose, a sequence or a feedback cycle. Remove one and the result goes with it.

  2. Adapted deliberately

    Changed openly, for a stated reason, and recorded, so everyone can see what changed and check what happened.

  3. Scales

    Spreads while keeping its effect, and lasts once the extra attention of the pilot has gone.

What it changes in practice

  • Before scaling, name the components you believe carry the effect, and say why.
  • Tell each new site what must stay fixed and what it may change.
  • Collect the adaptations sites make, and treat them as data.

A case, in general terms

A reading intervention works in a pilot with twenty-minute daily sessions led by trained volunteers. As it spreads, one site keeps the materials but runs the sessions weekly to fit its timetable. Another keeps the daily sessions but swaps the volunteers for teaching assistants. The first has cut a load-bearing part. The second has adapted a surface one. Only a list of the load-bearing parts tells them which is which.

Singapore