The Reads
Short, bounded diagnostics: the place to start.
A Read is a focused engagement: six to eight weeks, a fixed scope, a single question answered well. Independent by design: we don’t bid on the evaluations or programs we assess.
They differ only in the question: whether to generate new evidence, match your claims to the right grade of evidence, or stress-test a study before you commit.
The Counterfactual Hunt
Most programs already contain credible natural experiments (eligibility cutoffs, staggered rollouts, capacity limits) that are rarely looked for. We find the strongest, specify them fully, and say plainly where the data won’t support one.
The Evidence Map
We take every impact claim in your portfolio and establish the strongest version the evidence beneath it can support, so you hold a defensible map of where your evidence stands. Per-grantee findings stay confidential to you.
The Pre-Mortem
Before you commit to a major evaluation, we stress-test the design for what will go wrong (statistical, contextual, ethical) and pair each risk with a fix, while there’s still time to act.
After a Read
The larger engagements.
Many engagements begin with a Read. When a program is live and worth measuring continuously, the next step is Field, an ongoing measurement function built into how the program runs. It runs on the same method, and to the same standard, as the Read.
How to start
A short conversation first.
Tell us the decision in front of you and when an answer would still be useful. The first conversation is a fit check: we’ll tell you which Read fits, or that none does. I take these myself.
Start a conversation