Prova for foundations

Decision-grade evidence across the whole portfolio, not only the grants big enough to evaluate.

You already know what good evidence is worth, and you can rarely afford it past a few flagship grants. Prova exists to change that arithmetic: rigor you can extend across the portfolio, on a timeline that fits how you fund, from a partner you can hold to a published standard.

The problem you know well

Rigor gets rationed to the grants that can carry its cost.

A real evaluation costs enough that you can commission it for the flagship and little else, so most of the portfolio is known by self-report. The studies that do run often land after the funding decision they were meant to inform. And evidence paid for by the funder carries an incentive problem you can see as clearly as anyone.

None of this is a failure of will. It is the price of proof, the constraint Prova is setting out to move.

From the Founder

Ravi Narain

Where I’ve sat

For nine years I designed and measured social programs funded by corporations, foundations, family offices, and individual donors, including standing up philanthropic foundations from a blank sheet and building the multi-year theses behind them.
The work was, in effect, a privately funded policy testbed: real interventions that generated evidence built to be legible to the people deciding whether to scale them, and we competed against the Big Four on the rigor of that evidence.
I had our work evaluated independently, and I co-built a 42-factor framework that forced outcome-accountability and learning-from-failure into the structure of every engagement.

What it taught me

Why rigor gets rationed to the flagship grants: a credible study has cost enough that most of a portfolio ends up measured only by what grantees report about themselves.
The incentive problem in funder-paid evidence, and why an independent read, from someone with no stake in the answer, is worth more than a flattering one.

What Prova does

Rigor you can extend across the portfolio.

Prova brings decision-grade evidence to grants that could never have justified a traditional evaluation, by working from the records programs already keep and bringing the cost of credible reading down far enough to cover more of the portfolio. The standard of proof does not move; what moves is the price of reaching it.

The standards behind the work are written down: how we grade a claim against its evidence, how we find a credible natural experiment, how we stress-test a design before it runs. Read them, or hand them to your evaluation lead.

We are independent, and we do not bid on the evaluation work we review, so the read you get has no stake in its own conclusion.

This extends what your team can reach, and it does not replace their judgment. The hardest calls, what an effect means in a place and what a portfolio should do about it, stay with the people who hold them.

AI does the heavy lifting, under human supervision, which is what makes the cost fall; the judgment stays human.

The method → Read a standard →

How to start

Start with a Read.

A Read is a short, bounded first step, measured in weeks. The natural place to begin is the grant that has always been too small to evaluate properly: a single program, read credibly, for a fraction of what a full study would cost. It is the cost-of-proof thesis, tested on one of your own grants.

From there, the same method scales: Field for a live program worth following, and a full causal study when a question warrants one. All to the same standard.

See the Reads → Field →

Start a conversation

Bring your hardest grant to measure.

If there is a corner of the portfolio you have never been able to see clearly, that is the conversation to have. The first one is a fit check, with no obligation.

Start a conversation