Key takeaways
- Exclusions are hard rules; intended outcomes and who benefits are weighted preferences in a separate group (Pilot).
- An impact claim needs a definition as much as a revenue claim does: who counts as reached, over what period.
- Outcome evidence comes through follow-up forms and file requests, scanned and filed against the claim (Pilot).
- Delsivo produces no impact score or rating and checks no certification. People judge the impact case.
Planned feature: not in Delsivo yet, and details may change before it ships.
How impact screening differs from financial screening
An impact fund asks two questions of every deal: will it return the fund, and will it produce the outcomes the fund exists for. The second question has its own exclusions, its own vocabulary and its own kind of claim. A deck that says the product serves underbanked households is making an impact claim that needs as much definition as an annual revenue figure, and often has less evidence behind it.
That changes the work in three ways. The rubric needs a group for impact criteria. Impact claims need definitions and sources. And follow-up requests have to ask how outcomes are measured, not only what they are.
Writing impact investing screening criteria as rules
| Criterion | Type | Sample rule |
|---|---|---|
| Excluded activities | Hard rule | No revenue from activities on your exclusion list |
| Who benefits | Weighted preference | Primary users are in the groups your fund targets |
| Intended outcome | Weighted preference | A named outcome in your impact thesis, stated by the company |
| How it is measured | Question | Asked whenever the deck names an outcome without a method |
| Impact risk | Question | What could make the product's outcomes worse for the people it serves? |
If your fund follows a published impact framework, use its metric names in the rubric and forms, so criteria and claims use the same words as your reports.
Impact claims need definitions too
Take a made-up claim: reached 40,000 smallholder farmers. Reached could mean registered, active last month or paid at least once. Farmers could include buyers who are not smallholders. In Delsivo the claim keeps its assertion, a metric definition, a period, a source and an owner, and starts as Founder reported.
When the founder's survey data or a partner's records come in, an analyst marks it Corroborated with the document as its source, or Disputed if the numbers don't match, keeping both statements. The point is not to doubt founders. It is to make sure the impact case your committee reads rests on the same kind of evidence as the financial case.
Asking founders for outcome evidence
- 1Start from the impact criteria marked Unknown and the impact claims still Founder reported.
- 2Send a follow-up form asking for the measurement method, the period covered and who collected the data (Pilot).
- 3Add a scoped file request for the underlying survey, dataset or third-party report (Pilot).
- 4Use conditional questions, such as asking about a control group only when the founder says they ran one, in the first release.
- 5Review each upload and link it to the claim it supports.
Limits for impact funds
- No impact score, rating or label is computed. The rubric shows criteria and sources; your team judges.
- Delsivo doesn't check certifications or connect to impact data providers.
- Portfolio impact monitoring after investment is outside the first releases. Portfolio handoff is planned for later.
- English first at launch.