Impact Ventures analytics platform
Of hundreds of startups tackling the world's biggest problems, which ones deserve capital first?
The question
A philanthropic venture fund runs an open competition for startups addressing global problems. Hundreds apply. The partners needed a defensible, repeatable way to rank them, and to explain that ranking to the people writing the checks.
Collect
Application data arrived as long-form text, decks and spreadsheets. We built a data-entry and enrichment layer that normalized every submission into one schema, pulled public signals on each company, and kept every raw source alongside the structured record so any score could be traced back to its evidence.
Analyze
Each startup was scored across a rubric the fund could edit. Language-model classifiers handled the first pass on the unstructured material; human reviewers handled the calls that mattered. Scores were stored as full distributions, not just a final grade, so the partners could see how confident the model was and where reviewers disagreed.
Present
An interactive dashboard let the partners sort, filter and compare the field, drill into any company’s evidence, and export a one-page brief for the investment committee.
Impact
The platform guided the fund’s investment decisions across the competition and became the template for its later cohorts.