3 ms·
Show HN: Measuring brand share in AI answers – a Y Combinator case study
After working in data science at Google, I built GeoVector to systematically measure how brands appear in AI-generated answers. Our approach is research-based, using position-adjusted scoring grounded in published GEO literature. The Y Combinator report is one example of the analysis we run.
We ran 150 prompts across ChatGPT and Gemini, tracking 21 brands. Three things that surprised us:
1. Techstars outranks YC on ChatGPT despite YC's far stronger Google presence
2. YC's own site accounts for just 8 of 940 AI source references
3. The single most-cited source driving competitor visibility is a blog post on pitchwise.se — not any accelerator's own website
Full report at the link, no signup. GeoVector runs this analysis for any brand or vertical.
- zinc_philip 7mo agoThe YC report is based on 150 prompts. However when I request a report, it's only 15?
- roman10 7mo agoYeah, 15 is the free tier prompts limit. Unfortunately we can't support 150 prompts free of charge because of the infra cost etc. If you're keen to collaborate and explore, please reach out to hello@geovector.ai, and we can work something out. :-)
- zinc_philip 7mo agoMakes sense. 15 prompts are fine to get started for my case. Thanks for offering it for free! Will teach out if I need to expand the coverage.
- zenon_paradox 7mo ago[flagged]