A few years ago the worst boss I ever had told me “I don’t encourage side projects.”
So obviously I spent a few evenings recently to build Pythia (live beta at https://pythia-rating.com).
It’s a single letter grade (AAA down to C) that combines five indices into one executive-friendly score: •Performance (40 %) – real CrUX field data + Lighthouse lab •Security (20 %) – modern HTTP security headers •Privacy & Tracking (20 %) – cookies, trackers, 3rd-party domains, consent signals •Sustainability (15 %) – page weight, green hosting, CO₂ estimates, image bloat •Infrastructure (5 %) – CDN usage + aggressive caching headers
Plus a fun “Amazon lost revenue” calculator (100 ms delay ≈ 1 % revenue impact, adjustable by your ARR and e-comm %).
Some illustrative point-in-time scans (as of ~13:00 GMT today, public data only, indicative):
AI companies •OpenAI: CCC •xAI: B •Anthropic: CCC •Google DeepMind: BB
(Keeping Elon happy.)
Universities • MIT: CCC • Harvard: CCC • Oxford: BBB • Cambridge: BBB (Oxbridge quietly winning on digital health.)
Everything runs on public sources (CrUX, Lighthouse, custom resilient HTTP agent for headers/cookies). No logins, no tracking on the scanner itself. Built purely as a side project because I was tired of 47-metric dashboards that no CEO ever reads. Wanted one number + dollar impact + actual fix suggestions.
Would love your brutal feedback on: •Scoring weights (do they feel right?) •Missing metrics you’d want to see •Edge cases it gets wrong •Who else this could actually be useful for (agencies doing client audits? PE doing diligence? in-house teams benchmarking competitors?)
Happy to answer questions or add anything people ask for. Thanks for looking!