Accuracy is the claim Fingerprint leads with, and the chart that proves it ("Accuracy dropoff, days after initial identification") is drawn by script. In the HTML a crawler or an AI assistant reads, both lines say "- 0.0 %". A buyer who asks ChatGPT "how accurate is Fingerprint versus the alternatives" gets no number from the page built to answer it.
First action
View the homepage source and list every number that only appears after the script runs.
How to measure it
Homepage demo requests
- Now
- –
- Target
- +10% relative
- By
- 6 weeks
Steps
List the hidden numbers
View source on the homepage and the main product pages; list every stat that renders only in script (the accuracy chart first).
Director of Brand & Creative 2 hours
Get the source data
Pull the accuracy test behind the chart from research: days, rates, comparison set, sample, date.
Senior Web Browser Researcher 1 day
Write the method note
One page: how accuracy is measured, against whom, and when it was last run.
Senior Data Scientist 1 day
Put numbers in the HTML
Render the values as text under the chart, with a link to the method note.
Director of Brand & Creative 1 day
Fix the comparison page
Link the TruValidate table's speed and integration claims to the same method note.
Director of Brand & Creative 2 hours
Check what assistants say
Ask ChatGPT, Perplexity and Gemini "how accurate is Fingerprint" before and four weeks after.
Director of Brand & Creative 1 hour
Why it works
Write the chart's numbers into the page as text: Fingerprint's accuracy at 30, 60, 90 and 120 days, the comparison set, the sample and the date of the test, with a link to a one-page method note. Keep the animation on top. Do the same for the other script-only stats on the homepage, and add the method note to the TruValidate comparison so its "~400ms" and "Hours to days" claims carry a source.
The Director of Brand & Creative is measured on pipeline, and a number that can be quoted is pipeline: sellers paste it into emails, analysts and AI assistants repeat it, and the enterprise buying committee (security and engineering) asks where it came from. Writing it as text costs a day.
Ready to use
ChecklistAccuracy claim checklist
[ ] The accuracy at 30, 60, 90 and 120 days is in the page HTML as text [ ] The comparison set is named (which vendors or methods) [ ] The sample size and test date are stated [ ] A method note is linked from the chart [ ] The TruValidate comparison links the same note for speed and integration time [ ] The numbers are re-tested every quarter and the date updated
Risks and how to handle them
Competitors challenge the numbers
Publish the method and the date, and re-run the test each quarter.
The real numbers look less dramatic than the animation implies
Show them anyway; a sourced number beats an unsourced curve with enterprise buyers.
Evidence
- Your websiteHomepage: "Trusted by 6000+ companies"; "$1 Billion + in fraud losses prevented in the last 12 months"; "100 Million + real-time device intelligence API events per day"; 28K GitHub stars and 7.6M monthly downloads for the open-source library; an "Install Fingerprint with AI" block (a get-started.md guide for AI agents, `npx skills add fingerprintjs/skills`, open in Claude Code, Codex or Cursor); a banner "We now detect Meta Muse". The accuracy chart ("Accuracy dropoff, days after initial identification") reads "Fingerprint - 0.0 %" and "Competitors - 0.0 %" in the page HTML, so crawlers and AI assistants see no numbers.Source
- Your websiteFingerprint vs. TruValidate: a feature table claiming "~400ms" against "Up to 2 seconds +", integration in "Hours to days" against "Days to weeks", a 14-day free trial against none, and self-service pricing against none.Source
- HiringDirector of Brand & Creative (remote, North America, posted September 2026): a new function bringing brand, content and design together under one leader for the first time. "Brand exists to drive pipeline and generate demand"; pipeline is "the scoreboard, not an afterthought". Run "a portfolio of bets, not a list of safe deliverables", own the editorial calendar, feed the demand generation testing cadence with creative concepts, and close content gaps for target accounts with Demand Gen and Product Marketing.Source
- HiringStrategic Enterprise Account Executive (US, posted September 29, 2026) and two Enterprise AEs: create net-new opportunities "rather than relying solely on inbound or SDR-generated pipeline"; multi-thread across "executives, engineering, product, security, and other key stakeholders"; run proofs of concept with Solutions Engineering.Source