The homepage hero carries the firm's four headline numbers, and in the HTML the server delivers they read 3,390 healthcare orgs processed, 2,400 contacts surfaced, up to 2 decision-makers per org and 450 outbound emails a day. The case studies say 11,300, 8,000, up to 5 and 1,500. Every hero figure is exactly 30% of the real one, which is what a count-up animation looks like before it runs. A browser presumably animates up to the full figures; anything that reads the page without running the script does not. On the custom engines page the same pattern leaves every counter at 0: "Relationships scanned 0", "0 meetings booked". Meanwhile robots.txt says "AI search and answer engines are welcome" and names ten of them, and most of those readers take the text as delivered. The firm sells data accuracy, and the first numbers a machine reads about it are wrong by 70%.
First action
Open the homepage with JavaScript disabled and write down the four hero numbers next to the four in the case study.
How to measure it
Proof numbers correct without scripts
- Now
- –
- Target
- 5 of 5 pages
- By
- 1 week
Steps
Reproduce it
Load the homepage and /gtm-engines with scripts off; record every number that differs from the case studies.
Web developer 30 minutes
Render final values
Put 11,300, 8,000, up to 5 and 1,500 in the server HTML; the count-up becomes a visual effect that ends on the same text.
Web developer 2 hours
Fix the engine counters
Replace the zeros on /gtm-engines with labeled sample figures or remove the numerals.
Web developer 2 hours
Respect reduced motion
Skip the animation when the visitor's device asks for less motion.
Web developer 1 hour
Add structured data
State the same figures in the organization and case study markup.
Web developer 2 hours
Reconcile every mention
Check the hero, case studies, service pages and llms.txt use the same numbers and labels.
Founder 1 hour
Ask the assistants
A week later, ask three AI assistants what the firm has done in healthcare and note the figures they give.
Founder 30 minutes
Why it works
Render the final values in the server HTML (11,300, 8,000, up to 5, 1,500) and let the animation start from there visually, or animate with CSS only, so the text in the page source is always the true figure. Do the same for the engine counters on /gtm-engines: show a labeled sample figure or remove the number, never 0. Then check all five pages with scripts turned off, add the same figures to structured data, and confirm the hero, the case studies page and llms.txt all agree.
Search snippets, link previews, AI answers and many accessibility and reader tools use the delivered text. One consistent set of numbers across the hero, case studies and llms.txt gives those systems a single fact to repeat, and it removes an easy objection for a buyer whose own problem is inaccurate records. Visitors who prefer reduced motion also get the right number immediately.
Ready to use
ChecklistNumbers-without-scripts check
[ ] Homepage hero reads 11,300 / 8,000 / up to 5 / 1,500 with scripts off [ ] /gtm-engines shows no counter at 0 [ ] Sample figures keep the "invented for illustration" label [ ] Animation skipped for reduced-motion visitors [ ] Same figures in structured data [ ] Hero, case studies and llms.txt agree [ ] Link preview and search snippet re-checked after a week
Risks and how to handle them
The animation loses its effect
Keep it as a visual layer over text that is already correct.
Sample figures on the engines page are read as real results
Keep the existing "invented for illustration" label next to each one.
Evidence
- Your websiteHomepage (read twice on November 18, 2026): title "Blue Modern Advisory | GTM Engineering & RevOps for Healthcare", H1 "We build the engines behind your pipeline." In the HTML as delivered, before any script runs, the hero statistics read "3,390" healthcare orgs processed, "2,400" contacts surfaced for validation, "up to 2" decision-makers mapped per org and "450" outbound emails a day. The case-study block further down the same page reads "11,300", "8,000" and "Up to 5". Each hero figure is 30% of the published one.Source
- Your websiteCustom engines page: warm network engine, Slack reply desk ("Auto-send drafts scoring 90+") and event and conference engine ("Meetings booked before the badge pickup."). In the delivered HTML the counters read "Relationships scanned 0", "Warm paths found 0", "0 attendees", "0 best-fit customer profile matches", "0 warm paths" and "0 meetings booked". The page states "People, companies and figures in these engines are invented for illustration." Its FAQ says that for a healthcare client the warm network engine matched contacts to Director-level and above buyers in Managed Care, Payer Relations and Payer Contracting.Source
- Your websiteCase studies page: two anonymous case studies. Healthcare: a provider data set that began with about 350 organizations and 3,000 child NPIs across three states and grew to "11,300" healthcare organizations processed, "8,000" contact records surfaced for validation and up to 5 decision-makers per organization, using 30,000+ enrichment credits. B2B GTM: a team that could enrich about 400 provider-group leads a week moved to a model supporting 1,500 outbound emails a day (about 750 net-new and 750 recycled), needing about 15,000 leads a month, starting from 35,000 existing leads. Both report volumes processed; neither reports meetings, pipeline or revenue, and no client is named.Source
- Your websiterobots.txt and llms.txt: robots.txt says "AI search and answer engines are welcome" and names ten AI crawlers; llms.txt describes a Brooklyn, NY GTM engineering and RevOps firm and gives the proof points as 11,300 healthcare organizations and 1,500 outbound emails a day.Source