Get technographic data on any market with Claude.

Technographic data is the software a company runs. Data vendors sell it. For most markets you can read it straight off each company's own website, and Claude will do the reading for you.

I did this for every golf course in the US. 12,339 courses, 3,044 booking systems identified, $0 in data costs. Here's what technographics are, why they're worth the effort, and exactly how the run worked.

What technographic data is

Firmographics tell you how big a company is and where it sits. Technographics tell you what it runs. The CRM, the website platform, the payment system, the booking widget, the chat tool.

Most of that leaves fingerprints in public. A booking button that links to the vendor's domain. A script tag that loads from the vendor's servers. A "powered by" line in the footer. If a browser can see it, Claude can read it.

Why it matters

The software a company runs tells you who they are. In golf, it tells you the business model.

Private clubs mostly run member software. 42% of the private clubs we could read are on ClubEssential, 26% on Jonas, 16% on MembersFirst. That's 83% on three platforms, and none of them sell public tee times. Public courses run tee-sheet software instead. foreUP alone has 36% of them.

So one scraped field sorts the whole market. You know who to pitch, which competitor you're replacing, and what the email should say. That's the difference between a list and a market map.

Real rows from the golf run. Where the site didn't show clear evidence, the cell stays blank.

How the golf run worked

1. Build the list. Every golf course in OpenStreetMap, pulled free through the Overpass API. After dropping unnamed polygons and duplicates, 12,339 named courses.
2. Find the websites. OpenStreetMap had some. Serper's free tier (real Google results as an API) found 2,401 more. 9,101 courses ended up with a website.
3. Read each site. Fetch the homepage, then follow one link that looks like booking or tee times. That second hop matters. The widget often lives on the booking page, not the homepage. The pass that added it found 1,089 more.
4. Match fingerprints, then let Claude read the rest. A simple pattern list (foreupsoftware.com, clubessential, jonasclub and so on) caught 2,800. Then Claude read the pages the patterns missed and found 244 more, including vendors I didn't have on the list.

Claude also caught 577 course websites that were dead, parked, or worse. 128 of them had been taken over by gambling and casino spam. A pattern match would have happily called those valid.

The prompt

This is the shape of what I give Claude Code. Swap in your market.

The last two rules do the most work. Evidence-only keeps the data honest. Blank beats wrong, because a wrong vendor sends the wrong email.

Scraping it yourself vs BuiltWith

BuiltWith is the best-known technographics tool. Single lookups are free. Lists start at $295 a month on the Basic plan. It's great for web-stack tech like analytics, ad pixels, and site builders.

The case for doing it yourself is control. You pick the fingerprints, so niche vertical software counts as much as the big names. You follow the booking link a crawler might skip. And you get the evidence for every row, so you can check it. If the thing you sell against is a vertical tool, look it up on BuiltWith before you pay. If it's not tracked, this is how you get it.

Where it breaks

Some widgets only load with JavaScript, so a plain request sees nothing. That was about 624 golf sites. A headless browser fixes most of those. Some sites block anything automated. CloakBrowser or a proxy network like Bright Data gets you through. And some companies just don't show their software in public. Those stay blank.

New to all of this? Start with how to scrape websites with Claude.