Technographic data is information about the technology a company uses: the software, platforms and tools that make up its tech stack, usually with a note on when each tool was seen and how sure the source is. Where firmographics tell you who a company is, technographics tell you what it runs.

For a B2B seller that's often the more useful half. A company paying for your biggest competitor, a company running three tools that plug into yours, and a company spending on ads with nothing to measure them all deserve a different first message, and none of that shows up in an industry code or a headcount band.

The catch is that nobody publishes their stack. Every technographic record is a detection made from the outside, and the teams that get burned are the ones who read those detections as an inventory. A tool that was spotted once, a year ago, by a crawler that only saw the homepage, is a hint and nothing more.

TL;DR

• A good record names the tool and its category, where it was detected, when it was first and last seen, and a confidence level. Without the dates and the source you can't tell a current fact from an old guess.

• Firmographic data narrows the market to companies that could buy. The stack shows which of them have a problem you solve right now, so use firmographics to filter first and technographics to sort what's left.

• Providers detect tools four ways: code on the website (BuiltWith, Wappalyzer), public DNS and email records like SPF, job postings (TheirStack and others), and partner lists, case studies or surveys. Each one sees a different part of the stack.

• A missing tool only counts as evidence when that tool would normally leave a trace. No ad pixel on a public site means something. No ERP detected means almost nothing, because back-office software rarely touches the website.

• The stack supports four plays: displacing a competitor, selling into an integration, filling an obvious gap, and acting when a tool is added or dropped. Let the detection shape the angle of your message, and don't quote the prospect's own scripts back at them.

• Store the handful of tools that change a decision as CRM properties, enrich inbound leads on arrival and outbound accounts before a sequence, and re-check detections before they drive a campaign.

What one technographic record actually contains

Strip a provider's export down and each row usually carries the same few fields:

• The tool and its category. HubSpot under marketing automation, Intercom under live chat, Snowflake under data warehouse. Categories matter more than you'd think, because most plays are built on "uses anything in category X" rather than on one product.

• The detection source. A script on the website, a DNS record, a job posting, a partner directory. The source tells you how the tool was seen and what kind of mistake is likely.

• First seen and last seen. When the tool was first spotted on the company and when it was last confirmed. BuiltWith's free API documentation goes a step further and separates technologies "considered live on the domain" from ones "considered no longer live", which is exactly the distinction you need before you pitch a replacement.

• Confidence. Some providers give a score or a high, medium and low label, built from how many signals agree and how recent they are.

The dates are the fields people skip, and they're the ones that bite. A row that says a company uses a marketing platform tells you very little until you know whether that was confirmed last week or eighteen months ago. When a provider's export doesn't include them, ask for them, and if they aren't available, treat every row as older than it looks.

Firmographics say who could buy, the stack says why

Firmographic data describes the company itself: industry, headcount, revenue, location. It's how you decide that a 200-person logistics software company in the Benelux belongs in your market. What it can't tell you is whether that company has the problem you solve, and two companies with identical firmographics can be completely different prospects.

The stack fills that in. If one of those two companies runs a competitor's product under a contract that renews next spring and the other runs nothing in your category at all, you now have two different conversations, and probably a different priority for each. The firmographic entry lays out how firmographic, demographic, technographic and intent data split the work. The practical order is simple: firmographics decide which accounts are eligible, and technographics decide which of the eligible ones you work first and what you say to them.

Where technographic data comes from

How a tool was detected tells you how far to trust it, so it's worth knowing the four routes most providers use.

Code on the website

This is the classic method behind BuiltWith and Wappalyzer. A crawler loads the company's pages and matches what it finds (script tags, cookies, response headers, meta tags) against a library of fingerprints. A HubSpot tracking script, a Google Ads tag or a Hotjar snippet each leaves a recognisable pattern. It's reliable for anything that has to run in a visitor's browser: analytics, ad pixels, chat widgets, consent banners, the CMS. The HTTP Archive, which uses Wappalyzer for its yearly Web Almanac, states the limit plainly in its methodology: unless a tool leaves detectable fingerprints in the website's code, its usage can't be measured.

DNS and email records

A company's public DNS says a lot about its email setup. The MX record shows who handles incoming mail, and the SPF record lists the services allowed to send mail for the domain. Google's own SPF setup guide tells Workspace customers to publish include:_spf.google.com, and to add the domains of any other senders they use. So an SPF record that also includes a marketing platform or a support tool is the company telling the world which systems send email on its behalf. We read these by hand more often than you'd expect, because they're free and current.

Job postings

Hiring ads name tools all the time: "Salesforce admin", "experience with dbt and Snowflake", "you'll own our Gong setup". TheirStack builds its whole technographic dataset this way, and its documentation is refreshingly direct about the limit, under the heading that a technology that isn't mentioned is not a technology that is absent. Job postings are the best route to back-office tools (ERP, finance, data, CRM admin work) that never appear on a website. They're also the weakest for companies that aren't hiring.

Partner lists, case studies and surveys

Vendors publish customer logos and case studies, integration marketplaces list who has installed what, and some providers survey buyers directly. Larger providers combine these with their own models. HG Insights, for example, says its method does not rely solely on web scraping or digital signatures, and claims to surface technology used behind the firewall. Ask any provider which of these routes a given detection came from, because the answer changes how much weight it deserves.

Absence is only evidence for tools that leave a trace

Most of the useful plays rest on a gap: the company has paid ads but no attribution tool, or a growing sales team but no sales engagement platform. That means you're trusting what the data didn't find, which is the shakiest thing a technographic record can tell you.

The rule we use is to ask whether the missing tool would normally be visible from the outside. Attribution and analytics tools usually run a script on the public site, so if a crawler checked the site recently and found none, the gap is probably real. An ERP, a finance system or a data warehouse almost never shows up on a website, so "not detected" says almost nothing about whether they have one. A CRM sits in between. You'll often catch it through a tracking script, an embedded form or a job ad, and you'll often miss it entirely.

The same caution applies to things the data did find. Stacks change at contract renewals, and a detection that hasn't been reconfirmed in a year may describe a tool the company dropped. A script can also outlive the subscription, sitting in the site template long after anyone logs in. So read each detection as a probability. High confidence and seen recently is close to a fact. A single sighting from last year is a reason to look, not a reason to write.

When two providers disagree about the same company, that's usually two methods seeing different things rather than one of them being wrong. A crawler and a job-postings source can both be right. Running them in sequence through a waterfall enrichment setup, and keeping the source on each detection, gives you a fuller stack than either one alone.

Plays that start from a tech stack

Once you trust a detection, it turns into a reason to reach out. Most technographic targeting falls into one of four plays.

Displacing a competitor

Build a list of companies running the product you replace, and write to the specific weaknesses your buyers complain about. The timing matters as much as the list. A lot of B2B software runs on annual contracts, so a displacement message tends to land best in the months before a renewal, and a first-seen date gives you a rough guess at when that might be.

Selling into an integration

Companies already using tools that connect to yours have less to rebuild and fewer reasons to say no. If your product syncs with HubSpot, a HubSpot detection belongs in your lead scoring as a fit signal, and in our experience those deals tend to move faster than ones that need a new integration first.

Filling an obvious gap

Say you sell attribution software. Instead of mailing every mid-market e-commerce company, pull the ones with ad pixels from Google and Meta on their site and no attribution tool detected. The list gets much smaller, and every company on it has the same visible problem: money going into ads with no clear view of what it returns. That's a message you can write with confidence, because the evidence is sitting in their page source.

Catching a change

A new tool appearing, or an old one disappearing, often means a project is underway. A company that just added a data warehouse is probably rethinking reporting. One that removed a competitor's script may be between vendors. Those changes are a trigger event like a funding round or a new hire, and they sit alongside other buying signals in a signal-based selling setup. Pairing a stack change with intent signals, such as research on your category, tells you both that something changed and that someone is shopping.

Writing the email without sounding like surveillance

The fastest way to waste a good detection is to quote it. "I noticed you have the Hotjar script on your pricing page" tells the prospect you've been inspecting their site, and most people find that odd rather than impressive.

Use the stack to choose the problem, and then write about the problem. A team on a competitor's product gets an email about the thing that product does badly. A team running ads without attribution gets an email about not knowing which campaigns pay for themselves. Mentioning a public, widely known tool is fine when it's the point of the message ("we sync with Salesforce in both directions"). Naming something only a crawler would know is where it starts to feel like being watched.

Where technographic data belongs in the CRM

A provider can sell you hundreds of tools per company. Almost none of them should land in your CRM. Pick the handful that change a decision (your competitors, the tools you integrate with, the gaps your plays depend on) and store each as a property with its last-seen date. Everything else can stay in the enrichment tool, where you can pull it when a new play needs it.

Those few fields then do real work. They tighten your ideal customer profile from "B2B software, 50 to 500 people" to "B2B software, 50 to 500 people, running a CRM we integrate with". They filter outbound list building, feed fit scoring, and help routing send displacement deals to the rep who knows the competitor.

On outbound, enrich the account before it enters a sequence, not months ahead. On inbound, run inbound enrichment the moment a form comes in, so the rep sees the stack before the first call and two leads that look the same on paper can be ranked by which one has the gap you fill. Both are part of normal data enrichment, and both suffer from data decay like any other field, so re-check the detections that drive a campaign before the campaign goes out.

Remember what kind of data this is, too. Technographics nearly always come from a data provider, which puts them on the third-party side of the first-party vs third-party data line. When a prospect tells your rep on a call which CRM they use, that answer beats any detection. Give it its own field and don't let the next enrichment run overwrite it.

Wrapping up

Technographic data is one of the few ways to learn how a company works before you've spoken to anyone in it. Its value comes from reading it like a practitioner: know which route each detection came from, weigh it by how recently it was confirmed, and only trust a gap when the missing tool would normally be visible.

Used that way, a stack does two jobs at once. It tells you which accounts to work first, and it tells you what problem to open with, which is the part most cold outreach gets wrong. So stack data belongs in the targeting layer of your GTM system, and it hands you a specific opening line once prospecting runs automatically, instead of a template with a company name swapped in.

Our technographic builds for clients put three things in their own CRM and enrichment accounts: the detection sources, the short list of tools worth storing, and the refresh rules. Each field gets a note saying why it's there, and the client's team keeps adjusting it after the project ends.

Frequently asked questions