Guide / Buying Signals

Tech stack signals: targeting companies by what they run

If hiring tells you what a company is about to spend on, the tech stack tells you what it has already committed to. Technographic data is one of the most useful filters in B2B outbound and one of the easiest to get wrong, because the data goes stale fast and the temptation to over-target is constant. In this guide I will walk through what stack signals actually tell you, where the data comes from, how we use it for displacement and complement plays, and the failure mode that quietly kills most technographic motions.

What technographics actually tell you

A company runs the tools it can afford, configured by the people it has hired, integrated with the rest of its stack. So when you know what a target uses, you can usually infer four things with reasonable confidence: budget level, technical maturity, what integrates cleanly with your product, and where the displacement opportunities sit.

• Budget. A company running Snowflake plus Looker is in a different price bracket than one running Postgres and Metabase. Neither is wrong, but they will react to different price points and different sales motions.
• Maturity. Tools betray org age. A modern observability stack on a three-year-old startup says the engineering team has scaled. A wall of legacy Java on a fifteen-year-old company says you are selling into a different conversation.
• Integration fit. If your product depends on a particular CRM, warehouse or identity provider, technographic data tells you whether the integration story is short or long.
• Displacement opportunity. If they run a competitor, you have a specific wedge. If they run an adjacent tool that complements yours, you have a different wedge.

Where the data comes from

• BuiltWith and Wappalyzer. The two best-known providers, both based on what runs in the public-facing website (analytics, tag managers, frontend frameworks, payment processors, marketing tools). Strong for anything that touches the website. Useless for tools that live behind the login.
• Job posts as stack leaks. Engineering job descriptions list the tools the team uses. A senior backend role at a target account tells you the language, the database, the cloud, the CI provider and often the observability stack in one paragraph. This is some of the highest-quality technographic data on the internet and it is free.
• Public repositories. Companies that work in the open through GitHub or similar leak their stack constantly through manifests, configs and READMEs. Useful for dev-tools sellers, narrower for everyone else.
• Case studies on vendor sites. Most B2B vendors publish a customer list. Scraping these gives you a high-confidence list of accounts running the named tool. The displacement plays often start here.
• Specialist databases. Several vendors aggregate technographic data across categories, including tools that never touch a public website. Coverage varies wildly by category and by region. Always trial on a sample of accounts you can verify by hand before paying.

In our enrichment surface (usegrit.io, or Clay), we typically combine two sources for any stack signal that drives a workflow: one breadth source for coverage and one verification source for accuracy. A single source is almost always either too narrow or too noisy on its own.

Displacement plays

A displacement play targets companies running a competitor. It is the highest-intent technographic motion you can run, because you already know the buyer has a budget for the category, has been through the buying decision once, and is somewhere on a curve between honeymoon and quietly disappointed.

A displacement play works when you have a specific, defensible wedge. "We are better" is not a wedge. "We are faster on the workflow that consumes 60 percent of your power users time" is a wedge. "We do the thing the incumbent stopped investing in two years ago" is a wedge. The message has to sound like you understand exactly what is frustrating about the incumbent on a Tuesday afternoon, not like you read their G2 reviews.

A common mistake on displacement is timing. New customers of the incumbent are still in honeymoon. Long-term customers have either stopped caring or have made the incumbent unremovable by integrating around it. The sweet spot is usually customers between 12 and 36 months in, who have hit the limits the incumbent does not solve, and have not yet rebuilt their world around it.

Complement plays

A complement play targets users of an adjacent tool. If you sell a reverse ETL product, users of a particular warehouse are your TAM. If you sell a billing product, users of a particular CRM are an easier conversation than users of one that does not integrate with you.

Complement plays are lower intent than displacement (the buyer has not declared a need in your category) but higher conversion than cold ICP outbound (you know the wiring works). The message should lead with the integration, not the product. "You already run X. Here is what most teams that run X struggle with next, and here is the thing we built for it" is the pattern.

Data freshness: the main failure mode

The single biggest failure mode of technographic motions is stale data. A tool detected on a website last quarter may have been ripped out two months ago. A competitor relationship that looked confirmed in a year-old database may have ended. The buyer who reads "I saw you use X" when they switched to Y in February will mark your email as spam and move on.

Two cheap habits help. First, date-stamp every technographic record and expire anything older than 90 days for the signals you actually fire on. Second, build a verification step into the workflow itself. Before a message sends, an agent does a five-second check (a recent job post, a recent integration listing, a recent case study mention) to confirm the stack is still real. Skipping this step is how you end up apologising for a mistake your CRM made on autopilot.

Layering technographic filters on top of firmographics

Technographic data on its own produces a worse list than ICP firmographics on their own. The combination is where the value sits. The pattern we use in usegrit.io is to start from a firmographic TAM (industry, size, geography, growth), then layer the stack filter as a probability boost, not a hard gate. A target-fit company without the stack signal stays in the workflow on a slower nurture cadence. A target-fit company with the stack signal moves into the high-intent workflow this week.

Treating technographics as a hard gate shrinks your TAM aggressively, often more than the data quality justifies. Treating it as a probability layer lets you prioritise without overfitting to data that might be three months out of date.

Common mistakes

• Over-trusting a single source. One provider is one opinion. Verify on a sample by hand before scaling.
• Ignoring stack freshness. Date-stamp records, expire them, verify before sending.
• Leading the message with the detection. "I see you use X" sounds like surveillance. Lead with the implication, mention the tool as a passing reference.
• Confusing presence with depth. A pixel on the website does not mean the company has rolled out the tool seriously. Combine technographic data with hiring or social signals to gauge real adoption.

Where this fits in the broader signal stack

Tech stack signals work best paired with motion signals like hiring or funding, which tell you something is changing this quarter, and with job changes, which tell you a new decision maker may be reviewing the stack she inherited. For the wider framework, see how to build a GTM system and what is GTM engineering.