What is technographic data? It is the record of which tools a company runs: its CMS, analytics, CRM, ecommerce platform, email service and the rest of its stack. Most of it sits in plain view in the company's own website code, which makes it cheap to collect and easy to misread. A website shows some tools and hides others, and most lists treat the hidden ones as a no.
Technographics are one of the six common kinds of data enrichment, next to firmographic, contact and the first-party facts from your own billing and support tools. This post stays on the stack: what counts, where it comes from, and how to use it to pick and rank accounts.
If you would rather see a real stack than read about one, look up the tools behind your own accounts free.
What is technographic data?
Ask what is technographic data in a sales team and in a marketing class, and you get two answers. In B2B sales, it describes the technology a company uses: which products, made by whom, since when. The class will tell you the word began in consumer research, where it sorted people by how they adopt technology. Sales teams narrowed it to one thing, the stack.
A useful technographic record holds four facts about each tool:
The product, such as WordPress, HubSpot or Shopify.
Its category, such as CMS, marketing automation or ecommerce.
When it was first seen, and whether it is still there.
Where the fact came from: a website read, a job post or a note from a call.
Most lists drop the last one. It is the field that tells you how far to trust the other three, so keep it.
Firmographic data is a different question, drawn from firmographic data sources of its own. Industry, headcount, location and revenue describe the company's shape, and shape says whether an account could buy at all. The stack answers what happens after that: whether your product connects to something they have, replaces something they pay for, or would sit beside nothing.
What is technographic data used for? Examples
Here are the technographic facts that change what a seller does next. Each one is checkable on a real domain.
Technographic fact | Example | What a seller does with it |
|---|---|---|
Runs a tool you integrate with | Storefront on Shopify, HubSpot tracking code on every page | Target first: your product works on day one |
Runs a tool you replace | A competing help desk widget on the support page | Open on the switch, not on your feature list |
Recently adopted a tool | Marketing automation first seen two months ago | Offer setup help while the project is still open |
Dropped a tool | An analytics script that disappeared in the spring | Ask what replaced it, and why |
Email service | MX records pointing at Google Workspace or Microsoft 365 | Check compatibility before the demo |
No sign of a tool | A 200-person company with no chat widget on its site | Qualify by hand: missing from the site is not missing from the company |
Those facts end up in four jobs:
Targeting. Build a list of companies that run a tool you fit with, and leave out the ones running a tool that rules them out. Marketers call this technographic segmentation.
Scoring. Add the stack to the fit half of your score, beside industry and size.
Messaging. Open with the tool they run. "You run HubSpot" is a better first line than anything about your roadmap.
Timing. Adoption and removal dates are triggers. A tool added last month means a project is live, and someone owns it.
One opinion, and it does not sell anything. The count of technologies on a site is close to useless. A domain with 80 detected tools usually has a marketing team that installs every tag it reads about, and it is no more likely to buy from you than a domain with 12. Sort by the tools that matter to your product, never by the total.
Technographic data sources: where it comes from
Nearly all technographic data sources fall into three kinds. They see very different parts of a company, which is why a list built from only one of them has blind spots in the same place every time.
Website reads
Every page a browser loads pulls in scripts, stylesheets and tags, and many of them name the company that makes them in the file path or a global variable (the script element reference on MDN shows how a page loads them). A crawler reads the same code your browser does. DNS adds more: a domain's MX records name the service that handles its email.
This is the bulk of what gets sold as technographic data. It is repeatable and it carries dates. It also sees only what touches the public website, so the CRM, the data warehouse and the billing system behind the login usually stay invisible.
"Sales comes down to timing and to having the right data, and most of that data is already online. Years of building web scraping infrastructure showed me that."
Utku Zihnioglu, CEO of Oneprofile
Job posts
A posting for a Salesforce administrator says more about the back office than any crawl. Job descriptions name the CRM, the warehouse, the ERP and the support desk a team works in, because candidates need to know. The catch is that a post states a requirement or a plan, and posts linger after the role is filled. Hiring signals work best as a second opinion on a website read, or as the only evidence for a tool a website never loads.
Your own CRM and product
This is the kind nobody can sell you. Discovery call notes record what a prospect said they use. Closed-lost deals often name the tool that won. For accounts that already pay you, your own product knows which integrations they connected, which is the most accurate technographic data you will ever hold. Most teams never read it as technographics at all.
Source kind | What it sees | What it misses |
|---|---|---|
Website code | CMS, analytics, tag manager, chat, ecommerce platform | Anything behind the login |
DNS records | Email service, some hosting | Which apps run on top of them |
Job posts | CRM, warehouse, ERP, support desk | Whether the tool is still in use |
CRM notes and deal fields | What the prospect told you, the tool that beat you | Every account you never spoke to |
Your product's integrations | What your customers actually connected | Prospects, until they sign up |
Anyway, back to the blanks from the opening. Each kind of source fails quietly, and a missing answer reads the same as a no unless you record which sources were asked.
Using technographic data to target and score accounts
Say you sell a reporting add-on that reads HubSpot data, and your best customers are agencies of 10 to 50 people. HubSpot's tracking code loads on the public website, so this is the easy case: the fact you need is one a website read can see.
Build the list from the stack
In Oneprofile, open Find companies and use the Technology section. Type HubSpot into Technologies, and put a CRM that rules an account out into Exclude technologies. Under Advanced, Still using it since drops companies that ran HubSpot once and stopped, and First started using it since finds recent adopters.
The Technology section reaches companies whose website has been read, which is about one in six, and the preview says so once you add a technology. Pair it with industry and headcount rather than running it alone. Counting the matches is free, so shape the search before you spend.
Fill in the stack for accounts you already have
For a list you imported from your CRM or a CSV, add Find technology stack. It reads a domain and returns the CMS, analytics, CDN, ecommerce platform and JavaScript libraries it finds, with the categories they fall into. It costs up to 5.8 credits a company, and a lookup that finds nothing costs nothing. Its data sources are asked in an order set by accuracy, not price, the same kind of waterfall the main company lookups use.
For back-office tools, use Find companies hiring instead. Type the tool's name into Keywords to match the text of the posting, not just its title, and put your accounts' domains under Employer to keep the search to them.
Here is what the search and the lookups might show for five invented agencies:
Account | What the lookups found | How to read it |
|---|---|---|
juniperstudio.example | WordPress, HubSpot tracking code, 31 technologies | Strong fit: the add-on works on day one |
dovetaildesign.example | HubSpot, matched by a First started using it since date two months back | Strong fit, and the timing is now: open on setup |
sundaycreative.example | WordPress and an analytics tag, no CRM script | Unknown CRM: check job posts and your notes |
granitecreative.example | A competing CRM's forms on the contact page | Poor fit until they switch |
cedarcreative.example | Nothing found, not charged | No answer yet, which is not a poor fit |
The last two look alike in a spreadsheet and mean opposite things. granitecreative.example was checked and does not fit, while nobody has an answer for cedarcreative.example yet. A rep who cuts both throws away an account nobody looked at.
Score technographic data against your ICP
Write the stack rule into your ideal customer profile as a plain bullet: "Runs HubSpot. A competing CRM is a poor fit until they switch." In Oneprofile it goes under Ideal customer profile in Business context, one fact per bullet.
The built-in ICP scorer reads the company domain, name, description, industry, employees, company type, who they sell to and job title against that profile. To make it weigh the stack too, open the template gallery with See all templates, select the ICP scorer, click Apply template, then Save. In your copy, click Add input, name it Technologies, and map it to the Technologies that Find technology stack returned when you add the agent to the table.
Each account comes back with a band (strong, possible or poor), a score from 0 to 100, and the facts that drove it, each naming the field it came from. An account with too little evidence gets no band and no score, only a list of what was missing, because poor is kept for accounts that were checked and do not fit. It is the same line that separates granitecreative.example from cedarcreative.example in the table above. More on building the profile itself is in how to score accounts against your ICP.
The Free plan renews 200 credits a month for 1 seat, with up to 3 integrations, 3 tables and 100 rows per table. That covers at least 34 stack lookups, enough to test your stack rule on accounts you have already won or lost. Look up the stack on 30 of your own accounts free, then sort by the tool that matters to you and read the top ten.
