--- title: How we source tech stack data description: Discover how TheirStack extracts tech stack data from millions of job postings worldwide using confidence scoring to identify company technologies. url: https://theirstack.com/en/docs/data/technographic/how-we-source-tech-stack-data --- ## Inferring tech stack data from job postings We source our tech usage data from job postings. We gather job listings from millions of companies monthly by tracking company websites, job boards, and other hiring platforms. When a job mentions a specific technology it hints the company is using that technology. The mention of the technology could be in the: - **Title**: e.g. "Senior React Developer", "Java Developer", "Odoo Developer"... - **Description**: e.g. "It requires experience with Hubspot", "We use GraphQL to query our data", "You have at least 3 years of experience with React"... - **Job URL**: Jobs posted in a ATS have the domain of the ATS in the url. Eg: [https://join.com/companies/xxx/xxx](https://join.com/companies/xxx/xxx), [https://jobs.workable.com/view/xxx](https://jobs.workable.com/view/xxx)... ## Measuring the likelihood of technology usage Our **Confidence Score** (low, medium, high) helps us estimate how likely it is that a company actually uses this technology. It's calculated based on - the number of mentions - the frequency of the mentions relative to similar technologies in the category. If a company mentions 40 times Hubspot and 4 times Pipedrive, we consider it more likely they use Hubspot. - where this mention is found. We give more weight to mentions in the title or in the url. - the date of the mentions. We give more weight to the most recent mentions. ## Tracing technology signals back to evidence We're obsessed with transparency and giving you access to the evidence that led us to identify a company's technology usage. That's why everywhere we show a technology signal, we display all the information we have about that signal, allowing you to trace it back to its original source. ### In the app When you're in a [company search](/en/docs/app/company-search) filtering by technology or in the [technology tab for a company](/en/docs/app/company-lookup), we show cards highlighting the technologies they use. These aren't just simple tags – each card contains a complete audit trail of how we detected that technology usage. If you hover on the technology card, you can see some details on the relationship between that company and that technology: - **Category** this technology belongs to - **Confidence score (high, medium, low)** - see the full explanation [here](#measuring-the-likelihood-of-technology-usage) - **First date found** - when a job by the company mentioned the technology for the first time - **Last date found** - when a job by the company mentioned the technology for the last time - **Rel. occurrence** - number of mentions of this technology compared to the total mentions of all the technologies from the same category. See the full list of categories here and browser the technologies we track here. - **N. jobs mentioned** - number of jobs mentioning this technology, with link to those jobs. By clicking on `N.jobs mentioned`, you'll be taken to a [Job Search](/en/docs/app/job-search) view showing the jobs from that company where the technology was found. ### Through the API Through our [Technographics endpoint](/en/docs/api-reference/companies/technographics_v1) and [Company Search endpoint](/en/docs/api-reference/companies/search_companies_v1), you'll get access to the same detailed technology information that's available in our app. ## FAQs #### Why are there fewer companies listed for a technology than I expected? Our data comes from job postings. If a company doesn't mention a specific technology in their job ads, doesn't have open positions, or the technology isn't relevant to their hiring needs, it won't appear in our results. For some roles or categories, the actual technology used by the person hired may not be specified in the job description, or it might not be relevant. As a result, even if many companies use a certain technology, they may not show up in our database unless it's explicitly mentioned in their job postings. If you're seeing fewer companies than expected, it's likely because of how companies describe their needs in job ads, not necessarily a reflection of the true adoption rate of that technology. In any case, and because you can make unlimited searches without consuming [credits](/en/docs/pricing/credits) unless you reveal data, you can always see the number of companies that would be returned by any given search. #### Can I trust low/medium confidence technologies? The short answer is yes — low and medium confidence technologies still come from clear evidence. Some companies don't post many jobs, so our confidence that they use a given technology will naturally be lower than for companies with lots of signals. If you're doing outbound and prefer to minimize false positives, and you don't target startups or smaller companies with fewer jobs, stick to **medium and high** confidence signals. Otherwise, we recommend using **all** confidence levels.