Where do tech stack signals actually come from?
The phrase "tech stack data" suggests a neat list of installed software, but what providers actually capture is a set of observable signals that imply software use. Four distinct signal types exist, and they differ substantially in reliability, freshness, and coverage.
DNS records are the most trustworthy source. A company's MX records reveal their email provider (Google Workspace, Microsoft 365, custom mail servers). SPF records name the sending services authorised to send on behalf of the domain. CNAME records point to third-party platforms including marketing automation systems, customer portals, and CDN providers. DNS queries are cheap, near-real-time, and require no access to the company's internal environment. If a company moves from one email provider to another, the DNS records update within 24 to 48 hours of the change propagating.
Page-source markers cover the client-side technology visible in a website's HTML, CSS, and JavaScript. JavaScript tags identify Google Analytics, Hotjar, HubSpot, Intercom, Drift, Segment, and dozens of other tools. This signal class is reliable for anything rendered in the browser, but it is entirely blind to server-side tooling. A company running an on-premise Dynamics 365 instance with no web-facing integration will leave no page-source trace whatsoever.
Job advertisement requirements are the most widely used signal class and the one most likely to mislead you. A job listing requiring five years of Salesforce experience tells you the company wants someone who knows Salesforce. It does not confirm Salesforce is currently deployed, that the deployment is active, or that the company is not mid-migration to something else. Hiring signals can lag reality by six to twelve months in either direction. They are best treated as supporting evidence alongside a stronger signal, not as primary confirmation.
Public filings, press releases, and case studies provide the highest-confidence evidence when a vendor publishes a named customer case study or a company references a specific platform in an annual report or investor presentation. These mentions are sparse and irregular, but when they exist they are usually accurate at time of publication. Freshness is the problem: a case study published in 2023 tells you the company used Salesforce in 2023.
Which technology categories are well covered and which are not?
Coverage is not uniform across the technology landscape. The table below summarises what you can reasonably expect from commercially available tech stack data.
| Technology category | Primary signal source | Data quality | Typical freshness |
|---|---|---|---|
| Email service provider (GSuite, M365) | DNS (MX / SPF records) | High | Near real-time (days) |
| Web analytics (Google Analytics, Adobe) | Page-source JavaScript tags | High for client-side installs | Weeks |
| Marketing automation (HubSpot, Marketo, Pardot) | Page-source tags, DNS CNAME | Medium-high | Weeks to months |
| CRM (Salesforce, Dynamics 365, Pipedrive) | Page-source, job ads, case studies | Medium | Quarterly |
| E-commerce platform (Shopify, Magento, WooCommerce) | Page-source markers | High | Weeks |
| Cloud infrastructure (AWS, Azure, GCP) | DNS, page-source, job ads | Medium | Quarterly |
| Payment processing (Stripe, Braintree, Worldpay) | Page-source JavaScript | Medium-high (client-side only) | Weeks to months |
| On-premise ERP (SAP, Oracle, Sage on-prem) | Job ads, filings | Low | 6+ months |
| Legacy telephony / PBX systems | Job ads only | Very low | 12+ months |
| Internal BI / data warehouse tools | Job ads, filings | Low | 6+ months |
The practical takeaway: if your product integrates with or replaces something in the top half of that table, tech stack data is a credible qualifier. If your product targets the bottom half, treat any signal you receive with scepticism until you can verify it directly with the prospect.
Where does tech stack data genuinely earn its place?
There are three prospecting scenarios where tech stack data adds clear value. Outside of these, it tends to add noise rather than signal.
Displacement prospecting
If your product replaces a specific competitor's platform, knowing which companies run that competitor is the most direct targeting filter you can apply. A company identified as running a particular CRM is a legitimate prospect for a competing CRM vendor, and the opening of the sales conversation is immediately more specific: "We work with a number of teams that have migrated from [category platform] to ours, and I wanted to share what that transition typically looks like." That specificity is only possible because you know the tech.
The caveat is that displacement signals can be stale. A company flagged as running an older accounting platform may have already migrated by the time your sales team calls. Build that assumption into your outreach; acknowledge it is possible they have already moved on, rather than stating their current stack as a certainty.
Integration-led prospecting
If your product integrates with HubSpot, Salesforce, or Shopify, companies already running those platforms are pre-qualified in a meaningful way: they have the infrastructure your integration requires, and the conversation starts from "this will work in your environment" rather than "first you would need to buy X, Y, and Z." This is probably the cleanest use case for tech stack data because the signal (platform usage) maps directly to a product benefit (the integration), not just a sales assumption.
Security and compliance gap targeting
A company's visible technology profile can reveal gaps. A firm running a large web presence on an older CMS without an obvious web application firewall or CDN may be a prospect for security tooling. One whose job ads consistently reference legacy on-premise infrastructure alongside cloud migration requirements may be mid-transformation and open to conversations about modernisation tooling. These are probabilistic inferences, not certainties, but they can sharpen the prioritisation of an outbound list.
The accuracy and freshness problem
The single most common complaint about tech stack data is that it is wrong. Not occasionally, but at a rate that matters at scale. There are three distinct failure modes.
False positives occur when a signal appears to confirm platform usage but does not. A residual JavaScript tag from a previous vendor left in a site's source code is a classic example; the old tag fires, the provider records a positive hit, and the company appears to still be running that platform two years after they cancelled the contract. Page-source signals are particularly prone to this. DNS records are cleaner because a decommissioned service will eventually stop appearing in DNS, but even those can lag by weeks.
Departmental tools are routinely missed. Large UK companies often run different CRM instances for different business units. The headquarters website may signal Salesforce while the Northern European division runs Dynamics 365 and the legacy B2C business still operates on a bespoke system. Tech stack data aggregates to company level and flattens this complexity. The department you want to sell into may be running entirely different tooling from what the company-level record suggests.
Stale records affect any platform category that refreshes quarterly rather than continuously. Cloud-native SaaS tools are easier to track because deployments leave persistent web signals, but on-premise software changes are nearly invisible between refresh cycles. For a company that completed an ERP migration in February, a quarterly-refresh file queried in March may still show the old platform through June.
In our experience, the most reliable posture is to treat tech stack data as a probabilistic filter, not a confirmed fact. Use it to prioritise and segment, not to make definitive claims in the opening line of an outreach message.
How to combine tech stack data with firmographics
Tech stack data is an overlay, not a foundation. The correct sequencing is to build a qualified firmographic universe first, then apply tech stack filters to rank or segment within it.
Start with industry targeting by UK SIC 2007 code, adding employee headcount, turnover band, and geography. That base universe reflects companies that are the right size, in the right sector, and within your serviceable UK geography. Apply tech stack signals on top: "of the 4,200 companies in this universe, 900 show a HubSpot deployment signal and 1,100 show a Salesforce signal." That ordering gives you a denominator to work from.
What you should not do is start with a tech stack filter and then assume everything in the result set is a qualified prospect. "All UK companies running Salesforce" is not a target market; it is a list of 30,000-plus companies with no qualification on sector, size, or relevance to your product. The tech stack signal has done no firmographic work for you.
For account-based approaches where you are targeting specific job functions within technology-qualified accounts, the combination becomes genuinely powerful: firms in financial services (SIC 64-66) with 200 to 2,000 employees, running a specific cloud infrastructure platform, where you want to reach the Head of IT or CTO directly. That is a tight, qualified universe where the tech stack signal does real work alongside the firmographic criteria.
UK GDPR note on outreach to tech-qualified accounts
Tech stack data itself is observational and does not constitute personal data under UK GDPR. However, the moment you use it to identify companies and then contact named individuals at those companies, you are processing personal data. B2B prospecting to named corporate contacts requires a lawful basis under Article 6(1)(f) UK GDPR (legitimate interests), and you must complete a Legitimate Interests Assessment before commencing outreach. This requirement exists regardless of how you identified the accounts.
What tech stack data cannot tell you
The most important limitation is the one that is hardest to work around: tech stack data tells you the technology, not the intent.
A company running a legacy CRM is not necessarily unhappy with it. A company running your competitor's product is not necessarily evaluating alternatives. Knowing the tool does not tell you whether there is a budget cycle open, whether there is an internal champion looking for change, or whether the IT Director you want to reach has any authority over that particular purchasing decision. Tech stack data answers the question "what are they using?" It cannot answer "are they ready to buy something else?"
For that reason, tech stack data used alone as a cold outreach filter typically produces lower conversion rates than the same list further qualified by intent signals (topic-based research activity from B2B content networks) or by a warm inbound signal (the company has visited your pricing page). The tech stack tells you who the relevant companies are. Intent signals tell you which of those companies are active right now. The combination is more valuable than either alone.
There is also a coverage floor problem for smaller UK businesses. A fifty-person professional services firm may have a website, a business email address, and a LinkedIn page, but if their tooling is entirely cloud-based SaaS with no public integrations and their job ads are infrequent, the tech stack coverage may be sparse or absent entirely. Account-based approaches targeting smaller UK businesses should account for the fact that tech stack data coverage thins out significantly below around 100 employees for most technology categories.
A realistic view of what you are buying
Tech stack data from any provider is a probabilistic inference, not a confirmed audit. The better providers are transparent about this and express their outputs as confidence scores or signal counts rather than binary "yes/no" deployment flags. If a provider presents tech stack data as definitively confirmed rather than signal-derived, that is a red flag about their methodology.
For UK B2B targeting, the practical value of tech stack data is real but bounded. It earns its cost when used as one layer in a multi-signal qualification model: firmographics establish the universe, seniority and job function targeting (see our guide to job function and seniority data) identifies the right contacts within it, and tech stack signals rank accounts by fit. Used that way, it reduces wasted outreach calls and makes the first conversation more specific. Used as a primary targeting tool with no firmographic foundation, the precision is illusory.