TL;DR
- BuiltWith is a candidate for website technology discovery and historical web detections; check the list and API entitlement you actually need.
- Wappalyzer is useful for technology lookups and web-based list building. Its URL observation is not proof of every internal system at the company.
- HG Insights is worth evaluating for a broader enterprise technology footprint and organisational context; request a sample and quote for your specific products and accounts.
- PredictLeads offers technology detections through API and other delivery methods, with source and timing context. A public mention remains evidence to assess, not proof of a purchase.
- Apollo combines technology filters with company and contact prospecting; confirm filter entitlement, supported technologies and export rights.
- Treat “not detected” as unknown unless your evidence supports a stronger conclusion. Test 5,000 target accounts for useful, recent evidence rather than buying the largest count.
What does a technographic signal actually mean?
A technographic record describes a technology associated with a company. The evidence might be a script on a public website, a DNS record, a job advert mentioning a tool, a vendor case study or another source. Those observations answer different questions. A browser tag can show that a page exposed a product signature; it cannot by itself prove the company's total spend, contract owner or deployment across every subsidiary.
For prospecting, specify the decision first. A website analytics company may need to identify domains with a competing tag. An enterprise software seller may need evidence that an Indian subsidiary uses a particular ERP. A team tracking displacement may need changes over time, with enough evidence to distinguish a real removal from a failed crawl.
The providers below offer different ways to approach that problem. Their descriptions are based on published materials, not a comparative benchmark. The BuiltWith vs Wappalyzer comparison goes deeper on two website detectors; this guide covers the wider procurement decision.
Technographic providers at a glance
| Provider | Starting fit | Evidence question | Commercial question |
|---|---|---|---|
| BuiltWith | Website technology lists and history | Which domain and observation date support the detection? | List plan, upload allowance and API rights |
| Wappalyzer | URL lookups and technology-based lists | Was the technology detected on the exact URL recently? | Lookup credits, list scope and plan |
| HG Insights | Enterprise technology and organisational context | Which source, product and location support the install record? | Contract, delivery and minimum commitment |
| PredictLeads | API-delivered detections and change context | Which public evidence and timestamps underlie the signal? | Request credits, minimums and delivery format |
| Apollo | Technology-filtered company and contact prospecting | Which technologies, source context and observation dates are available? | Plan entitlement, contact credits, exports and rights |
A supplier can be strong for public web tags and weak for private back-office software, or vice versa. Ask for evidence at the level where you will make a sales decision.
| Buying question | Evidence to request | What that evidence cannot establish alone |
|---|---|---|
| Does a specific site expose a web technology? | Exact URL, detected signature, fetch time and response status | That every company site or subsidiary uses it |
| Is enterprise software associated with an operating entity? | Product, entity/location, source type and last-supported date | Current contract scope, spend or renewal date |
| Has visible use changed? | Comparable observations on the same source and detection method | A purchase, cancellation or displacement without corroboration |
| Can a seller act on the result? | Correct CRM account, relevant workflow and evidence a rep can inspect | A verified buyer, budget or permission to contact |
Use this matrix to separate a website-detector purchase from enterprise-install research. A supplier may cover both, but the evidence and contractual fields still need to be checked for the actual product and account.
1. BuiltWith: for website technology lists
BuiltWith provides website technology information and lists. It is a useful candidate when a team wants to find domains associated with a web-visible product, segment those domains and inspect historical information.
Its plan explanation lists Basic at $295 per month and Pro at $495 per month, with different technology-list and domain-upload allowances. Confirm the live checkout and whether the exact API or list operation is included; a report subscription and an API credit package are not necessarily interchangeable.
BuiltWith fits web-technology prospecting where the observed domain is central. It is less conclusive for software that never appears on a public site. Ask for first and last observation dates, how subdomains are associated with company records and how historical or temporarily absent detections are labelled. A tag disappearing after a redesign is a reason to investigate, not automatic proof the customer cancelled a contract.
2. Wappalyzer: for website lookups
Wappalyzer identifies technologies associated with websites and offers lookups and lists. Its lookup API documentation requires a Business plan and lists one credit per URL for the standard lookup. A request using both live and recursive options costs five credits per URL and runs asynchronously; the documentation requires a callback for that combination. Historical results may include technology no longer in use.
The default lookup can include observations from the current and two previous calendar months, and Wappalyzer says it can begin a live crawl when no cached record exists. An initial response without technologies may therefore be a pending crawl rather than an empty technology stack. Preserve the initial and callback responses, authenticate callbacks as documented, and distinguish pending, failed and completed-with-no-detection states. Set the age and live options deliberately when comparing observations over time.
Its pricing page shows a free allowance and paid plans with different list and lookup limits. Confirm the current monthly or annual selection, whether the plan supports the desired combination of technologies, and how many URLs a single account requires. Five thousand companies with several domains each can require more than 5,000 lookups.
Wappalyzer is a natural candidate for an operational workflow that starts with known websites. It is a weaker sole source for private ERP or CRM installations without public signatures. Test the precise domain and time of observation before writing a result to the CRM as a current account fact.
3. HG Insights: for enterprise technology context
HG Insights describes technology-install data derived from multiple public signals, including company-produced documents as well as digital evidence. Its documentation describes product, location and verification context. Those fields can help a team evaluate enterprise software that may not be visible in a website tag.
Ask HG to explain the evidence and confidence for the exact product and subsidiary in your sample. A job posting may describe an existing installation, a planned migration or a desired skill. A historical verification date is not a contract end date. Request data lineage, refresh cadence, delivery format and the commercial terms for the fields you need.
HG is a candidate when organisational context and enterprise software matter enough to justify a broader data purchase. It may be excessive for a small website-only prospect list. Compare a bounded sample and contract against a narrower web detector rather than assuming the two products measure the same thing.
4. PredictLeads: for evidence-led detections
PredictLeads describes technology detections associated with companies, delivery through an API and other formats, and technology metadata. Its own explanation of the dataset distinguishes a detection from a confirmed internal installation and describes first- and last-seen timestamps.
That distinction is useful for a workflow that needs to prioritise research rather than assert certainty. A public job post mentioning a data warehouse can be a relevant signal, but it may refer to a future project or an unrelated client. Store the evidence type and date alongside the account record so a rep can decide whether to investigate.
The pricing page explains credits and request-based pricing. Confirm the minimum monthly charge, free allowance, pagination and how multiple datasets or refreshes consume credits. PredictLeads can suit an API-first team combining technology with other signals; it still requires a matching, validation and activation process in your own stack.
For the Technology Detections endpoint, PredictLeads describes one credit per request with up to 1,000 results and a separate credit for each paginated request. Its discovery endpoints use a per-result model instead. Do not price 5,000 company lookups as five requests merely because one request can return many detections for a company; confirm the endpoints, pages and selected tier for the actual job.
5. Apollo: for technology-filtered prospecting
Apollo's search documentation describes a Technologies filter for finding companies or people associated with companies using selected products. It says the dataset draws from multiple sources and is updated regularly. A team can combine that filter with its other prospecting criteria instead of buying a separate technology list and then finding contacts elsewhere.
That convenience does not prove that a returned account currently runs the product across every business unit. Ask which technology names are supported, whether the relevant plan includes the filter, what evidence or observation dates are visible, and how export or enrichment rights apply. Separate any contact-data credits and outreach costs from the account-search decision; do not assume every returned person is a verified buyer.
Apollo is a practical candidate when the team wants technology context inside a broader prospecting workflow. A buyer needing auditable, installation-level evidence or an event feed should test that requirement separately against a specialist source. Compare the same difficult accounts and retain an unknown state where the evidence is insufficient.
What does a 5,000-account programme cost?
Assume 5,000 target accounts, one mapped primary domain per account and a monthly refresh. In practice, multiple domains, redirects, subsidiaries, failed lookups and separate technology queries can increase the count. A provider that charges per URL and one that charges per company or export are not using the same unit.
For a synthetic model, suppose a data plan costs $300 per month and an operator spends five hours at $60 per hour on matching, review and exceptions. The monthly total is $600. If the team obtains sufficiently recent, relevant evidence for 2,000 of the 5,000 accounts, the cost is $0.30 per evidence-backed account. This is a scenario, not a price or observed coverage result for any supplier.
If 500 accounts require a second domain check at an assumed $0.05 each, add $25, making the total $625, or $0.3125 per evidence-backed account if the accepted count remains 2,000. If that second check finds 100 additional valid accounts, the denominator becomes 2,100 and the cost is about $0.298 per account. Both outcomes are assumptions; measure the extra yield before buying the second pass.
For a public-plan reference, BuiltWith's documented Basic list plan is $295 per month and includes 2,000 monthly domain uploads, so this example's 5,000-domain refresh would need another allowance or arrangement. Under Wappalyzer's documented rates, 5,000 standard API lookups require 5,000 credits, while 5,000 live recursive lookups require 25,000; check Business-plan entitlement, available credits and billing before converting those units to dollars. PredictLeads uses its own API response units. Do not divide a subscription by 5,000 until you know whether the selected plan supports all 5,000 required operations and the required technology filters. Include minimum commitments, API overages, storage, monitoring and the labour to match company hierarchies.
How do you test the evidence before activation?
Create a fixed sample of public websites from the actual target list: for example, 100 accounts across your major sectors, company sizes and regions. Record each company, expected domain, subsidiary relationship and the technologies the sales team wants to identify. Include a separate category for internal enterprise software whose presence cannot be established from the public site. This is a proposed test, not research we have performed.
For each returned signal, record the technology, exact domain or document, source type, first and last observation where available, and whether the evidence supports current use, historical use or only a mention. Have a reviewer classify supported, contradicted or unknown using permitted sources. Do not mark an unobserved back-office product absent simply because the site has no tag.
For a real, source-linked example, WordPress's Capgemini showcase describes a move from Drupal to WordPress for Capgemini's web publishing and identifies capgemini.com; the entry is dated August 2018. Record this as a historical, WordPress-published account of a web-platform migration, with the entry date and the date your team retrieved it. It is a useful lead for a current check, not proof that every Capgemini website still runs WordPress, that a particular supplier detected it, or that the company is currently buying a replacement. A back-office CRM remains unknown on this evidence.
For procurement, ask each supplier to return the same ten difficult accounts before a full paid pilot: two with several regional domains, two whose parent and subsidiary have different brands, two with public technology signatures, two with only job-posting evidence and two for which no public evidence is expected. Keep the last group as unknown. Record whether the supplier can explain a result or only supplies a label. This small exercise does not establish market-wide accuracy, but it can reveal a mismatch between the product and your intended decision.
Then test change detection. If a tag is present in one observation and missing in the next, check whether the page was accessible, the same subdomain was examined and the detector's coverage changed. Seek a second observation or a different source before triggering a displacement campaign. A “new detection” can reflect a newly indexed page rather than a new purchase.
Keep raw evidence separate from the CRM's sales action. For example: “A public job advert mentioned Product A last month” is a traceable research cue; “the company just bought Product A” is a stronger claim that needs additional evidence. Record the source and review date, expire stale signals and allow a rep to correct a mistaken entity match.
Consider a fictional manufacturer, Northstar India, whose public careers page mentions Product A in a desired-skills list. A provider matches that page to the global parent and reports Product A as detected. The useful CRM observation is the specific page, its date, the matched entity and a job-posting mention. It does not establish that the Indian subsidiary bought the software, that a particular deployment is active or that its contract is expiring. Route the record to research unless another permitted source supports the decision your sales team intends to make.
For a later refresh, retain the earlier observation rather than replacing it with an unexplained blank. Record whether the new page was retrieved successfully and whether the same source still supports the claim. A failed fetch, an accessible page without the mention and a confirmed end of use are three different states. This is a proposed evidence workflow, not an observation about any named supplier.
The waterfall enrichment guide can help decide when a second source is worth the extra cost. For continuous monitoring rather than initial data procurement, see our competitor monitoring tools guide.
Which provider should you shortlist?
Start with BuiltWith or Wappalyzer for a website-led question, HG Insights for a broader enterprise technology investigation, PredictLeads for evidence delivered into an API workflow, and Apollo when technology filters need to sit inside a broader prospecting process. Request the same account sample and evidence fields from each.
Technographics are one layer of account research. Our B2B company data provider comparison covers the firmographic and company-matching work needed to turn a domain observation into a usable account record.
FAQ
Does a missing technology detection mean the company stopped using it?
No. The page may not expose the technology, the crawl may have failed, or a different domain may be in use. Treat it as unknown until additional evidence supports a change.
Can website detectors find every enterprise application?
No. Internal CRM, ERP and data systems may not leave a visible public signature. Other sources can provide useful evidence, but a job advert or public mention still needs context and does not prove a current contract.
What should a technographic record contain?
At minimum, retain the matched company and domain, product, evidence source, observation date and any relevant confidence or status. Separate a detected signal from the sales team's interpretation and record corrections.





