# Intent data providers compared > Bombora, G2 Buyer Intent, 6sense, Demandbase and site de-anonymisation tools compared on signal source, accuracy, price and what each is genuinely good for. Source: https://saas-marketing.net/guides/intent-data-providers-for-saas/ Topic: SaaS Demand Generation Type: listicle Published: 2026-09-11 Last updated: 2026-09-11 Publisher: SaaS Marketing (saas-marketing.net) License: CC BY 4.0. Quote or republish with attribution and a link to https://saas-marketing.net/guides/intent-data-providers-for-saas/ ## Short answer Intent data providers for B2B SaaS split into four groups by signal source. Publisher co-ops such as Bombora report topic surges across thousands of member sites. Review platforms such as G2 Buyer Intent report category and comparison behaviour. Aggregators such as 6sense and Demandbase blend co-op, bidstream and first party signals with orchestration. De-anonymisation tools resolve your own traffic to companies. Review site signals are the most precise and the narrowest. Buy any of them only if sales can act within 48 hours. ## Key takeaways - Signal source predicts precision better than any feature list, so compare collection methods before you compare dashboards - Review site intent usually beats co-op topic surges for mid market SaaS because the buyer named your category and often your competitor - Intent data pays back only when an account gets a human touch inside 48 hours of the signal firing - Expect 8 to 15 percent of a raw co-op surge feed to survive ICP filtering, and budget SDR time against the filtered number - Platform contracts from 6sense and Demandbase run $55,000 to $130,000 a year, roughly triple a standalone feed - Validate any provider by replaying 60 days of signals against accounts you already closed, not against the vendor's sample list --- Every intent vendor demos the same screen. A list of accounts, a surge score, a green arrow next to a topic you sell into. The screen is real. What the demo leaves out is that the score describes a company rather than a person, that it is a probability rather than a fact, and that its value decays to almost nothing if nobody contacts the account within a couple of days. ## What you are actually buying when you buy intent data You are buying a prioritisation input, not a lead source. Every provider here observes behaviour somewhere on the internet, attributes it to a company, compares it against that company's own baseline, and flags the difference. [Intent data](/glossary/intent-data/) is a ranking adjustment applied to accounts you already care about. That distinction decides the budget question. A $60,000 contract does not create demand and it does not create accounts. It re-sorts a list. If you have 400 named accounts and three SDRs who can work 60 a week properly, re-sorting is worth real money. If you have no account list, you have bought an expensive spreadsheet with a nice colour scheme. Three things intent can honestly tell you: which named accounts are researching this month, which topic to open with, and which quiet account has re-entered a cycle. Everything past that is the vendor's own marketing, and you should price it at zero. ## Where the signal comes from, and why that decides everything Collection method predicts precision better than any feature comparison. Five methods dominate the market and they behave nothing alike. | Signal type | How it is collected | What it tells you | Main weakness | |---|---|---|---| | Publisher co-op | Member sites share anonymised content consumption, resolved to company by IP and registration data | Topic interest rose above this company's baseline | Topic level only, weekly cadence, no department | | Review site behaviour | Logged and anonymous behaviour on G2, TrustRadius or Gartner Digital Markets | Someone compared products in your category, often against a named rival | Narrow coverage, only accounts that use review sites | | Bidstream | Ad exchange bid requests carrying IP and page URL | Rough topical interest at scale, cheaply | Noisy, duplicative, most exposed under EU rules | | First party de-anonymisation | Your own site traffic matched by reverse IP and identity graphs | This company read your pricing page twice this week | Match rate is partial, and remote workers break IP matching | | Technographic and hiring change | Install detection, job postings, funding records | A structural reason to buy appeared | Slow, and not a timing signal on its own | The first two are the ones most SaaS teams should care about. Bidstream tends to arrive bundled inside a platform rather than sold on its own, and technographic change is better handled inside an enrichment workflow than bought as an intent product. Treating a co-op topic surge as a buying signal. A surge on the topic 'data governance' can be triggered by an analyst, a compliance hire reading up, or a competitor's employees. It is a reason to look, not a reason to say a company is in market. ## Provider notes, one at a time ### Bombora The category's utility layer. Company Surge measures content consumption across a co-op of thousands of B2B publisher sites against each company's own historical baseline, across a topic taxonomy running into the thousands. Most other vendors either resell it or compete with a version of it. Bombora is the right pick when your ICP is broad, your account list is long, and you want to steer topic clusters rather than chase individual accounts. It is the wrong pick when you sell a narrow product to 800 companies, because a weekly topic surge across a list that small produces too few usable events to justify the price. ### G2 Buyer Intent The highest precision signal available to most mid market SaaS. A G2 event tells you someone looked at your category page, your profile, or a comparison between you and a named competitor. That last one is the single most actionable signal in this market, and it is the reason [how B2B SaaS buyers find vendors](/research/how-b2b-saas-buyers-find-vendors/) keeps pointing back at review platforms. The catch is coverage. G2 only sees accounts that visit G2, which skews toward categories where review sites drive real research volume. Security, dev tooling and finance software do well. Very new categories do badly, because the category page barely exists. ### 6sense A platform, not a feed. 6sense combines co-op topics, bidstream, keyword activity and your own web and CRM data into an account level buying stage, then lets you run ads and orchestration against it. The 2021 Slintel acquisition added technographic and contact data underneath. Buy 6sense if you are already running structured [account based marketing for SaaS](/guides/account-based-marketing-saas/) with a budget for display. Do not buy it as a data source you will export to a CSV, because you will pay platform pricing for feed value. ### Demandbase The closest competitor to 6sense, with deeper roots in advertising delivery and a stronger pull with enterprise teams. Demandbase One bundles intent, account identification, ad delivery and sales insight, and its 2021 acquisitions of InsideView and DemandMatrix filled in company and technographic data. Pick Demandbase over 6sense when advertising is the primary activation and you want the media buying in the same system. Pick 6sense when predictive stage scoring and sales orchestration matter more than ad delivery. ### ZoomInfo Intent Worth considering only if you already pay ZoomInfo for contact data. The intent layer, built substantially on the 2020 Clickagy acquisition plus co-op sources, is priced as an add-on and lands in the same workspace your reps already use. Signal quality is middling. Workflow proximity is excellent, and workflow proximity is what actually drives usage. ### Website de-anonymisation tools RB2B, Warmly, Vector and HubSpot's Breeze Intelligence (the former Clearbit) resolve traffic you already paid for. This is the cheapest intent money you can spend and the first thing most teams should buy. Person level identification is restricted to United States traffic by several vendors for privacy reasons, and match rates run between a quarter and just under half of sessions in the mid market SaaS programs we have reviewed. **25% to 45%** Share of sessions resolved to a company by de-anonymisation tools on mid market SaaS sites ## The comparison table The pattern in that table is the argument of this page. Precision and price move in opposite directions. The cheapest signals, your own site visits and review site behaviour, are the most specific about purchase intent. The most expensive products buy you coverage, modelling and delivery, which matter when you have a large account list and an ads budget and matter very little when you have neither. ## What accuracy you should actually expect Lower than the demo implies, and that is survivable if you plan for it. Company level attribution is solid enough to act on. Person level claims should be treated as marketing unless the vendor shows you the identity source. In the pilots we have reviewed, 8 to 15 percent of a raw co-op surge feed matched the buyer's own ICP definition before any filtering. That number is not a scandal, it is the point: you are meant to filter. The failure comes when a team accepts the raw feed as a work queue and burns four weeks of SDR time on companies that were never targets. Two honest failure modes to budget for. The first is baseline drift, where a company that always reads a topic never surges, so your best-fit accounts go invisible. The second is remote work breaking IP resolution, which quietly hollows out de-anonymisation match rates for distributed companies and makes your data look worse in exactly the segments that buy software fastest. If a surging account will not get a human touch within two business days, do not buy intent data this year. Buy sales capacity instead. Every program we have seen fail failed at this step, not at vendor selection. ## How to validate a provider with a 60 day pilot Vendors will offer you a sample list of surging accounts. Refuse it. Run the test backwards against outcomes you already know, because that is the only version a CFO will believe. **A 60 day intent pilot that produces a defensible answer** Write the success threshold down before the pilot starts and give it to the vendor. A pilot without a pre-agreed number becomes a negotiation about vibes in week nine, and the vendor wins that negotiation every time. ## The plays intent data should trigger A signal with no attached play is a notification. The programs that work define four or five plays in advance and route each signal type to exactly one of them, which is the same discipline that makes [signal based outbound for SaaS](/playbooks/signal-based-outbound-saas/) work at all. **Plays worth wiring before the contract starts** Notice how few of those plays are new emails. Most of the return comes from suppression and routing rather than volume. The single highest return play in the list is the last one on the pricing page, and it costs nothing beyond a workflow rule. Intent also improves media efficiency, which is where the platform vendors earn their premium. Feeding a surging account list into paid social and into [newsletter and podcast sponsorships for B2B SaaS](/playbooks/podcast-and-newsletter-sponsorships/) tightens frequency against companies actually in a cycle instead of spraying the full ICP every month. ## What this costs, and who should walk away Budget the contract at roughly a third of the real cost. The rest is people. A platform deployment takes 6 to 10 weeks of someone's attention, plus ongoing ownership of ICP filters, topic mapping and CRM field hygiene. Feed-only products are lighter, but somebody still has to maintain the filter logic monthly or the list degrades into noise by quarter two. Three types of company should skip intent this year. Pre product market fit companies, because your ICP will change twice before the contract ends. Companies under about $3M ARR with fewer than two dedicated sellers, because capacity is the binding constraint. And any team whose CRM cannot reliably match leads to accounts, since every intent product assumes account level truth you do not yet have. Fix that first, and read [B2B attribution software compared](/guides/b2b-attribution-software-compared/) before you spend on either. For everyone else, the sequencing is straightforward. Start with de-anonymisation of your own traffic for a few thousand dollars. Add review site intent next, since it is the most specific signal per dollar for mid market SaaS. Only add a platform when you have ad budget to orchestrate and a defined set of [buying group](/guides/buying-group-demand-generation/) plays waiting for it. ## What to do next Pick one thing this week: install a de-anonymisation script and route company-level pricing page visits into a Slack channel your AEs already read. It costs a few hundred dollars a month and it will tell you more about whether your team can act on signals than any vendor pilot will. If that channel produces conversations inside 48 hours for a month straight, you are ready to buy a real feed and can plan the spend against your wider [B2B SaaS demand generation strategy](/guides/b2b-saas-demand-generation-strategy/). If the channel goes unread, you have saved yourself a five figure contract and learned something more useful than any surge score. Start there, and see [SaaS demand generation](/saas-demand-generation/) for how this fits the rest of the program. ## Frequently asked questions ### What is the best intent data provider for a B2B SaaS company? It depends on motion. Mid market SaaS with a named competitor set gets more from G2 Buyer Intent, because the buyer explicitly viewed a category or comparison page. Enterprise teams running coordinated ads and outbound get more from 6sense or Demandbase. Teams with meaningful site traffic and no budget should start with de-anonymisation of first party visits. ### How much does B2B intent data cost per year? Standalone feeds start around $18,000 to $45,000 a year depending on topic count and account volume. Full platforms including advertising and orchestration run $55,000 to $130,000 for a mid market SaaS company, more at enterprise seat counts. Website de-anonymisation tools start near $3,000 a year and rarely pass $25,000. ### Is G2 Buyer Intent better than Bombora? For precision, yes. A G2 signal means someone compared products inside your category on a review site, often against a named rival. A Bombora surge means content consumption on a topic rose above that company's baseline across a publisher co-op. Bombora covers far more accounts. G2 covers fewer accounts with far higher purchase relevance. ### How accurate is B2B intent data? Company level attribution is reasonable and person level is not. In pilots we have reviewed, 8 to 15 percent of a raw co-op feed matches the buyer's ICP before any filtering, and match rates on website de-anonymisation land between 25 and 45 percent of sessions for mid market SaaS. No provider can tell you which individual researched, and most will not claim to. ### Does intent data actually increase pipeline? Only through speed and sequencing. Intent does not create demand. It reorders accounts you had already targeted, so the lift comes from contacting a surging account before a competitor does and opening on the right topic. Teams that route signals into a queue nobody works see no lift at all, which is the most common outcome. ### Should a seed stage SaaS company buy intent data? Almost never. Below roughly $3M ARR you usually lack both the target account list that makes intent useful and the sales capacity to work it inside two days. Spend the same money on first party de-anonymisation of your existing traffic and on a better account list. Revisit intent once two or more reps are working named accounts full time. ### Is intent data legal under GDPR? Company level intent built from consented publisher co-ops is generally workable in the EU. Bidstream derived data and person level identification of EU visitors carry more exposure, and several de-anonymisation vendors restrict person level output to United States traffic for this reason. Get your data protection officer to review the vendor's lawful basis before signing, not after.