# Intent data for SaaS lead generation > What first party, second party and third party intent data can and cannot tell you, plus the plays that turn a signal into a real conversation. Source: https://saas-marketing.net/guides/intent-data-lead-generation/ Topic: SaaS Lead Generation Type: guide 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-lead-generation/ ## Short answer Buyer intent data is a prioritisation input, not a lead source. First party data observes your own site and product behaviour, second party data comes from review marketplaces like G2 and Capterra, and third party data infers topic consumption across publisher co-ops. Match rates for anonymous visitor identification typically run 25 to 55 percent for US traffic, and most signals decay within 14 to 30 days. Intent spend rarely pays back below roughly $15K ACV. ## Key takeaways - Intent data ranks accounts you already had reason to contact; it does not manufacture accounts you did not. - Visitor identification match rates run 25 to 55 percent on US traffic and far lower in the EU because of consent rules. - Second party review marketplace signals convert best because the buyer is actively comparing vendors, not reading. - Most third party signals decay inside 14 to 30 days, so a weekly export into a quarterly nurture wastes the data. - Below roughly $15K ACV the per-account cost of intent tooling and follow-up rarely pays back. --- Intent data gets sold as a lead source and bought as a lead source, which is why most programmes disappoint within two quarters. It's a ranking function. It tells you which of the accounts you were already going to work deserve attention first, and it says almost nothing useful about accounts outside your list. Get that framing right and intent earns its cost. Get it wrong and you've paid $45,000 for a spreadsheet. ## What each type of intent data actually observes Three categories, and they observe genuinely different things. Confusing them is the root of most bad expectations. First party data is behaviour on surfaces you own: pricing page visits, docs searches, feature usage inside a trial, seat additions, a second user from the same domain signing up. It's the highest fidelity signal you will ever get because you observed it directly, and it's free apart from instrumentation work. Second party data comes from review marketplaces. When an account browses your category on G2 or views a competitor's Capterra profile, the marketplace knows. That buyer is comparing vendors, not idly reading, which makes this the highest converting external signal available. Third party data is inferred. Providers run co-ops of publisher sites, watch content consumption by IP, resolve the IP to a company and report topic surges. The account is usually real. The topic is usually right. What you don't get is a person, a role, or any sense of where they are in a buying process. **25% to 55%** Account-level identification rate for US business web traffic in 2026, down sharply since distributed work fragmented office IP ranges "What is your match rate on my traffic, measured against total sessions, not against sessions you already filtered?" Vendors quote match rates against qualifying traffic, which excludes consumer ISPs, VPNs and bots. That's how a 32 percent real rate becomes a 78 percent headline. Run a two week pilot on your own traffic and compute it yourself. ## Why intent is a prioritisation input and not a lead source Because a signal is not a person and a person is not a conversation. An intent record tells you an account showed activity around a topic. It does not tell you the CFO approved a budget line, or that anyone at the account would take your call. Run the arithmetic. A mid-market team gets 400 surging accounts a month from a third party provider. Suppose 55 percent are already in some stage of your funnel or outside your ICP, leaving 180 genuinely new. Your SDR team can work maybe 60 accounts a month properly. So the data's practical job is choosing which 60, which is exactly what a ranking function does. That's a real and valuable job. It's just not the job people buy it for. The [lead source mix research](/research/b2b-saas-lead-source-mix/) shows how small a share of pipeline pure outbound produces at most companies, and intent doesn't change that ratio, it improves the hit rate inside it. ## Five plays that actually produce conversations Each of these has a specific trigger, a specific window and a specific message. Generic "we noticed you were researching" emails fail because they're creepy and vague at the same time. ### Marketplace category surge follow-up Trigger: your account appears in a G2 or Capterra buyer intent feed for your category or a named competitor. Window: 72 hours, no longer. The buyer is shortlisting now. Message: lead with the comparison, not the observation. "Most teams evaluating us against Monday are deciding on the resourcing view. Here's how ours works in 90 seconds." Send a Loom, not a calendar link. Conversion from this play typically runs 3 to 8 times better than cold outbound in the same segment. ### Deanonymised account retargeting Trigger: an identified account visited pricing, security or integration docs without converting. Window: 7 days. Play: a tightly targeted LinkedIn campaign to 5 to 15 named roles at that account, running creative specific to the page they visited. Don't email them saying you saw them on your website. That converts badly and generates complaints. Advertise instead. The ad feels like coincidence, the email feels like surveillance. ### Product usage triggers Trigger: inside a trial, a second user from the same domain invites a third, or usage crosses a threshold you know correlates with conversion. Window: hours. This is the highest yield play in the list and it needs no vendor. It needs an event in Segment or PostHog, a rule, and a Slack alert to whoever owns that account. Companies routinely pay $50K for third party intent while ignoring the free version sitting in their own product telemetry. ### Job change alerts Trigger: a champion who used your product at a previous company starts a new role. Window: 30 to 90 days, because new leaders buy tools in their first two quarters. Clay and Apollo both surface this. The message writes itself and it's the only outbound in this list where a cold email genuinely feels welcome, because you have a real prior relationship. ### Competitor page visits Trigger: an identified account reads your "versus competitor" or alternatives page. Window: 14 days. This buyer is deep in evaluation. Route these to a salesperson directly rather than into nurture. A comparison page visit from a target account is worth more than most demo requests, and it usually gets treated as a pageview. **Standing up an intent programme in 30 days** If your SDR team takes three days to work a queue, intent data buys you nothing. The signal will have decayed and the buyer will have booked demos with three vendors. Fix response time before you fix targeting. This is the single most common reason intent contracts do not renew. ## What it costs and where the payback breaks Build the model before the pilot, not after the invoice. A typical mid-market setup: a third party intent platform at $45,000 a year, visitor identification at $18,000, plus the SDR capacity to work the output. Call it $63,000 in tooling. If the programme prioritises 720 accounts a year and your team works 500 of them, the tooling cost per worked account is $126, before any human time. Now the conversion side. If signal-prioritised outbound converts to opportunity at 6 percent versus a 2.5 percent baseline, those 500 accounts produce 30 opportunities instead of 12.5. The 17.5 incremental opportunities cost $3,600 each in tooling alone. At a 25 percent close rate that's 4.4 incremental customers, or roughly $14,300 of tooling cost per customer. That table is why the ACV floor sits around $15,000. Below it, put the money into first party instrumentation and [inbound lead generation](/guides/saas-inbound-lead-generation/), which scales without a per-account cost. Model your own volumes in the [lead goal calculator](/calculators/lead-goal/) first. ## The privacy constraints you cannot design around In the EU, IP addresses and cookie-derived identifiers are generally personal data under GDPR. Person-level deanonymisation of a site visitor without consent is not defensible, whatever a vendor's sales engineer tells you. Several providers now disable EU person-level matching entirely rather than argue the point. Account-level resolution from a corporate IP range sits on firmer ground, because you're identifying an organisation rather than an individual. That's still a processing activity you need to document, and your cookie banner has to reflect it honestly. Practical positions that hold up: - Account-level identification for EU traffic, person-level only for US and Canada. - No emails that reference observed browsing behaviour. Advertise to those accounts instead. - A data processing agreement with every intent vendor, and a note in your privacy policy naming the category of processing. - Suppression lists honoured across providers, not just in your ESP. The reputational risk is larger than the legal one. One prospect posting your "I saw you were on our pricing page" email to LinkedIn does more damage than the tooling ever earned. ## How intent fits alongside your other sources Intent improves outbound efficiency. It does not replace demand creation, and it competes for budget with things that do. If you're choosing between a $60,000 intent contract and a $60,000 content programme, the content programme creates demand that intent tools can later detect, and the intent contract detects demand someone else created. At most companies under $20M ARR the content programme wins that argument. Above that, when you already have category presence, intent starts earning its place. Compare it honestly against [outbound lead generation](/guides/saas-outbound-lead-generation/) and [referral lead generation](/guides/referral-lead-generation-saas/), both of which produce higher converting conversations at lower tooling cost. For a side by side of the platforms themselves, see [intent data providers compared](/guides/intent-data-providers-for-saas/) and the [tools directory](/tools/intent-data-providers-b2b-saas/). The strongest intent programmes I've seen run a control group permanently. Ten percent of ICP accounts are worked without signal prioritisation, all year. When renewal comes round, the comparison is a fact rather than an argument. It costs a little efficiency and buys you the only defensible answer to "is this working". ## What to do next Instrument your own product and comparison pages before you buy anything, then run a 14 day parallel pilot on two providers and compute the match rate yourself. Pick the play that fits your motion, give it one owner and a response window measured in hours. Then set up the control group on day one. Two quarters from now it's the only thing that will settle the renewal conversation, and it costs nothing to start. For the wider picture of where pipeline comes from, start at [SaaS lead generation](/saas-lead-generation/), and check the platform landscape in [intent data and visitor identification tools](/guides/b2b-intent-data-providers/). ## Frequently asked questions ### What is buyer intent data? Buyer intent data is behavioural evidence that an account is researching a problem or category you sell into. It comes in three forms: first party signals from your own site and product, second party signals from review marketplaces where buyers compare vendors, and third party signals inferred from content consumption across publisher networks. It indicates research activity, not purchase intent. ### Does third party intent data actually work? It works as a prioritisation layer for outbound, and disappoints as a lead source. Third party providers infer topic surges from IP-level content consumption across co-op publisher networks. The account is usually real and the topic is usually right, but you learn nothing about who at the account is researching or how far along they are. ### What is a good match rate for website visitor identification? For US business traffic, 25 to 55 percent account-level identification is typical in 2026, down from the higher figures vendors quoted before remote work fragmented office IP ranges. European traffic resolves much lower because consent requirements limit what can be processed. Anyone quoting above 70 percent is measuring against filtered traffic. ### How quickly do intent signals decay? Most third party topic surges are meaningful for 14 to 30 days. Review marketplace signals, where a buyer views your category or your competitor's profile, decay faster, often inside 7 to 14 days because the buyer is already shortlisting. Product usage triggers from your own data decay fastest of all, sometimes within hours. ### What ACV do you need for intent data to pay back? Roughly $15K and above in most cases. A mid-market intent platform costs $25K to $70K a year, and the follow-up requires SDR capacity. At $6K ACV with a 4 percent conversion on prioritised accounts, the arithmetic breaks. Below that threshold spend the money on first party product signals, which cost nothing but engineering time. ### Can you use intent data legally in the EU? With care. GDPR treats IP addresses and cookie-derived identifiers as personal data in most contexts, so deanonymisation of individuals requires a lawful basis and usually consent. Account-level firmographic resolution from business IP ranges is more defensible than person-level identification. Many providers restrict or disable EU person-level matching entirely.