Intent data for outbound works best as a stack, not a single source. Category surge tells you what topic an account is researching. Comparison signals reveal late stage evaluation. Behavioral data shows who visited your site. Company event data captures the hire, the funding round, the launch. The unified play layers all four and routes each signal type to a different sequence.

That sounds obvious until you watch a team buy one feed and walk away saying intent data does not work. It works, just not the way vendors sold it.

Why one intent data source is never enough in 2026

Intent data is any behavioral or contextual signal that suggests an account is getting closer to a buying decision. The classic version is third party research data, but the operator version is broader. Anything that says "this account moved" counts. No single provider covers the full picture because buying behavior happens in too many places at once.

A company researching a topic on Bombora's publisher network is not the same company hitting your G2 comparison page, which is not the same company quietly visiting your pricing page through RB2B, which is not the same company that just announced a new VP of Sales on LinkedIn. All four can be true in the same week. Buy one and call it intent data, and you see a tenth of the real surface area.

Most ranking articles on this topic explain first party, second party, and third party data and stop there. That taxonomy is fine but not actionable. Operators do not buy "third party data." They buy a surge feed, a comparison feed, a visitor ID tool, or a company events feed. Each is a different product at a different price point. The shift in 2026 is treating intent as a layer cake instead of a category. The operator playbook for signal based outbound only compounds when the layers are stacked.

The four layers of intent (and which provider owns each)

There are four useful layers of intent data. Each one answers a different question. Each one fires on a different cadence. Each one routes to a different kind of message.

1. Category surge (topic research at the account level)

The classic third party feed. A network of publishers detects spikes in content consumption on a given topic, attributes the spike to an account via IP and cookies, and sells you a weekly account list ranked by surge intensity. Bombora invented the category, and most B2B sales platforms now resell or repackage it under their own brand.

What it tells you: this 800 person fintech is suddenly reading three times more about fraud prevention than its industry baseline. What it does not tell you: which contact is reading, or whether the account will buy this quarter. Useful for nurture queues and ABM audiences. Wrong tool for "send today and ask for a meeting tomorrow."

2. Comparison and evaluation (late stage public signals)

A narrower, hotter layer. G2 buyer intent, TrustRadius, Capterra. Signals fire when a known account views your product page, your competitor page, or a side by side comparison. By the time someone reads three competitor pages on G2, they are evaluating, not researching.

The buying readiness here is much higher than category surge but the signal density is lower. Useful for warm outbound the same week and competitive deal alerts.

3. Behavioral (your site, your inboxes, your funnel)

Your own first party data. Visitor identification, form fills, repeat pricing page views, content downloads. RB2B sits squarely in this layer for B2B teams without enough traffic to do server side identification themselves.

The strongest signal on the list because there is no inference involved. This person, at this company, did this thing on your property. Useful for immediate outreach inside a 24 to 72 hour window. The thinness of your own funnel is the only ceiling.

4. Company events (job changes, hires, funding, launches)

The newest and fastest moving layer. Hiring announcements, job change tracking, funding rounds, product launches, executive turnover, technographic shifts. PredictLeads and Crustdata sit on top of this layer for most operators. The signal is not "they read your topic." It is "their reality just changed."

This account just hired their first head of growth last Tuesday. This account just raised a Series B. This account quietly switched away from Marketo. The reasoning chain to outreach is much shorter than surge data, and the message almost writes itself. Worst when treated as a list of accounts to dump into a generic sequence.

Provider map per layer

The mapping is cleaner than most articles make it look.

  • Category surge: Bombora's Company Surge plus resellers who repackage it. Pricing is enterprise only, with no public rate card on the Bombora product page as of June 2026.
  • Comparison and evaluation: G2 buyer intent, TrustRadius, Capterra. Pricing routes through demo requests, not a public card.
  • Behavioral: RB2B for visitor identification, plus your own product analytics. Per RB2B pricing fetched June 2026, the Pro plan starts at $149 per month for 600 monthly resolutions with email addresses, and a free tier covers 150 anonymous resolutions.
  • Company events: PredictLeads for hires and funding, plus Crustdata for live company and people data. PredictLeads runs on pay as you go credits, priced at $0.04 per credit for the 101 to 5,000 tier with a $40 monthly minimum, dropping to $0.01 at 10,001 credits.

Pick one provider per layer, not four. The trap is buying two products in the same layer because they were sold in different demos.

How signals overlap, and how to deduplicate

The unified stack creates a new problem the moment you turn it on. Two providers fire on the same account in the same week and your queue starts double sending. The prospect gets two notes from two different angles, which reads worse than no notes at all.

The rule that works: latest signal as trigger, older signal as enrichment context. When a new signal fires on an account that already has an open signal in the last 21 days, do not start a second sequence. Use the new signal as the cadence trigger, the most recent as the message angle, and the older one as context inside the same first touch. One outreach, two reasons for it.

The other rule is suppression. Any account with an active conversation, an open opportunity, or a recent reply in either direction gets removed from every signal triggered queue. Intent data into existing conversations is how you sound like an algorithm, not an operator. Treat the active pipeline as a hard suppression list before any sequence runs.

This is painful inside a vendor's UI and trivial in a markdown configured operator stack. The leads qualification skill handles exactly this dedup pass before any signal triggered queue gets sent.

Routing logic: which signal triggers which sequence

The biggest leverage point in a unified intent setup is not the data, it is the routing. Different signals warrant different cadences. Treating them all the same is how teams burn $2,000 a month on intent feeds and book the same number of meetings they booked last quarter.

The routing pattern that compounds:

  • Category surge fires: route to a slow nurture. Three touches across three weeks. No meeting ask in touch one. The reader is in research mood. Pitch the artifact, earn the second read.
  • Comparison page fires: route to a same week reply ask. Two touches in seven days, both reference the named competitor without trashing them. Make the comparison concrete and offer the call.
  • Behavioral visit fires: route to a 24 to 72 hour outreach. One LinkedIn touch within a day, one email within three days. The window matters more than the cadence. Acknowledge what the visitor saw without naming the tracking.
  • Company event fires: route to a personalized one to one. No sequence. The operator writes the note. This is last mile work, not automation.

Four signals, four sequences. The data feeds are commodities. The router is the moat. The buying trigger outbound playbook covers the trigger framing in more depth; this piece is about wiring the stack underneath it.

Yalc as the unified orchestrator

The four layer stack does not work without an orchestration layer underneath. That layer reads the signals, runs the dedup pass, picks the sequence, writes the first touch, and logs the result. Most teams try to build this inside a CRM with custom fields or a workflow graph with 40 nodes. Both break for the same reason: intent signals change shape every quarter, and you do not want to refactor a graph every time PredictLeads adds a new event type.

Yalc replaces the orchestration layer instead of the data. It pulls signals from Crustdata, PredictLeads, RB2B, and G2 through real APIs, runs a markdown configured router, and drops the qualified queue into your sending tool. The router gets sharper every week as you mark which signals converted. The middle mile runs in the background. The operator writes the company event notes, takes the discovery calls, and owns the angle.

The architectural property doing the work is interoperability. A new event type from PredictLeads becomes a five line update to the router file, not a vendor sponsored integration project. That is the difference between a stack that compounds and one that has to be rebuilt every quarter.

Cost stack for a unified intent setup

Ranking articles imply intent data is an enterprise budget item. It is not, if you buy by layer. A workable unified stack for a small operator team in 2026 sits inside a sub $500 monthly software budget.

  • Behavioral: RB2B Pro at $149 per month for 600 monthly resolutions with business email addresses, per the RB2B pricing page.
  • Company events: PredictLeads pay as you go, typically $40 to $200 per month for a small operator workload, per the PredictLeads pricing page.
  • Comparison: G2 buyer intent is quoted through sales; a free profile gets you started.
  • Category surge: optional. Add Bombora or a reseller after the other three layers are routing reliably.

That is roughly $200 to $350 in real monthly spend, less than a single ZoomInfo seat on most contracts. The leverage is not in spending more on intent. It is in routing the signals you already have. The operator's view on AI SDR tools goes deeper on where the budget actually compounds.

Worked example: three signals on one prospect in one week

Monday morning, Crustdata flags that a 240 person logistics SaaS just hired its first VP of Demand Generation. The router pulls the new VP and three other plausible buyers.

Wednesday afternoon, RB2B identifies that someone at the same company hit your pricing page from a Chicago IP. Same domain. Same week.

Friday morning, G2 flags that the company viewed your competitor's comparison page twice in 48 hours. Same domain. Same week.

Three signals. One prospect company. Without dedup, you would have run three sequences in five days and the same buying committee would have received six to nine disconnected notes. That is the algorithmic smell that kills outbound.

With routing, the company event lands first. The operator writes a one to one note to the new VP referencing the team they are building. The behavioral signal updates the message: "saw your team is checking pricing already, happy to set up an account so you can poke around without us in the room." The comparison signal becomes a P.S.: "if you are comparing us against X, our differentiator is Y, and the engineering lead will be on the call." One message. Three reasons it lands. One operator decision.

Common mistakes that waste intent spend

Four mistakes show up in almost every intent program that fails to convert.

The first is buying two products in the same layer. Two surge feeds, two visitor ID tools, two job change feeds. The overlap looks like coverage and feels like waste two months in. Audit by layer, pick one per layer, cancel the duplicate.

The second is running every signal through the same sequence. Category surge in a same week meeting ask reads as creepy. Behavioral pricing visits in a slow nurture reads as asleep. The cadence has to match the signal heat.

The third is treating intent feeds as lead lists. A surge account is not a lead, it is a hint that a lead exists inside that account. You still have to source the contacts and qualify the buying committee. Apollo, ZoomInfo, Crustdata, and similar tools handle that layer.

The fourth is no feedback loop. The router never gets sharper because nobody marks which signals converted to replies and which converted to pipeline. Weight the router weekly. That is the compounding step most teams skip.

What to run this quarter

Pick the one layer you have not built yet and add it. Most teams already have some behavioral identification. Add company events next; the data is cheaper than category surge and converts faster. Build the suppression list before the new signal triggers anything.

Then write the router in markdown. Not in a vendor UI, not in a 40 node workflow graph. A folder of markdown files that says "if Crustdata fires this event, run this sequence, with this dedup rule." Your future self will read that folder in 20 minutes.

Intent data for outbound is not a category to buy. It is a layer cake to stack. The operators winning this quarter are the ones whose router is sharper than their data.

FAQ

What is intent data?

Intent data is any behavioral or contextual signal that suggests a B2B account is moving closer to a buying decision. It includes third party topic research, comparison and review site activity, first party site visits, and company events like hires and funding rounds. The useful definition treats intent as a layered phenomenon, not a single product category.

What are the different types of intent data?

The classic taxonomy is first party (your own site and product data), second party (someone else's first party data shared through a partnership), and third party (research and consumption data sold by a network of publishers). The operator taxonomy is more useful: category surge, comparison signals, behavioral visits, and company events. Each layer has different providers, different cadences, and different message angles.

How do you use intent data for outbound?

Pick one provider per layer of intent. Wire each signal type to a different sequence: surge to a slow nurture, comparison to a same week reply ask, behavioral visits to a 24 to 72 hour outreach, company events to a one to one personalized note. Run a hard suppression check against your active pipeline before any sequence sends. Dedup overlapping signals by treating the latest signal as the trigger and older signals as enrichment context.

Where do you get intent data?

Category surge data comes from Bombora and its resellers. Comparison signals come from G2 buyer intent, TrustRadius, and Capterra. Behavioral data comes from your own product plus a visitor identification tool like RB2B. Company event data comes from PredictLeads and Crustdata. Most operators end up buying one provider per layer rather than a single bundled intent platform.

Is intent data useless for outbound?

No, but it is incomplete on its own. Account level intent tells you which company is moving, not which person is in the buying committee or what message will land. The teams that get poor results from intent data usually bought one feed, ran one generic sequence, and skipped the routing and contact layer. Treated as a layer in a stack, intent data lifts reply rates meaningfully. Treated as a magic lead list, it underperforms basic ICP filtering.

How do SDRs and AEs use intent data to prioritize outbound?

SDRs use category surge and behavioral signals to sort their daily account queue and pick which prospects to research first. AEs use comparison signals and company event triggers to time their outreach into named accounts already in flight. The shared pattern is signal density on the calendar: when an account fires multiple signals inside a 14 day window, that account jumps to the top of both queues.

What are the benefits of intent data when stacked correctly?

The benefit of a properly stacked intent program is signal density, not signal quantity. One signal on an account is a hint. Three signals across three layers in a 14 day window is a near certainty that the account is in market. The unified stack converts the hint into a routed sequence with the right cadence, the right message angle, and a suppression check against active pipeline. That is where the lift compounds.