# Intent Data and Buying Signals for Outbound Sales

> Canonical: https://www.yalc.ai/blog/intent-data-buying-signals/

Buying one intent feed covers a tenth of the surface area. The win is a four layer stack with one router deciding which signal fires the outreach.

Intent data is any behavioral or contextual signal that an account is moving closer to a buying decision. It spans third party topic research, comparison and review site activity, first party visits to your own site, and company events like hires and funding rounds. Buying signals are the individual data points inside that feed. The useful definition treats intent as four layers, not one product category.

That distinction decides whether intent data works for you. Most teams buy a single feed, run one generic sequence, and conclude intent data is hype. It is not hype. It is mislayered.

## What intent data actually is, and where the term gets muddy

The textbook taxonomy splits intent into 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 publisher network). That split is accurate and close to useless for an operator, because nobody buys a product called "third party data." You buy a surge feed, a comparison feed, a visitor identification tool, or a company events feed. Each is a separate product at a separate price.

The reason no single source is enough is that buying behavior happens in too many places at once. An 800 person fintech can be reading about fraud prevention on a publisher network, viewing your competitor on a review site, hitting your pricing page through a visitor ID tool, and announcing a new VP of Sales on LinkedIn, all inside the same week. Each event is true. Each is invisible to the other three feeds. Buy one and you see roughly a tenth of the surface area.

There is a hard reason this matters in 2026. According to 6sense's [2024 B2B Buyer Experience Report](https://6sense.com/science-of-b2b/2024-buyer-experience-report/), 81 percent of buyers have a preferred vendor before they ever speak with sales, 85 percent have established their purchase requirements first, and 69 percent of the purchase process is over before a seller is involved. If most of the decision happens before contact, the only way to enter early is to catch the signal that the account is moving, and one feed catches a fraction of those signals. The broader [operator playbook for signal based outbound](/blog/signal-based-outbound/) only compounds once the layers are stacked.

## What are the four layers of intent data

There are four useful layers. Each answers a different question, fires on a different cadence, and routes to a different message. The non-obvious decision rule that follows from this is the spine of the whole piece: cadence has to match signal heat, because a hot signal in a slow nurture wastes the window, and a cold signal in a same week meeting ask reads as creepy.

### Category surge: topic research at the account level

A publisher network detects spikes in content consumption on a topic, attributes the spike to an account via IP and cookies, and sells a weekly account list ranked by surge intensity. Bombora created the category and most B2B sales platforms resell or repackage it. What it tells you is that an account is suddenly reading three times more about a topic than its baseline. What it does not tell you is which contact is reading, or whether the account buys this quarter. It is right for nurture queues and ABM audiences, and wrong for "send today, ask for a meeting tomorrow."

### Comparison and evaluation: late stage public signals

A narrower, hotter layer. [G2 Buyer Intent](https://sell.g2.com/buyer-intent), TrustRadius, and Capterra fire when a known account views your product page, a competitor page, or a side by side comparison. G2 surfaces category page views, competitor comparisons, pricing engagement, and review consumption as verified activity. By the time someone reads three competitor pages, they are evaluating, not researching. Buying readiness is high, signal density is low, and it earns warm outbound the same week.

### Behavioral: your site, your inboxes, your funnel

Your own first party data. Visitor identification, form fills, repeat pricing views, content downloads. [RB2B](/tools/rb2b/) sits here for B2B teams that lack the traffic to identify visitors server side themselves. This is the strongest layer because there is no inference. This person, at this company, did this thing on your property. It earns immediate outreach inside a 24 to 72 hour window. The only ceiling is the thinness of your own funnel.

### Company events: hires, funding, launches, churn

The fastest moving layer. Hiring announcements, job change tracking, funding rounds, product launches, executive turnover, technographic shifts. [PredictLeads](/tools/predictleads/) and [Crustdata](/tools/crustdata/) cover this for most operators. The signal is not "they read your topic," it is "their reality changed." They hired their first head of growth last Tuesday. They raised a Series B. They switched off Marketo. The reasoning chain to outreach is short and the message nearly writes itself. It is worst when dumped into a generic sequence as a flat list of accounts.

## Provider map and 2026 pricing per layer

The mapping is cleaner than most articles make it look. Pick one provider per layer, never two in the same layer. The expensive mistake is buying two products in one layer because they were sold in separate demos, then calling the overlap "coverage."

| Layer | Lead provider | Buying readiness | Public 2026 pricing |
|---|---|---|---|
| Category surge | Bombora Company Surge plus resellers | Low | Enterprise only, no public rate card |
| Comparison and evaluation | G2 Buyer Intent, TrustRadius, Capterra | High | Demo only, free 30 day lookback offer |
| Behavioral | RB2B plus your own analytics | Highest | Free tier 150 resolutions, paid from $79/mo |
| Company events | PredictLeads, Crustdata | High | PredictLeads pay as you go, $40/mo minimum |

The two layers with public pricing are the two you should build first. Per the [RB2B pricing page](https://www.rb2b.com/pricing) fetched June 2026, the free tier covers 150 monthly resolutions at company level, Starter is $79 per month for 300 resolutions, and Pro is $149 per month for 600 resolutions with business email addresses. Per the [PredictLeads pricing page](https://predictleads.com/pricing), the model is pay as you go credits: 100 free per month, then a $40 monthly minimum at $0.04 per credit for the 101 to 5,000 tier, dropping to $0.01 per credit above 10,000. Most company API requests cost one credit regardless of records returned, though discovery endpoints charge per company returned.

## How buying signals overlap, and how to deduplicate them

The unified stack creates a new problem the second you turn it on. Two providers fire on the same account in the same week and the queue starts double sending. The prospect gets two notes from two angles, which reads worse than silence.

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

The second rule is suppression. Any account with an active conversation, an open opportunity, or a recent reply in either direction leaves every signal triggered queue. Firing intent data into a live conversation is the clearest way to sound like an algorithm rather than an operator. Treat the active pipeline as a hard suppression list that runs before any sequence. This is painful inside a vendor UI and trivial in a markdown configured stack. [The leads qualification skill](/skills/qualify-leads/) runs exactly this dedup and suppression pass before anything sends.

## Which buying signal triggers which sequence

The biggest gain is not the data, it is the routing. The data feeds are commodities and the router is the moat. Treating every signal the same is how a team spends real money on feeds and books the same number of meetings it 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 mode, so pitch the artifact and earn the second read.
- Comparison page fires: route to a same week reply ask, two touches in seven days, both referencing the named competitor without trashing it.
- Behavioral visit fires: route to a 24 to 72 hour outreach, one LinkedIn touch within a day, one email within three. Acknowledge what the visitor saw without naming the tracking.
- Company event fires: route to a one to one note written by the operator. No sequence. This is last mile work, not automation.

Four signals, four sequences. The [buying trigger outbound playbook](/blog/buying-trigger-outbound/) covers trigger framing in more depth. This piece is about wiring the stack underneath it.

## Why a markdown router beats a CRM workflow graph

The four layer stack needs an orchestration layer that reads the signals, runs the dedup and suppression pass, picks the sequence, drafts the first touch, and logs the result. Most teams try to build this inside a CRM with custom fields, or a workflow graph with dozens of nodes. Both break for the same reason. Intent signals change shape every quarter, and a node graph forces a refactor every time a provider adds a new event type.

The operator stance is to replace the orchestration layer, not the data. yalc is open source and built on Claude Code, so the router lives in a folder of markdown files. It pulls signals from Crustdata, PredictLeads, RB2B, and G2 through real APIs, runs the markdown router, and drops the qualified queue into the sending tool. A new event type from PredictLeads becomes a five line edit to a router file, not a vendor sponsored integration project. That interoperability is the difference between a stack that compounds and one that gets rebuilt every quarter. The [operator's view on AI SDR tools](/blog/ai-sdr-tools/) goes deeper on where the budget actually compounds.

## What a sub $500 monthly intent stack costs

Ranking articles imply intent data is an enterprise line item. Bought by layer, it is not. Using the verified public pricing above, a workable stack for a small operator team sits inside a sub $500 monthly software budget.

- Behavioral: RB2B Pro at $149 per month for 600 resolutions with business email addresses, per the [RB2B pricing page](https://www.rb2b.com/pricing).
- Company events: PredictLeads pay as you go, roughly $40 to $200 per month for a small workload at the $0.04 per credit tier, per the [PredictLeads pricing page](https://predictleads.com/pricing).
- Comparison: G2 Buyer Intent quoted through sales, with a free 30 day lookback to start.
- Category surge: optional, added after the first three layers route reliably.

That is roughly $200 to $350 in real monthly spend. The edge is never in spending more on feeds. It is in routing the signals you already have.

## Worked example: three signals on one prospect in one week

Monday, Crustdata flags that a 240 person logistics SaaS hired its first VP of Demand Generation. The router pulls the new VP plus three plausible buyers. Wednesday, RB2B identifies someone at the same company hitting the pricing page from a Chicago IP. Same domain, same week. Friday, G2 flags that the company viewed a competitor comparison twice in 48 hours. Same domain, same week.

Three signals, one prospect company. Without dedup, that is three sequences in five days and a buying committee buried under six to nine disconnected notes, which 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 about the team they are building. The behavioral signal updates the message ("saw your team is already checking pricing, 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 weighing us against X, the differentiator is Y, and the engineering lead will be on the call"). One message, three reasons it lands, one operator decision.

## Four mistakes that waste intent spend

Four failures show up in almost every intent program that does not convert.

First, buying two products in the same layer. The overlap looks like coverage and feels like waste two months in. Audit by layer, keep one per layer, cancel the duplicate. Second, running every signal through the same sequence, which ignores the cadence to heat rule above. Third, treating intent feeds as lead lists. A surge account is a hint that a lead exists inside that account, not a lead. You still source the contacts and qualify the committee through tools like Apollo, ZoomInfo, or Crustdata. Fourth, no feedback loop. The router never sharpens 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.

## Frequently asked questions

### What is intent data?

Intent data is any behavioral or contextual signal that 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 it as four layers, each with different providers, cadences, and message angles, rather than one product category.

### What are buying signals in sales?

Buying signals are the individual data points that suggest readiness to purchase. Examples are a surge in topic research at an account, a visit to your pricing page, repeated views of a competitor comparison, a new executive hire, and a funding round. A single signal is a hint. Several signals across different layers inside a 14 day window is a near certainty the account is in market.

### How do you use intent data for outbound?

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

### How much does intent data cost?

Two of the four layers publish pricing. RB2B for visitor identification starts free for 150 resolutions and runs $79 to $149 per month for paid tiers, per its pricing page. PredictLeads for company events is pay as you go with 100 free credits, then a $40 monthly minimum at $0.04 per credit. Category surge from Bombora and comparison data from G2 are quoted through sales with no public rate card.

### Is intent data worth it for outbound?

It is worth it when stacked and routed, and weak 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. Teams that get poor results usually bought one feed and ran one generic sequence. With 81 percent of buyers picking a preferred vendor before talking to sales, per 6sense's 2024 report, catching the early signal across multiple layers is often the only way to enter the deal in time.