Intent data providers compared, by someone who does not sell intent data

Almost every comparison of intent data providers is published by an intent data provider. That is not a conspiracy, it is just who has the commercial incentive to write one. It does mean the genre has a predictable shape: four vendors are described fairly, and the fifth, which happens to be the publisher, solves the problems the other four have.

We sell go-to-market work. We buy this category, we do not sell it. So here is the comparison with the incentive removed.

The short version: the category is real, most teams buy the wrong tier of it, and a meaningful minority should not buy it at all.

The four things sold as "intent data"

The first problem is the label. Four genuinely different products share it, and they are not substitutes.

Most confusion in this market comes from teams comparing a review-intent product against a co-op content product as though they do the same job. They do not. One tells you someone is comparing vendors in your category. The other tells you a company's employees read more articles about a topic than they usually do.

Co-op content consumption

The largest and oldest segment. A network of publishers shares reader data, IPs resolve to companies, and a surge is flagged when a company's consumption on a topic exceeds its own baseline.

Real strength: breadth. It sees activity across the open web that you could never observe yourself, including accounts that have never touched your site.

Structural weakness, and it is not fixable by picking a better vendor: the reader is anonymous, so you cannot know whether it was a decision-maker, an analyst, a student, or a competitor doing research. Topic mapping is coarse enough that adjacent categories blur. And the baseline is statistically noisy at companies under a few hundred people, which is where the false positives cluster.

Use it as a prioritisation tiebreaker between two otherwise equal accounts. Do not use it as a trigger to launch outreach, which is exactly how it is most often deployed.

Review site intent

A buyer comparing vendors on a review platform has declared evaluative intent in a way that reading an article never does. The behaviour is much closer to purchase.

Real strength: precision. Category and comparison browsing is genuinely hard to misread.

Weakness: narrow and late. You only see buyers who reached a review platform, which means a process is already running and someone else likely framed it. You are joining a race in progress.

Best value-per-dollar in the category for most B2B SaaS teams, and the one most often skipped in favour of the bigger, broader, more expensive option.

Website de-anonymisation

Identifies the companies visiting your own site.

Real strength: it is your own first-party behaviour, which is the most predictive signal class there is. Repeat pricing page visits from several people at one company is the strongest signal in B2B, as we argue in the parent piece on B2B buying signals.

Weaknesses to know before you buy: company resolution is good, person resolution is poor and getting worse. Remote work has broken IP-to-company matching badly, and vendor-quoted match rates are measured generously. Expect materially less than the demo suggests, particularly outside the US. Privacy regulation is also tightening around this, and the compliance position varies by jurisdiction.

Cheapest meaningful entry point into the category. Covered in more depth in website visitor identification.

Predictive account scoring

A model that ingests intent, firmographics and your CRM history and returns a ranked account list.

Real strength: when it works, it works. It is the only tier that combines sources into a single decision.

The catch nobody puts in the deck: the model learns from your closed-won data. If you have fewer than a few hundred closed deals, there is not enough signal to learn from and you are buying a confident-looking output built mostly on generic priors. Small teams pay enterprise prices for a dressed-up firmographic filter.

Only worth it above meaningful deal volume and with clean CRM data. The second condition disqualifies more teams than the first.

What the sales process will not volunteer

Four things worth raising in the demo. The answers are more informative than the product tour.

1. "What share of accounts you flag as surging enter a buying process within 90 days?"

This is the only number that matters and it is rarely published. Vendors who will answer honestly are worth taking seriously. Vendors who redirect to case studies have told you something.

2. Coverage outside North America. Frequently much weaker, and rarely stated up front. If a material share of your pipeline is European or APAC, ask for coverage by region rather than in aggregate.

3. Refresh latency. Some feeds update daily, others weekly. Combined with the fact that a surge already lags the underlying research, a weekly feed can mean acting on behaviour from three weeks ago. Signals decay fast, and stale intent is worse than none because it feels actionable.

4. What happens on cancellation. Whether scores persist in your CRM or evaporate. This affects switching cost far more than the contract price does.

Who should not buy this

Written plainly because no vendor will say it:

• Teams that have not instrumented first-party signals. If pricing page visits, product usage and email engagement are not landing somewhere a human sees them, you already own better data than you are about to buy. Fix that first. It is free.
• Teams with no capacity to act. A signal creates value only if something happens when it fires. If acting on an alert takes eleven manual steps across six tools, more alerts will not help. That is an orchestration problem, not a data problem.
• Teams with fewer than roughly 50 target accounts. At that size you should know all of them personally. Intent data solves a scale problem you do not have.
• Teams under about $15k ACV. The cost per incremental deal rarely clears.

What to buy, in order

If you are starting from nothing:

1. Instrument first-party signals. Costs nothing but attention. Most teams find they already had the best data in the building.

2. Add website de-anonymisation. Cheapest paid tier, highest precision, directly extends what you already own.

3. Add review site intent if your category has meaningful review platform presence. Best precision-per-dollar of the third-party options.

4. Consider co-op content data only when you need reach into accounts that have never touched your properties, and only as a prioritisation input.

5. Consider predictive scoring only above a few hundred closed deals with clean CRM data.

Most teams jump straight to step 4 or 5 because those are the products with the largest sales teams. The order above is roughly inverse to the order you will be sold.

Whatever you buy, the signal still needs weighting and decay applied before it drives anything. That model is in signal-based lead scoring.

Frequently asked questions

What are the main types of B2B intent data?

Four: co-op content consumption across publisher networks, review site intent from comparison platforms, website de-anonymisation of your own traffic, and predictive account scoring that blends these with firmographics. They are commonly compared as substitutes but they observe different behaviour and are not interchangeable.

How much does intent data cost?

Roughly $5k to $30k a year for website de-anonymisation, $10k to $50k for review site intent, $15k to $60k for co-op content networks, and $40k to $150k or more for predictive account scoring. Pricing usually scales with target account volume rather than seats.

Is intent data accurate?

It depends on the type. Review site intent and first-party website data are precise because the behaviour is explicitly evaluative. Co-op content data is structurally imprecise: the reader is anonymous, topic mapping is coarse, and detection lags the underlying research by weeks. That is a property of the method, not of any particular vendor.

Do you need intent data for ABM?

No. Intent is one of four account selection dimensions alongside fit, reachability and winnability, and first-party signals cover it adequately for most programs. Buy third-party intent when your first-party volume is genuinely too thin to fill a pipeline.

Which intent data provider is best?

There is no single best, because the four product types solve different problems. For most B2B SaaS teams the best value-per-dollar is review site intent combined with website de-anonymisation. Predictive scoring only earns its price above several hundred closed deals with clean CRM data.

---

*NomiOS is RZLT's GTM and ABM engine. Point it at a target and get back finished, branded work built on a real read of that company.*

[See how NomiOS works →](https://nomios.rzlt.io)