Reviews
How to evaluate B2B intent data tools
A practical guide to B2B intent data tools, with checks for signal sources, sample data, match rates, score definitions, vendor testing, and contract terms.
What to take away
- Intent data answers a different question than lead scoring, and vendors that sell both rarely separate them.
- Ask for the source of the signal, the refresh interval and the identity match rate before you discuss price.
- Three answers should end the conversationno sample data, no definition of a qualified account, and no named data source.
- The contract must state what happens when match rates fall below the level you tested.
- Test on your own closed-won accounts for 30 days before you sign anything annual.
Scopes that are actually different jobs
B2B intent data tools collect signals from publishers, review sites, job boards and content networks. They tell you an organization showed interest somewhere. B2B lead scoring tools rank named people against your own historical conversion data. These are separate products with separate failure modes, and a vendor selling one as the other is a warning sign.
Three intent data scopes
Buyer intent data
- What it collects
- Publisher and review signals
- What it tells you
- Organization showed interest
- Failure mode
- Signal without person
- Best for
- Top-of-funnel research
Lead scoring
- What it collects
- Historical conversion data
- What it tells you
- Ranks named people
- Failure mode
- Ignores buying committee
- Best for
- Inbound lead triage
Account-level intent
- What it collects
- Multiple contacts at one firm
- What it tells you
- Account-wide buying activity
- Failure mode
- Aggregation hides contacts
- Best for
- US pipeline teams
A third category, account-level intent, aggregates multiple contacts at one firm. That is closer to what most US pipeline teams actually need, because B2B buying involves several people. Our guide to running account-based marketing covers why account scope changes the tooling question.
Five tools worth testing
The tools below draw on different intent signals and offer different scoring features. Their descriptions are starting points, not evidence that any score predicts pipeline. Use the tests in this guide to evaluate each vendor.
Five intent tools compared
Bombora
- Signal source
- B2B content syndication
- Scope
- Account-level intent
- Pricing model
- Annual, account volume
6sense
- Signal source
- Intent plus predictive model
- Scope
- Account and contact
- Pricing model
- No public per seat
Demandbase
- Signal source
- Ad network and IP data
- Scope
- Enterprise ABM
- Pricing model
- Custom pricing
ZoomInfo Intent
- Signal source
- B2B database and web
- Scope
- Sales teams on ZoomInfo
- Pricing model
- Add-on fee
- BomboraPublishes Company Surge scores from a B2B content syndication network. Suits teams that want named publisher sources and account-level intent. Pricing is usually annual based on account volume and seats.
- 6senseCombines intent with predictive buying stage and buying center. Good for ABM teams that need account and contact scoring together. No public per seat price.
- Demandbase OneSources intent from its advertising network and IP data. Fits enterprise ABM programs that want advertising and intent in one system. Custom pricing.
- ZoomInfo IntentDraws from ZoomInfo's B2B database and web activity. Works for sales teams already paying for ZoomInfo. Usually sold as an add-on fee.
- G2 Buyer IntentCaptures first-party review and comparison activity on G2. Useful for software vendors targeting accounts researching categories. Seat based pricing.
Questions to ask
Ask each vendor these questions in writing, and keep the replies:
- Which publishers, sites, or other sources contribute to each signal?
- How often are signals refreshed?
- What is the identity match rate against our domain list, and how is it calculated?
- What does each score mean, and how does an intent score differ from a lead score?
- Will you provide raw sample rows and run a 30-day test against our closed-won accounts?
Vendor questions to ask in writing
- Name the underlying source of each signal
- State refresh frequency and oldest record age
- Match rate against 500 US domains
- Define high-intent account and false positive rate
- Show scores for ten closed accounts
The answer to the sample-and-test question matters: a vendor that can provide usable records gives you evidence to evaluate. If it relies on a licensed feed and cannot expose underlying records, ask what it can provide before deciding whether the test is meaningful.
Evidence to request
Ask for a sample file, not a dashboard screenshot. You want raw rows with domain, date, signal type and score. Import it into your CRM and compare against closed-won accounts from the last two quarters.
How to test vendor evidence
- Request a sample file, not a dashboard screenshot
- Get raw rows with domain, date, signal type, score
- Import into your CRM
- Compare against closed-won accounts from two quarters
- Request a written definition of every field
Request a written definition of every field. If the vendor cannot explain how a score of 80 differs from a score of 60, the number is decoration. The NIST Privacy Framework is a useful structure for documenting how you handle the personal data inside those rows.
A score you cannot explain to your sales team is a score they will ignore within six weeks.
Answers that should end the conversation
Some replies are disqualifying on their own. Walk away when you hear them.
Disqualifying vendor replies
Vendor says it cannot share sample data for privacy reasons
walk away, data is usually thin
ask for a percentage against your domain list
- "We cannot share sample data for privacy reasons." Privacy law allows you to test a vendor's own records under an NDA. This answer usually means the data is thin.
- "Our match rate is industry leading." No number, no deal. Ask for a percentage against your own domain list.
- "The model learns automatically, so we cannot explain individual scores." Unexplainable scores cannot be defended to sales leadership or to a regulator.
What the agreement must say
Put four things in the contract. First, the minimum identity match rate, stated as a percentage, with a remedy if performance drops below it for two consecutive months. Second, the permitted uses of the data, including whether you may append it to CRM records or only view it in the vendor interface.
Third, a data processing addendum naming the categories of personal information collected and the retention period. The FTC's data security guidance sets out the expectations regulators apply to firms holding contact data. Fourth, a termination clause that lets you exit within 30 days if the match rate falls short.
How a US team compared two tools
A mid-market software firm in Ohio compared Bombora and 6sense on closed-won accounts. It chose the tool with the lower false positive rate from the sample export, despite a higher cost per seat.
That trade-off is normal. Higher coverage usually means lower precision, and precision is what your sales team feels. Our notes on B2B marketing analytics explain how to measure both without inflating the numbers.
