How many sales conversations start like this, even when it was layered with intent data?
The data was not necessarily wrong. Someone did read the content, visit the page, or research the category. The system fails somewhere after that. Behavior becomes intent, intent becomes purchase readiness, and purchase readiness becomes an SQL.
Three inferences stacked on top of each other. Then the dashboard presents the result like a fact.
Of course, intent data works. It is one of the few ways marketing and sales teams can observe a buying journey that mostly happens without them. But it does not find sales-qualified leads on its own. It finds a disturbance in the market, a shift in attention, and asks your team to understand why it happened.
That difference is the entire strategy.
Intent data is a map of curiosity, not a buying contract
Let’s start with what third-party intent actually measures.
Bombora explains Company Surge as a change in how many people from an organization are researching a topic, how frequently they are reading, and how deeply they are consuming content compared with their normal activity. The data is mapped to a business and organized into topics.
Read that carefully.
A surge tells you that an account is behaving differently from its own baseline. It does not tell you which person is leading a purchase, whether a budget exists, or whether the organization prefers you. It does not even guarantee that the research is attached to an active buying project.
It is a probability marker.
And the companies operating these systems know the danger of treating it as more. In its 2026 guide to scoring Bombora signals, Demandbase recommends keeping third-party intent scores intentionally low because their volume can inflate account engagement and create noise. It also separates broad educational topics from competitive research because the two do not carry the same commercial meaning.
The people selling the technology are warning teams not to worship the score.
First-party intent looks stronger because it happens on your properties. Pricing-page visits, case-study downloads, webinar attendance, chat interactions, and demo requests are attached to your brand rather than a general topic.
Yet, the same problem remains.
A content download can mean the content was useful. A pricing-page visit can mean your pricing was required for a competitor analysis. A webinar attendee can be a practitioner learning about the market with no authority, timeline, or desire to buy.
The interaction is real. The conclusion is still yours to prove.
The buyer may have decided before your “high-intent” alert
Here is where the intent conversation becomes uncomfortable.
The 6sense 2025 Buyer Experience Report surveyed nearly 4,000 B2B buyers and found that the winning vendor was already on the Day One shortlist 95% of the time. Ninety-four percent of buying groups ranked their shortlist before speaking with sellers, and buyers initiated 79% of seller engagements themselves.
Yes, 6sense sells revenue intelligence. But the report’s methodology is public, and its findings reinforce something buyers have been telling the market for years: they are not waiting for your SDR to begin forming an opinion.
This creates the intent-data paradox.
By the time an account displays the cleanest signals- comparison research, pricing visits, multiple solution pages, and demo activity- the buying group may already know which vendors it trusts. Your team might be celebrating the discovery of demand after someone else has shaped it.
Then what is the point of intent data?
Not just speed. Speed has become the default answer because it gives software a clean use case: detect the signal, trigger the sequence, call within the hour.
But if the buyer has spent months building a preference, a fast email will not reverse it. Intent data has to do something more valuable. It must help you understand which problem is taking shape, which accounts are entering the conversation, and what value you can give them before asking for the meeting.
Intent should influence your content, advertising, account research, and nurture long before it becomes an excuse for sales outreach.
What is a Sales-Qualified Lead, really?
Salesforce defines an SQL as a potential customer assessed by marketing and sales, demonstrating purchase intention and meeting qualification criteria.
Fair enough.
But who confirmed the intention to purchase?
If the answer is the same behavioral score that moved the lead through marketing, the definition becomes circular. The person is ready for sales because the model says they are ready, and the model says they are ready because they behaved like someone the organization previously called sales-ready.
A feedback loop can become an echo chamber very quickly.
An SQL should be an account where sales has enough context to spend human attention, and the buyer has a plausible reason to accept that attention. Not guaranteed revenue. Not a perfect BANT form. Just a defensible business case for a conversation.
Intent data contributes to that case. It should not be the entire case.
The multi-layered intent model
The problem with lead scoring is not that it uses numbers. The problem is that one number hides the quality of the evidence underneath it.
A better model is layered. Each layer answers a different question and removes a different type of uncertainty.
Layer 1: Can this account realistically buy from you?
This is fit, not intent.
Industry, geography, company size, technology, regulation, operating model, and the scale of the problem establish whether an account belongs in your market. If the company cannot use, afford, or implement the solution, a surge in research does not magically make it qualified.
ICP fit creates the boundary. It prevents teams from confusing attention with commercial possibility.
Layer 2: What changed in its research behavior?
Now the intent signal becomes useful.
Is the account consuming broad educational content, investigating a specific problem, comparing approaches, or researching named vendors? Did the activity appear once, or has it continued? Is it recent? Is it materially different from the account’s baseline?
Build topic clusters around the problems your operation solves. A cybersecurity vendor should not treat every “cloud security” signal equally. Research into general trends is different from research into zero-trust migration, implementation costs, named competitors, and compliance requirements.
The topic tells you what the account may care about. The pattern tells you how seriously to investigate it.
Layer 3: Does the account know you exist?
Third-party intent can reveal market interest. First-party behavior shows whether some of that attention has reached your brand.
But even here, create a hierarchy.
An anonymous blog visit is not equal to repeated engagement with a relevant case study. A case-study view is not equal to a direct response. A demo request is not equal to a procurement question. Treating these actions as interchangeable is how a scoring model becomes theatre.
There is also a relationship question: If your salesperson calls, will the person know who you are and why the conversation might be useful?
If not, you may have an account worth marketing to. You do not yet have an SQL.
Layer 4: Is the problem visible across the buying group?
The same 6sense research found that typical B2B purchases involve more than ten people. Yet, most lead systems still celebrate one active contact, as if that person carries the organization’s budget, risk, and politics in one browser session.
They do not.
Look for a cluster. Is someone in operations researching efficiency while IT examines integration and finance reads about total cost? Are several people engaging with the same problem from different angles? Does the pattern reflect the shape of a possible buying committee?
This does not mean ten contacts must download ten assets. It means qualification must account for the organization, not mistake one person’s curiosity for collective agreement.
Multi-threading begins here, before sales inherits the account.
Layer 5: Why would the organization act now?
This is the layer most dashboards cannot give you.
A leadership change, compliance deadline, expansion, hiring pattern, product launch, technology migration, cost-reduction program, or public failure can create a reason to act. But a trigger is not intent either. Funding does not mean the company wants your product. Hiring does not mean the current system has failed.
Context becomes valuable when it explains the behavior.
A surge around data governance, visits to implementation content, engagement from legal and IT, and an approaching regulatory deadline tell a coherent story. Now the team has a hypothesis worth validating.
And validation requires a person.
The first outreach should not say, “We saw your account researching data governance.” That is not personalization; it is a surveillance notification.
Ask a useful question. Share a benchmark. Offer an argument about the problem. Give the buyer something that helps them think, even if they never buy from you.
Their response, or refusal, will tell you more about readiness than another ten clicks.
Not every intent account should become an SQL
This layered model creates three practical states.
- Fit plus a weak topic signal: The account deserves awareness and monitoring, not a sales sequence.
- Fit plus sustained research, first-party engagement, and buying-group activity: The account is a pre-SQL. Marketing and sales can coordinate nurture, research, and low-pressure outreach.
- The same evidence plus a validated problem, stakeholder, or initiative: Now it can become an SQL because the organization has a reason for a sales conversation.
The word validated matters.
Maybe the buyer confirms that the problem exists, but the timing is wrong. Good. That is not a failed lead; it is accurate qualification. Maybe the company has an active initiative, but your solution cannot satisfy a non-negotiable requirement. Also good. The disqualification saves both parties from a painful sale.
Qualification is not the art of forcing more records into the pipeline. It is the discipline of deciding where attention is justified.
The sales handoff should admit uncertainty.
Do not hand sales an intent score of 92 and call it context.
Tell the story:
- Which topics surged and whether they were educational, problem-led, or competitive
- Which first-party interactions occurred and how recently
- Which roles or departments appear involved
- What changed inside the organization
- What the team believes the problem might be
- What evidence contradicts that belief
- Which question sales should ask first
This gives the salesperson an intelligent opening. It also gives them permission to test the hypothesis instead of performing a pitch written by the automation system.
Then sales must return the truth. Wrong person. No project. Existing contract. Problem confirmed. Timing later. Competitor preferred. Opportunity created.
Without that feedback, the intent model only learns from clicks, and the organization keeps mistaking its own marketing activity for buyer reality.
Intent data finds the moment of change. But it’s your teams that find the SQL.
Intent data is not a crystal ball, and it is not useless. It is an observation system, a way to notice that an account’s attention has moved before the account explains why.
Used badly, it creates false confidence, contextless outreach, and another argument between sales and marketing about lead quality.
Used properly, it helps teams ask better questions. What is changing? Who cares? Why now? Do they know us? Is there a real problem? Have we earned the right to begin a conversation?
The dashboard cannot answer all of these.
That is not a limitation you need to automate away. It is the work.
Sales-qualified leads are not hiding inside the intent score. They emerge when market behaviour, account context, and human conversation finally agree.
