How to Use Intent Data to Identify Sales Qualified Leads

Introduction

Your reps are working a list of accounts that someone, somewhere, flagged as in-market. Some of those accounts are genuinely evaluating a solution this quarter. Most are not. The difference between a team that gets pipeline out of intent data and a team that gets a busy dashboard comes down to what happens between the signal firing and the call being made.

That gap is a qualification problem, and it is solvable. Having run outbound programs for 2,000+ B2B brands across 50+ verticals, we have watched plenty of intent programs produce activity without producing sales-qualified leads, and the pattern behind the failures is consistent enough to design around. Getting the basics of sales leads right matters more here than the sophistication of the data source, because a signal you cannot validate, attribute to a person, and act on inside its useful window is not a lead at all.

What follows is the qualification bar itself: which signals carry weight, how to filter the ones that do not, how to turn an account-level surge into a named buyer, and where to draw the line that makes something an SQL.

From Intent Data to Sales Qualified Leads: The Short Answer

1. Intent data identifies sales qualified leads only when it is combined with ideal customer profile fit, person-level resolution, and a recency window, because intent alone measures curiosity rather than buying authority.
2. Weight signals by what they prove: first-party behavior on commercial pages outranks review-site comparison activity, which outranks third-party topic surges, which outranks general content consumption.
3. Require corroboration before outreach, meaning two independent signals from different sources rather than a single spike from one provider.
4. Resolve the surging account to at least two engaged roles inside the buying group, weighted by function and seniority, before treating it as sales-ready.
5. Set an explicit numeric threshold at which a scored account becomes an SQL, publish the disqualifiers that void it, and route anything below the line to nurture rather than to a rep.
6. Validate the whole system quarterly against a holdout, comparing SQL rate and cycle length for flagged versus unflagged accounts.

What Changed in 2026

• Buyers now research with AI, and they check its work. The TrustRadius B2B Buying Disconnect Report found 63% used AI during their purchase journey and 94% of those buyers fact-check what it tells them. Research activity has moved into channels no intent vendor observes directly.
• Trust in online sources fell again. The same report tracked buyers who trust online resources less than before, rising from 39% to 47% in a single year, with the neutral middle shrinking. Signals that depend on content engagement are being generated by a more skeptical, faster-moving audience.
• Selling time recovered slightly. Salesforce’s State of Sales surveyed 4,050 sales professionals in August and September 2025 and found the average seller spends 40% of the week actually selling. The remaining 60% is where a bad intent list does its damage.
• Review-site intent consolidated. HG Insights completed its acquisition of TrustRadius in June 2025, bringing review-site behavioral data and technographic intelligence under one owner and reducing the number of independent sources available for corroboration.
• Rep-free preference held. Gartner’s 2025 buying research found 75% of B2B buyers prefer a purchase experience without a sales rep, which means the signals you can detect are increasingly the only invitation you will get.

Terms Worth Knowing

• Intent data is behavioral information suggesting an account or person is researching a problem your product solves.
• First-party intent is activity on properties you own, such as your website, documentation, product, or email.
• Third-party intent is aggregated research activity observed across external publisher networks, usually reported at account level as a topic surge against that account’s historical baseline.
• Topic surge is a measurable spike in an account’s research activity on a defined subject relative to its own baseline, rather than an absolute volume of activity.
• Person-level resolution is the step that identifies which individual at a surging account is doing the researching, or at minimum which roles are engaged.
• Signal decay is the loss of predictive value a signal suffers as time passes after it fires.
• Sales qualified lead (SQL) is a lead that has been vetted and is interested in a next step with sales, sitting one stage past MQL and one stage before a booked meeting.
• Corroboration is confirmation of a signal by a second, independent source before the account is treated as in-market.

How this was built: we reviewed the leading published research on B2B buying behavior and intent programs, compared it against practitioner discussion in sales and marketing communities, and interpreted the findings through our experience running outbound qualification for B2B clients. 

Why: teams keep buying signals and getting activity, and the missing piece is usually a stated threshold rather than better data.

What Intent Data Actually Tells You About a Sales Qualified Lead

Intent data tells you that behavior consistent with research occurred at an account. It does not tell you who acted, why, whether they can buy, or whether they are already committed elsewhere. Every reliable qualification system starts by naming that boundary, because the cost of ignoring it is a rep calling a stranger who has no idea why the phone is ringing.

The scale of the problem is structural rather than tactical. Research by the Ehrenberg-Bass Institute for the LinkedIn B2B Institute found that in categories where companies replace a provider roughly every five years, only about 5% of business buyers are in the market in a given quarter and around 20% across a full year. Your intent tool is looking for a thin slice of genuine demand inside a much larger population of accounts doing something else entirely.

What intent data is good at

Timing and topic. When intent data works, it narrows a large addressable market to a shortlist of accounts whose behavior has changed recently, and it tells you roughly what they were looking at. That is genuinely useful, and it is a better starting point than firmographic filters alone.

What intent data cannot do

It cannot establish authority, budget, or the absence of a competing contract. It cannot tell you that the person generating the signal is the person who will sign. And it cannot distinguish evaluation from idle interest, because both look identical in a page-view log.

Why a signal is not yet a lead

Martal’s lead taxonomy is deliberately strict on this point: a prospect is someone contacted or engaged, an MQL has responded and matches the ideal customer profile, and an SQL is interested in a next step. An intent signal sits upstream of all three, and the difference between an MQL and an SQL is the one definition worth settling before you touch intent data at all. A signal is an argument for spending attention on an account, and treating it as anything more inflates your pipeline with accounts that were never in a buying cycle.

Teams that consistently convert signals into qualified leads treat validation as staffed work rather than an automated step, whether that sits with an in-house SDR function or with an outsourced sales partner running the qualification layer. Someone has to look at the account, confirm the story holds, and decide. A scoring model narrows the queue down to the accounts worth that attention.

First-Party vs Third-Party Intent Data: Which One Identifies an SQL?

First-party intent identifies sales-qualified leads. Third-party intent identifies accounts worth investigating. First-party behavior happens on property you control, so you know the account, usually the person, and exactly what they looked at. Third-party data infers research from external networks, arrives at account level, and names a topic rather than a buyer.

• First-party intent
  • Where it comes from: Your website, product, docs, email, CRM.
  • Resolution: Account, and often the individual.
  • What it proves: Interest in your solution.
  • Latency: Real time.
  • Coverage: Only buyers who already found you.
  • Cost: Effectively free once instrumented.
  • Weight in an SQL decision: Decisive.
  • Main failure mode: Blind to buyers still researching anonymously.
• Third-party intent
  • Where it comes from: Publisher networks, content co-ops, ad exchanges.
  • Resolution: Account only.
  • What it proves: Interest in your category.
  • Latency: Days to weeks, reported as a surge above baseline.
  • Coverage: Accounts that have never heard of you.
  • Cost: Five to six figures a year.
  • Weight in an SQL decision: Supporting.
  • Main failure mode: False positives and misattribution.

The tradeoff is coverage against certainty. First-party data is highly reliable but arrives late, since a buyer on your pricing page has usually built a shortlist already. Third-party data arrives early, with a much higher error rate. Neither one qualifies a lead alone, which is why the corroboration rule below asks for one of each, and why sequence matters: third-party data decides where to look, first-party and second-party evidence decide whether to reach out.

Second-party intent sits between the two

Second-party intent is activity on someone else’s property that concerns you specifically, most commonly a buyer viewing your listing, your category, or a comparison involving you on a review platform. It carries the account-level resolution of third-party data with specificity closer to first-party behavior. The distinction is worth keeping straight, because review-site activity is routinely filed as generic third-party intent, and that understates how close those buyers are to a decision.

How to Collect Intent Data

Start with the signals you already own, add review-site activity if buyers research your category there, and evaluate third-party topic feeds last. Most teams work in the opposite direction: sign a substantial contract first, then find that the signals they needed were sitting in their own analytics.

First-party signals you can capture this quarter

• Event tracking on commercially meaningful pages: pricing, security and compliance documentation, integrations, implementation guides
• Repeat-visit and session-depth tracking, rather than raw pageview counts
• Company-level identification of traffic on those pages
• Demo, trial, quote and contact requests, timestamped and routed the moment they land
• Email and sequence engagement rolled up to the account, not just the individual
• Product usage patterns where a trial or free tier exists

An intent vendor is not a prerequisite for any of it. What it takes is someone deciding which pages carry commercial meaning and instrumenting them properly.

Review-site and second-party signals

Claim and complete your profiles on the review platforms your category uses, since the intent data those platforms sell back to you only works if buyers find you there in the first place. TrustRadius research puts peer reviews among the resources buyers rate most heavily at selection time, which makes the same activity useful in both directions.

Third-party topic feeds

Define a topic set narrow enough to mean something before you shop for a provider. A list of 10 to 25 topics tied to the problems you solve produces a usable signal; a list of 200 produces noise you will pay to store. Ask any provider two questions in writing: how the data is collected, and at what resolution it is reported. Vague answers on collection method usually indicate bidstream sourcing, which carries both the weakest compliance basis and the highest misattribution rate.

A sequence that avoids wasted spend

1. Instrument first-party capture and route alerts to a person, not a dashboard
2. Add review-site and second-party signals
3. Add trigger-event monitoring: champion job changes, funding, technology added or removed, relevant executive hires
4. Only then test third-party topic data, inside a defined window with a measurement plan attached

One pattern shows up repeatedly in outbound programs: the accounts that eventually convert were usually visible in first-party or trigger data before any third-party surge appeared for them.

The Four Signal Classes, Ranked by Qualification Weight

Not every signal deserves the same response, and the fastest improvement most teams can make is to stop treating them as interchangeable. Rank signals by what they prove about buying intent for your solution, then attach a different response to each tier.

• Class 1. First-party commercial behavior
  • Examples: Repeat pricing-page visits, security or compliance doc views, integration pages, demo or trial requests, quote requests.
  • What it proves: Interest in your specific solution, late-stage evaluation.
  • Weight: Highest
  • Response window: Same day
• Class 2. Comparison and review activity
  • Examples: Category and alternatives pages on review sites, competitor comparison views, RFP-style research.
  • What it proves: Active vendor evaluation; you may or may not be on the list.
  • Weight: High
  • Response window: 24–48 hours
• Class 3. Trigger and relationship events
  • Examples: A past champion changing employer, new executive hire in a relevant function, funding round, technology added or removed, hiring for roles that imply your problem.
  • What it proves: A reason for a conversation exists; timing is inferred rather than observed.
  • Weight: Medium to high
  • Response window: 3–5 days
• Class 4. Third-party topic surge
  • Examples: Account-level spike in category research across publisher networks, keyword-level topic consumption.
  • What it proves: Someone at the account is researching the category.
  • Weight: Lowest on its own
  • Response window: Nurture, then corroborate

First-party behavior earns the most trust

A visit to your pricing page from an account matching your ideal customer profile is the highest-confidence signal available to most teams, and it costs nothing beyond the analytics you already run. Repeat visits inside a short window matter more than a single session, and depth matters more than volume. Someone reading your security documentation is doing procurement work.

Review-site activity shows a live evaluation

Comparison behavior on review platforms indicates a buyer building or narrowing a shortlist. The signal is valuable precisely because it is late-stage, and it is worth noting where those buyers place their trust: the TrustRadius 2026 report found that buyers rank demos, free trials, prior experience, and peer reviews above other resources when they choose a vendor. An account comparing options is close to a decision that will be made largely without you.

Trigger events create a reason rather than a moment

Job changes are the signal practitioners rate most highly and buy least often, mostly because they require relationship data rather than a subscription. A champion who used your product at their last company arrives at a new one with an opinion and a budget cycle. That is a warmer starting point than a topic surge, and the outreach writes itself without referencing any tracking.

Third-party surges belong in triage

Topic-level intent widens your view beyond people who already know you exist, which is its real value. Weighed against stronger buyer intent signals, it works as a filter on where to look rather than a trigger to act. Treated as a standalone trigger, it produces the calls that damage credibility. Use it to decide which accounts get attention and enrichment, then require a second signal before a rep dials.

This is also where a consolidated data layer earns its place. Martal’s Agentic AI Platform tracks 10M+ intent signals against 300M+ verified contacts and 24M+ company accounts, which matters less for the volume than for what it removes: the gap between an account showing interest and a verified person to contact at that account. Choosing between sources is a separate exercise, and an honest comparison of intent data providers depends on which coverage gap you are trying to close rather than which vendor scores highest overall.

The Validation Layer: Why Most Signals Never Become Opportunities

Validation means checking, before a rep spends time on it, that a signal came from a plausible buyer at the account it was attributed to. It is the step that decides whether an intent program produces pipeline or activity. The practitioner consensus in sales and marketing communities is blunt: most purchased triggers get treated as unusable, and the complaints repeat with enough consistency to build a checklist from.

The cost of skipping validation is measured in rep hours. Salesforce’s State of Sales, 2026 edition, found the average seller spends 40% of the week on actual selling. Every hour spent on an account that was never evaluating anything is drawn from that 40%, not from the administrative remainder.

Who else triggers your signals

Buyers are not the only people researching your category. The recurring false-positive sources:

• Job seekers and current employees researching a company or a category before an interview
• Students and analysts doing coursework or market research
• Competitors reading your pricing and feature pages, often repeatedly
• Consultants and agencies researching on behalf of a client who may not match your profile
• Existing customers looking for documentation, which reads as fresh interest in an account-level report
• Automated crawlers inflating page-view counts

None of these are exotic edge cases. Together they explain most of the “they had no idea what I was talking about” calls that erode a team’s faith in the data.

Attribution has grown less reliable

Account-level third-party intent generally depends on resolving network activity back to an organization. Distributed work, VPNs, residential IP ranges, and shared cloud infrastructure have all degraded that resolution. The practical consequence is that a portion of your surging accounts are not the accounts named in the report, and no amount of downstream scoring can repair a misattributed input.

Single-source dependency hides the error

If one provider is your only view of the market, you have no way to distinguish a real surge from an artifact of that provider’s panel. Two independent sources agreeing is meaningfully stronger evidence than one source being emphatic. This is the practical argument for corroboration, and it is also the reason the June 2025 consolidation of review-site data under a single owner is worth noting: fewer independent sources means fewer opportunities to cross-check.

The corroboration rule

Require two signals from different sources, at least one of which is first-party, before an account is eligible for direct outreach. Teams without the capacity to run that check themselves often hand it over, since qualifying inbound leads against fit criteria is the same screening work applied to a different intake. A topic surge plus a website visit is a lead worth working. A topic surge plus another topic surge from the same provider is one observation reported twice.

What Practitioners Report About Intent Data in Practice

Community discussion on this topic is unusually consistent, and it converges on execution failures rather than on data quality alone. Four themes recur across B2B marketing and sales forums, and each one maps to a decision in the framework above.

Users in Reddit and community discussions often ask how to filter intent data down to real buyers without throwing away the signals that matter. The prevailing answer is to treat a signal as unusable until it can be tied to observable behavior at a known account, which is the corroboration rule stated in operational terms.

• Account-level reports are treated as too blunt to act on. Contributors to threads on filtering noisy intent data describe company-level surge reports as difficult to convert into a specific person worth calling, which is the gap the person-resolution step exists to close.
• Unpurchased signals outperform purchased ones. Discussions of underrated intent signals surface champion job changes, technology added or removed, competitor review activity, and new executive hires far more often than topic surges.
• A few minutes of manual account checking pays for itself. Practitioners in threads on what is working for account research describe verifying fit, funding, and technology environment before outreach, and report reply-rate improvements that justify the time. That is the argument for making ICP fit a floor condition rather than a scored input.
• Referencing the data irritates buyers. Threads on which signals actually work and Quora discussions on applying B2B intent data converge on using a signal to set timing and subject matter while keeping its source out of the message entirely.

Treat all of this as directional evidence about where programs break in practice, and look to published research for the benchmarks. Community threads report experience rather than measurement.

From Account Surge to a Named Buyer

Turn an account-level surge into a named buyer by mapping the functions that must be involved in a purchase like yours, enriching only those roles, then requiring two of them to show engagement before the account counts as sales-ready. This is the most common practitioner complaint about intent data: knowing that someone at a company is researching a category is too blunt to act on, because sales rarely extracts the specific person from it.

The structural reason is committee size. Gartner’s research on the B2B buying journey puts a complex purchase at six to ten decision makers, each arriving with four or five pieces of independently gathered information, and finds buyers spend only 17% of their total purchase time meeting suppliers, with as little as 5% or 6% going to any single vendor. A surging account is therefore several people forming separate opinions, most of it happening where you cannot see it.

Resolve to roles before you resolve to names

Start with the functions that would have to be involved in a purchase like yours: the operational owner who feels the problem, the technical evaluator, the budget holder, and the risk reviewer. That map is stable across accounts and lets you enrich deliberately rather than pulling every contact at a 400-person company.

Weight by function and seniority

A signal traced to a relevant function carries far more qualification weight than one traced to an unrelated department. A director-level or above title in the operational owner’s function is a strong indicator. Activity from a function with no plausible role in the decision is a reason to lower the account’s score rather than raise it.

Require two engaged roles

One engaged contact is a person with curiosity. Two engaged roles from different functions suggest an internal conversation is happening, which is the observable proxy for a real evaluation. Making two roles a condition of SQL status filters out a large share of single-researcher noise at no cost.

Multi-thread from the start

Once two roles are engaged, approach them differently rather than sending one message twice. The operational owner cares about the problem and the workflow. The executive cares about the business case and the risk. Running both conversations in a coordinated omnichannel sequence is more effective than routing everything through whoever answered first.

The SQL Threshold: A Scorecard for Turning Signals Into a Decision

A signal becomes a sales qualified lead at a stated number, not at a feeling. The scorecard below scores an intent-sourced account across five inputs and sets an explicit line where it becomes an SQL candidate, with floor conditions that no total can override. It works as an admission test for intent-sourced accounts, and it deliberately leaves out the general lead-scoring mechanics that belong to your broader prioritization model.

Score each account across five inputs. The weights reflect what each input proves about a buying decision rather than how easy it is to measure. The first input carries the most weight, which is why a documented ideal customer profile has to exist before the rest of the model means anything.

• ICP fit
  • Question it answers: Could this account buy from us at all?
  • 0 points: Outside profile on size, geography, or category.
  • 1–2 points: Adjacent to profile, one criterion off.
  • 3–4 points: Clean match on size, sector, geography, and technology environment.
• Signal class
  • Question it answers: What does the behavior prove?
  • 0 points: Class 4 only (topic surge).
  • 1–2 points: Class 3 (trigger event) or Class 2 (comparison activity).
  • 3–4 points: Class 1 first-party commercial behavior.
• Person resolution
  • Question it answers: Do we know who, or at least which roles?
  • 0 points: Account only, no contacts identified.
  • 1–2 points: One relevant role engaged.
  • 3–4 points: Two or more relevant roles engaged across functions.
• Recency
  • Question it answers: Is the window still open?
  • 0 points: Older than 30 days.
  • 1–2 points: 8–30 days.
  • 3–4 points: Inside 7 days.
• Corroboration
  • Question it answers: Does a second source agree?
  • 0 points: Single source.
  • 1–2 points: Two sources, both third-party.
  • 3–4 points: Two sources including first-party behavior.

Where the line sits. An account scoring 12 or above out of 20, with at least 2 points on ICP fit and at least 1 point on person resolution, is an SQL candidate and goes to a rep. The two floor conditions matter more than the total: a 14 built entirely from signal strength and recency, with nobody identified and a shaky fit, is the exact profile of the calls that go badly.

Accounts scoring 7 to 11 go to nurture with an enrichment task attached, because most of them are one missing input away from being workable. Below 7, leave them alone and let the signal build.

Automatic disqualifiers. These void an account regardless of score:

1. A known contract with a competitor with more than six months remaining
2. Signal traced solely to a function with no role in the purchase
3. Account already in an active sequence, to prevent duplicate outreach across teams
4. Firmographics outside your serviceable market, including geography you cannot support
5. A prior explicit request not to be contacted

Publish the threshold and the disqualifiers where both marketing and sales can see them. Most arguments about lead quality are arguments about an undocumented definition, and writing the number down converts a recurring dispute into a calibration exercise. Once an account clears the bar, the conversation moves to the substance of the opportunity, which is where standard criteria and a documented approach to how to qualify sales leads take over. Deciding how to prioritize sales leads once they clear the bar is a separate question with its own mechanics.

Treat the scorecard as a starting calibration. Run it for a quarter, compare the accounts that closed against their scores, and move the weights toward whichever input predicted revenue best in your market.

The MQL-to-SQL Handoff: What Moves With the Lead

An intent-sourced account becomes a sales qualified lead when it clears the threshold above, and someone at that account has confirmed interest in a next step. The handoff is the moment that judgment transfers between teams, and it needs three things agreed in advance: the evidence that justifies it, how long the receiving side has to pick it up, and what context travels with it.

The evidence that justifies a handoff

Qualification rests on authority and need, confirmed rather than assumed. Intent data adds three specific inputs to that confirmation: which signal fired, which roles engaged, and when. An MQL has responded and matches your ideal customer profile. It becomes an SQL when a conversation establishes that the need is real and the person can act on it, or bring in whoever can.

How long the receiving side has

Set a queue clock separate from your outbound response window. The outbound window governs how fast you react to a signal; the queue clock governs how long a qualified lead can sit unworked before it goes back to nurture. Harvard Business Review’s audit of first-response times found qualification odds falling away within the hour, and a lead that waits three days in an unattended queue has lost most of what qualified it.

What travels with the lead

A score alone gets a lead worked as a cold call. Send the story with it:

• The triggering signal, its class, and its date
• Which roles engaged, and what each of them looked at
• The basis for the ICP-fit judgment
• Which disqualifiers were checked and cleared
• The last touch, the channel, and what was said

What happens when sales rejects it

Rejection is the most useful data in the system, and most teams discard it. Require a reason code on every rejected SQL, recycle the account to nurture with its signal history intact, and review the reason codes monthly. Patterns in rejections tell you which input in the scorecard is miscalibrated, which is far faster than waiting two quarters for closed-won data to tell you the same thing.

How Long an Intent Signal Stays Qualified

Signals lose value quickly, and the loss is steepest at the top of the hierarchy. A pricing-page visit reflects a decision being made this week. A topic surge reflects research that may run for a month. Matching your response speed to the signal class is more useful than a blanket “act fast” rule.

The foundational evidence on response speed remains Harvard Business Review’s March 2011 audit of 2,241 US companies, which found that firms attempting contact within an hour of a query were nearly seven times likelier to qualify the lead than those trying an hour later, and more than 60 times likelier than those waiting a day or more. Average first response across the sample ran to 42 hours. The multipliers are frequently misattributed elsewhere, and the underlying finding has held up: qualification odds fall away sharply with delay.

Response windows by class

• Class 1, first-party commercial behavior: same business day, ideally within the hour for a demo or quote request
• Class 2, comparison activity: 24 to 48 hours, while the shortlist is still being assembled
• Class 3, trigger events: three to five days, since the window is weeks rather than hours
• Class 4, topic surge: no direct outreach until a second signal appears

When the window closes

An expired signal is not a dead account. Drop it back to nurture, keep the enrichment you gathered, and set the account to alert on its next signal. Accounts that surge, go quiet, and surge again are often mid-evaluation with an internal delay, and the second surge is usually a stronger buying indicator than the first.

Build for the window you can actually staff

A one-hour response commitment that nobody can honor on a Friday afternoon produces worse outcomes than an honest four-hour commitment, because the queue silently ages while everyone assumes it is handled. Set the window your coverage supports, then measure against it.

Writing to the Signal Without Sounding Like Surveillance

Reference the problem, never the tracking. The most consistent warning in practitioner discussions is that opening with an observation about someone’s browsing behavior reads as surveillance and gets messages blocked. Intent data should shape what you say and when you say it, and stay invisible in the message itself.

The reference test

Before sending, ask whether the recipient could reasonably wonder how you knew that. If the answer is yes, rewrite. “I noticed you were looking at our pricing” fails. A message about the specific problem your product solves for their role, sent the day they were looking at pricing, passes and performs better.

Match the ask to the stage

• Category or how-to research
  • What it implies: Problem awareness, no shortlist yet.
  • Appropriate ask: Share something genuinely useful, no meeting request.
• Comparison or alternatives activity
  • What it implies: Shortlist being built.
  • Appropriate ask: Offer specific differentiation and a relevant customer example.
• Pricing, security, or integration docs
  • What it implies: Late-stage validation.
  • Appropriate ask: Direct offer of a conversation about scope, terms, and implementation.
• Champion job change
  • What it implies: New context, existing preference.
  • Appropriate ask: Reconnect on the relationship, not the product.

Treating every signal as a demo request is the error that makes intent-based outreach feel indiscriminate. An account three weeks from a decision and an account three months out need different opening moves, and the signal class tells you which you are looking at. From there, the work of moving a qualified conversation forward is standard pipeline discipline, and how to convert leads to sales follows its own playbook.

Proving Intent-Sourced SQLs Are Actually Better

Prove it by comparing intent-flagged accounts against a matched holdout on SQL rate, meeting rate, and cycle length across a single quarter. Very few teams run that comparison, which is why so many renew a subscription on the strength of a dashboard.

Four metrics that answer the question

• Activation rate: the share of signals that produced a human action. A low rate means your problem is operational rather than a data-quality issue.
• Signal-to-SQL rate: the share of flagged accounts that clear your threshold and become SQLs. This is the number that tells you whether the scorecard is calibrated.
• SQL-to-meeting rate, flagged versus unflagged: the core comparison. If intent-flagged accounts do not book at a higher rate, the signal is not adding information.
• Cycle length, flagged versus unflagged: intent-sourced deals should close faster, since you are arriving during an active evaluation.

A holdout you can run this quarter

Hold back a matched sample of intent-flagged accounts from the intent-driven treatment and work them through your normal motion. Work the rest with signal-informed timing and messaging. Compare SQL rate, meeting rate, and cycle length at the end of the quarter. Keep the samples matched on firmographics so you are testing the signal rather than the segment.

The pattern worth aiming for is quality over volume. In one confidential engagement with a US artificial intelligence and machine learning company, a 13-month program converted 362 leads into 153 SQLs, an SQL rate near 42%. Rates in that range come from disciplined admission criteria rather than from a wider net, which is the same principle the scorecard encodes. Explore the AI and machine learning use case.

When to walk away from a source

If a data source has been in production for two quarters and the accounts it flags do not convert better than the accounts it does not, stop paying for it and reallocate the budget to enrichment or coverage. A source that adds no lift is not neutral, because it also consumes the rep attention that produces qualified opportunities elsewhere. This is worth checking before you expand a program, since the constraint is rarely how to generate sales leads and usually which of them deserve a call first.

Compliance and Data Provenance

Ask any prospective intent source how its data was collected, and get the answer in writing. Provenance is a commercial risk question as much as a legal one, because a source with a weak basis for collection is also usually a source with weak accuracy.

Two practical points. First, collection method matters: consented publisher networks and first-party analytics rest on firmer ground than data inferred from advertising exchange traffic, which has drawn sustained regulatory scrutiny in Europe. Second, a vendor’s compliance posture does not transfer to you. You still need a lawful basis for storing and acting on prospect data in your own systems, and that obligation sits with your organization.

Channel rules also differ by market, and the differences are not cosmetic. Outreach into the EU, the UK, and Canada operates under tighter constraints than outreach into the US, which changes how a validated signal can be actioned rather than whether it can be. Martal’s programs run GDPR compliant, SOC II certified, and CAN-SPAM and CASL aligned, with channel selection set by the market being targeted.

Where This Leaves You

Intent data earns its cost at the point where a signal becomes a decision, and that decision needs a written bar. Rank your signals by what they prove, require a second source before anyone dials, resolve the account to at least two relevant roles, set a numeric threshold with disqualifiers attached, and check the whole thing against a holdout each quarter. Those five habits are what turn a signal feed into sales-qualified leads, and each one is cheaper to build than the rep hours it protects.

If you would rather have that qualification layer run for you than build it, our team handles signal validation, enrichment, and outbound qualification as part of a managed program, with SQLs and booked meetings as the deliverables. Book a consultation, and we will walk through your current signal sources and where the qualification bar should sit.