Scoring accounts with intent data turns a noisy stream of signals into a ranked queue the team can actually work. The right score blends fit and intent into one number, calibrates against pipeline history, and refreshes on a cadence that matches the buying cycle. Built well, it routes attention where it converts. Built badly, it produces a leaderboard nobody trusts.
Disclosure: Abmatic AI is an account-based marketing platform, so we have a financial interest in B2B teams running structured ABM. The framework below is platform-agnostic and works regardless of whether the team's stack centres on Salesforce, HubSpot, a warehouse, 6sense, Demandbase, ZoomInfo, Clearbit, or another vendor.
See how Abmatic AI operationalises this framework, book a demo.
An intent score is a tool, not a deliverable. Before designing the formula, write down what decision the score is supposed to drive: routing, prioritisation, ad bidding, content choice, or all four. The score that drives routing is not the same as the score that drives ad bidding, and treating them as one number is the most common reason intent programmes stall.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
Intent without fit is noise. Build a fit score from firmographic and technographic data before layering intent on top. The fit score is the floor: an account that scores zero on fit will not convert no matter how strong the intent looks, so the intent layer can only re-rank within the fit-qualified universe.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
Intent data comes in three flavours: third-party (vendors like Bombora aggregate publisher consumption), first-party (the team's own website, content, and product telemetry), and partner-network (community, partner referrals, ecosystem signals). Pick a small set rather than a wide one. Three trusted sources beat ten noisy ones.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
Each intent source needs an explicit threshold below which the signal does not count. Without thresholds, the score is dominated by background noise and the team stops trusting it inside a quarter. Per Forrester research on intent data programmes, the strongest predictor of programme success is whether thresholds are written down.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
The composite score is what the team actually uses day to day. The simplest defensible formula is a weighted sum: 50 percent fit, 30 percent third-party intent, 20 percent first-party. Adjust the weights based on what the historical pipeline tells you, not based on what feels right to the loudest stakeholder.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
The score is only useful if it predicts conversion better than ICP filters alone. Pull the last 12 months of opportunities, calculate the score retroactively, and check that closed-won opportunities cluster in the high-score band. If they do not, the formula is wrong.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
A score that does not change rep behaviour is decoration. Wire the composite score into CRM routing rules, the SDR queue, the marketing automation lead-grade, and the ad audiences. The single most common failure mode is a beautifully designed score that nobody reads.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
Reps see things the score does not. Build a one-click feedback loop where reps can mark a high-score account as a false positive and a low-score account as a true positive. Roll the feedback into a monthly retraining of the weights. Without this loop the score drifts away from reality inside two quarters.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
Scores drift. Industries change, products change, and the buying committee changes. Run a quarterly audit that compares the score's predicted conversion to the actual conversion and flags drift early. The audit should also check for bias against under-served segments where the team has limited training data.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
A score is only useful if reps trust it. Publish a one-page explainer that shows what goes into the score, what threshold means what action, and how the score has performed against actual deals. Update the explainer when the formula changes. Reps who understand the score use it; reps who do not, ignore it.
The operational reading: this step is where most teams under-resource the work, because it looks like documentation rather than execution. In practice, the discipline of writing the artifact down is what allows the next step to compound. Skip the writing and the next quarter starts the conversation from zero.
The framework above sits inside a wider set of operating-model artifacts the Abmatic AI editorial library has documented. The links below cover the adjacent topics most teams reach for next, in plain English, with the same platform-agnostic stance.
The framework is informed by the public B2B research bodies that cover this space. The links below open in a new tab and point to the most useful starting pages on each.
Want to see this framework running on the Abmatic AI platform? Book a demo.
Most teams stall on a small set of recurring failure modes rather than on the framework itself. The list below names the patterns we see across B2B revenue teams in the under-500M ARR band, drawn from public customer reports and from Forrester and Gartner research on B2B operating models.
Each pitfall has the same fix: write the artifact, name the owner, set the date, and review on a fixed cadence. The framework above is the canonical reference; the pitfalls list is the recurring trap on the way to using it.
Fit scores who an account is (firmographics, technographics, ICP match). Intent scores what the account is doing (research, web activity, product engagement). Most defensible programmes combine both into one composite score, with fit acting as the floor and intent acting as the re-ranker.
Three is a good starting point: one third-party source (Bombora or similar), first-party deanonymised website behaviour, and either product or partner signal if available. More sources do not produce better scores; they produce more noise unless the team has explicit thresholds.
Quarterly, against the last 12 months of pipeline. Re-tune more often than that and the formula chases noise; re-tune less often and the score drifts away from reality. The audit and the re-tune are the same activity.
Yes, if the team has strong first-party signal and a high-traffic website with deanonymisation. First-party intent often outperforms third-party in the bottom funnel because it captures the actual behaviour the rep cares about (pricing, comparison, demo). Per Forrester research on first-party data, the bottom-funnel lift is meaningful.
The list says who the team works; the score says when. The list is firmographic and strategic; the score is behavioural and tactical. A target account with a low score still gets touched; a non-target account with a high score still does not. The score re-ranks within the list.
The shortest path from this page to a working operating model is to pick one section above, name a single owner, and ship the deliverable inside two weeks. Frameworks compound; the first artifact is the one that matters.