Account-level intent has plateaued. Buying-group intent — stakeholder-by-stakeholder signal — is the next decade of B2B targeting. Here's how to build it.
Account-level intent tells you something is happening. Buying-group intent tells you who, why and what to do about it. Stitching stakeholder-level signal across LinkedIn engagement, product usage, content consumption, employment moves and partner signals produces 4–7x the precision of account-level scores and dramatically faster pipeline acceleration.
Account intent has been the standard for a decade. Vendors compete on signal breadth and account-resolution accuracy. The problem: the average enterprise buying group is 11+ humans and the account-level signal cannot tell you which of them is moving — only that someone might be. Sellers are then asked to guess.
Buying-group intent flips the model. The unit of scoring is the stakeholder, not the account. Signals are joined at the contact level: LinkedIn role and content engagement, product-usage events, content consumption from your own owned channels, employment moves, partner signal. The output is a ranked list of named humans within named accounts.
The technique is harder to build than account-level — contact-resolution is messier, privacy considerations are higher, the operating model needs more cross-functional plumbing. The payoff is decisive: 4–7x precision uplift on 'will accept a meeting' and a sharply shorter sales cycle.
Precision uplift on 'will accept meeting' for buying-group vs account-level intent
WMA internal benchmark, 2025
Average enterprise buying-committee size
Gartner B2B Buying Report 2025
Average sales-cycle reduction when buying-group intent drives outreach
Forrester ABM Maturity 2025
The first hard problem. Use deterministic identifiers (work email, LinkedIn URL, CRM contact ID) wherever possible; only use probabilistic match where the confidence threshold is high. A single mis-resolved signal poisons the entire model for that account.
LinkedIn role + content engagement, product-usage events from your own platform, content consumption from owned channels, employment-move signal, and partner/integration signal. Adding more sources rarely helps; getting these 4–6 right always does.
Use historical opportunity data to weight each signal by its actual predictive contribution. SHAP-based importance + stability checks. The weights change by segment — score B2B-tech and B2B-finance separately.
A CFO consuming pricing content is a stronger signal than a developer doing the same. Layer stakeholder role into the score; a generic 'high-intent contact' label is fool's gold.
Output: ranked list of named humans within named accounts, with a one-line 'why this person, why now'. Push into Salesforce against the contact record. Make it impossible to ignore.
Document data sources, retention and consent. Every score must be explainable in plain English. Privacy review at quarterly cadence is non-negotiable.
Probabilistic identity resolution at low confidence
Mis-resolved signals poison the model. Set a high confidence threshold; accept lower coverage.
Treating all stakeholders equally
Layer role into the score. A CFO and a developer are not interchangeable signals.
Ignoring privacy and consent
Govern from day one. A buying-group programme that fails a privacy review is one that gets cancelled mid-quarter.
Buying-group intent is the next decade. Account-level intent has plateaued.
Identity resolution is the hard problem — solve it before the scoring problem.
Stakeholder role belongs in the score. Without it, you're back to account-level.
Delivery at the contact level in the CRM is non-negotiable.
Privacy and explainability are operating requirements, not afterthoughts.
Bring us your top problem in intelligence — we'll bring the playbook.