B2B deals close at the account level with a buying committee of 4-7 stakeholders. Scoring that treats each individual separately misses the signal. Strkr aggregates signals to the account, surfaces the committee shape, and scores the opportunity accordingly.
A B2B account typically engages Strkr with 4-7 people on the buying committee. Scoring each person independently produces seven separate scores and no deal signal. B2B scoring aggregates to the account, treating the account as the unit of analysis. Below is the aggregation model and the signals it reads.
Account aggregation
Signals roll up from contacts to account.
A contact-level signal (demo page view, whitepaper download, pricing-page visit) rolls up to the account. The account signal is the sum of contact signals, weighted by stakeholder role (champion weighted 1.5x, decision-maker 2x, user 1x). The account-level score reads the aggregated signal, not any single contact's view. Classic win: an account that scores 72 at the aggregate level has three contacts in signal; a 72 from one contact is a different qualification.
Committee shape
Four to seven stakeholders typical.
The account detail surfaces the identified buying committee: champion, economic buyer, technical buyer, user champion, blocker (if known). Each stakeholder displays their individual signals and recent engagement. The committee shape itself is a scoring input: an account with champion plus economic buyer identified scores higher than an account with only one contact identified.
ICP fit
Firmographic match to the ideal customer profile.
ICP fit scores the account against: industry (NAICS and custom tags), employee count band, revenue band, geography, tech stack from enrichment, funding stage, growth signal. Each dimension carries a configured weight. Hover shows per-dimension match: "Industry +18 (SaaS, matches), Size +12 (240 FTE, matches), Growth +8 (Series B 2026, matches)."
Intent signals
Account-level, not person-level.
Intent signals aggregate: pricing page visits across any contact at the account (sum, weighted by recency), demo page views, G2 competitor-comparison views, Bombora topic surges on the account domain. The account-level intent reads the whole buying committee's digital footprint, which is a more honest signal than a single person's browsing.
Engagement decay
Account-level inactivity.
Engagement decays when the account goes dormant: no contact-level activity in the trailing window. Half-life is configurable. An account at score 85 with no activity for 21 days decays to the mid-60s, which is accurate because the committee has gone quiet. The decay is at the account level, not the individual level, which matches the B2B motion.
Component weights
Tunable per segment and per vertical.
The three component weights (ICP fit 55 percent, intent 30 percent, engagement 15 percent is the default for mid-market B2B) are tunable per segment and per vertical. Enterprise verticals weight ICP higher; SMB-focused teams weight intent higher. Weight edits are versioned; the audit trail preserves the full history.
Band actions
Named segments drive automation.
Account score ranges carry named bands: Hot (85+), Warm (65-84), Cool (40-64), Cold (under 40). Each band has configured actions on entry. Hot opens an immediate AE and SDR notification plus the SE engagement; Warm enrolls in the nurture; Cool stays in the newsletter; Cold drops. One scoring model drives multiple downstream flows.
Hover explanation
Every account score explains itself.
Hovering the account score opens a popover with the three-component breakdown, the top five contributing signals across the committee, and the committee shape. The AE sees exactly which signals produced the number. No black box; defensible in a 1:1 and in a QBR.
What the model learns
Three calibration patterns the engine tracks.
B2B scoring models drift silently. The signals that predicted win rate last quarter may not predict next quarter. The engine measures three calibration patterns continuously and surfaces drift on the admin dashboard. Admins accept or reject recalibration suggestions explicitly.
Win rate by band
Hot should close faster than Warm.
The engine measures win rate by score band over the trailing six months. Hot leads should close at a materially higher rate than Warm. If the gap is less than 5 percentage points, calibration is suspect; the engine suggests a weight adjustment. B2B teams with good calibration see a 15-25 percentage-point gap between Hot and Warm.
Committee complete
Deals close faster with a mapped committee.
The engine measures close rate by committee completeness: 1 contact, 2 contacts, 3+ with roles mapped, 5+ with roles mapped. The data shows a nearly 2x close-rate improvement when the committee hits 5 identified roles. The score weights "committee completeness" as a signal, and the admin dashboard surfaces accounts missing the committee map.
Signal lift by segment
Pricing-page lift differs enterprise vs SMB.
The engine measures per-signal win lift separately for each segment. Pricing-page view produces a 3.4x win lift on SMB accounts and a 1.8x lift on enterprise accounts; enterprise buyers look at the pricing page as part of information gathering rather than buying intent. The model weights signals differently per segment, and the engine suggests shifts when the data diverges.
Decay rate fit
Enterprise decay should be slower.
B2B sales cycles vary by segment: SMB closes in 23 days average, mid-market in 54 days, enterprise in 112 days. The decay rate should match. The engine measures time from first-signal to close by segment and recommends decay half-lives that keep fresh accounts in Hot and age out dormant ones. Enterprise should use a 60-day half-life; SMB should use 14-21.
Firmographic drift
ICP shifts over time.
The engine tracks win rate against ICP dimensions over quarters. A team that was winning deals at 50-120 FTE in Q1 might see the win zone shift to 120-500 FTE in Q3. The ICP shift surfaces on the dashboard with a suggestion ("increase size weight on 120+ segment, decrease on sub-50 segment").
Rep acceptance
Measured by first-touch rate on Hot.
If AEs trust the scoring, they first-touch Hot accounts within the SLA at high rate. The engine measures first-touch rate on Hot specifically. A drop below 85 percent signals that the AEs are not treating Hot as Hot; the dashboard surfaces it as a calibration or trust issue, which the admin addresses in the next sales enablement cycle.
Where scoring shows up in the day
Seven surfaces that read the account score.
A score is useful only if it lands in the places reps and managers make decisions. Below are the seven specific surfaces where the B2B account score surfaces in the Strkr workflow. The surfaces matter more than the model itself.
Account list
Default sort by score, filter by band.
The account list defaults to score-descending sort. Filters support band selection, segment, vertical, and recent-activity window. The AE starts the day looking at the top of their book; the manager reviews the Hot plus Warm sections in the Monday roll-up. The surface replaces the "which accounts should I work this week" tribal-knowledge game.
Pipeline view
Score-weighted forecast.
The pipeline view includes a score-weighted forecast number alongside the raw sum of open deal amounts. The weighted number applies the account score as a confidence multiplier, which lands closer to actual close than the raw sum. CROs reporting to the board show the weighted number; the raw sum is for operations.
Routing input
High-fit accounts bypass SDR pool.
The routing engine reads the account score. A Hot account with ICP fit over 90 can bypass the SDR qualification pool and route directly to the AE pool. The pattern catches the case where a Fortune 500 inbound lead sits with an SDR because standard routing did not know the fit was high.
Sequence enrollment
Different cadences per band.
The nurture sequence enrollment reads the band. Hot accounts enroll in the AE direct-outreach cadence. Warm accounts enroll in the SDR nurture plus occasional AE touch. Cool stays in the marketing newsletter. Cold drops. One scoring model drives three cadences.
ABM campaigns
Target lists built from score bands.
Account-based marketing campaigns build target lists from score bands. The "top 100 accounts to invite to the executive dinner" is the top 100 by account score within the target vertical. The list is a filter, not a maintained spreadsheet. Changes propagate automatically.
Deal reviews
Score history as the deal story.
Deal reviews reference the account score history curve. Flat score means the committee is not progressing; climbing score means engagement is growing; dropping score means risk. Managers read the curve the way a doctor reads a chart. Specific prompts: "the score dropped two weeks ago; what happened?"
Marketing attribution
Which signals produced the Hot account.
Marketing attribution dashboards include a per-signal contribution to Hot accounts. "The whitepaper on security produced 23 Hot accounts last quarter" is a attribution; "pricing page views" is another. Marketing teams see which assets actually produce deal signal, not just traffic.
What this replaces
The three scoring models mid-market B2B teams outgrow.
B2B teams typically run one of three scoring models before migrating to a proper account-aware model: HubSpot predictive, Marketo scoring, or a homegrown Salesforce formula. Each has a failure mode in B2B context. Below is a critique with the tradeoffs named.
HubSpot predictive
Person-centric by design.
HubSpot predictive scoring is person-centric. It scores contacts individually without strong account aggregation. For B2C or product-led, the model fits; for B2B with 4-7 person committees, the model misses the account signal. Teams graduating from HubSpot predictive cite the "we have five Hot contacts at one account but the account looks lukewarm in reporting" failure mode.
Marketo scoring
Capable but opaque.
Marketo Scoring supports account-level aggregation but the configuration is dense and the hover explanation is thin. Reps struggle to defend a score to a sales manager. The explainability gap is the primary pain. Strkr's hover-explanation design was built specifically to address this.
Salesforce formula fields
Brittle and admin-dependent.
Many teams built their lead score as a Salesforce custom formula field. The approach works at the first level but breaks when the formula needs to aggregate across the account or incorporate dynamic enrichment. Changes require a certified admin. Teams migrating off Salesforce often cite the scoring brittleness as the pain.
6sense and Demandbase
ABM intent, not integrated scoring.
6sense and Demandbase supply intent signals (keyword surges, anonymous visitor identification, buying-stage prediction) but do not replace the full scoring model. Teams running these feed the intent signals into Strkr's scoring engine where they get weighted alongside ICP and engagement. The combination works; neither tool alone covers the full scoring job.
Not Zapier
Middleware does not do account aggregation.
A Zapier zap can update a lead score on a event (webhook received, form filled). It cannot aggregate across the buying committee, decay on inactivity, or weight signals by segment. B2B scoring is not a Zapier job. We do not recommend middleware for the scoring engine.
Vertical integrations
Industry data feeds via open API.
Vertical intent providers (industry-specific intent data for healthcare, construction, legal) feed into Strkr's scoring via the open webhook plus REST contract. The signals weight into the model like any other intent source. Teams in specialized verticals get vertical-specific scoring without Strkr building every vertical in-house.
Measurable outcomes from live customers
Eleven B2B teams measured scoring impact.
Eleven mid-market B2B customers measured scoring deployment outcomes against a pre-deployment baseline. The numbers are the median; variance across customers was in the +/- 20 percent range. B2B account scoring is one of the five highest-ROI changes a mid-market team ships, specifically because it changes behavior at the top of funnel where the compounding is largest.
Qualify rate
Up nine points on target accounts.
Median qualify rate on target accounts rose from 28 percent to 37 percent after scoring deployment. The lift came from AEs working the top of their book instead of first-in-first-out. The tightest customer saw a 15-point lift (fast-moving SaaS); the loosest saw a 4-point lift (slow enterprise with 110-day cycles where scoring plays less of a role).
Pipeline velocity
Hot accounts close 2.1x faster than Warm.
Hot accounts in the sample closed in 47 days median; Warm closed in 99 days. The 2.1x differential is the practical value: AEs invest the faster motion in Hot and the longer nurture in Warm. The differential held across all 11 customers with less than 15 percent variance.
ACV lift
Hot deals carry 24 percent higher ACV.
Hot accounts closed at an average 24 percent higher ACV than Warm accounts in the sample. The ICP-fit weighting lands on larger, better-fit accounts and the committee-aware aggregation surfaces the enterprise opportunities first. The lift is a mix of better qualification and better-targeted investment of AE time.
Committee mapping
From 2.1 to 4.7 contacts per deal.
The committee-shape surface changed rep behavior: AEs identify and map committee members more aggressively. Average contacts per open deal rose from 2.1 to 4.7 across the sample within one quarter. The deeper committee mapping directly correlates with higher close rate; the pattern reinforces itself.
Marketing budget efficiency
Cost per Hot account down 31 percent.
When marketing teams started reporting on Hot accounts generated per channel (rather than MQLs), budget shifted from top-funnel to mid-funnel. Cost per Hot account fell 31 percent across the sample. Marketing teams started measuring the output that matters (Hot accounts the sales team actually works) instead of the proxy (raw MQL count).
Admin cadence
Weight edits land at 1.4 per quarter.
The admin dashboard's calibration suggestions produce a 1-to-2-edit-per-quarter cadence. Teams that edit more frequently are often chasing noise. Teams that never edit are missing drift. The 1-to-2 cadence is the healthy range; the engine helps the admin land there.
See account-level scoring with a real buying committee in one meeting.
Bring your current scoring model (or lack of one). We will walk through the aggregation math and show the committee visualization on a sample account.
Can we score accounts that have only one identified contact?
Yes, with the caveat that the account score is less confident with one contact than with the full committee mapped. The engine surfaces the confidence level on the account detail. A one-contact account with high ICP fit and strong intent still scores Hot; the AE just knows the committee discovery is the next step. The engine does not inflate confidence on thin signal; the hover shows exactly which signals contributed and which committee roles remain unknown. The pattern catches the common first-touch moment when only the champion is identified.
How does committee role identification work?
Committee roles get identified three ways: explicit AE tagging (the AE marks a contact as "economic buyer" after a meeting), signal inference (a contact with a job title or seniority band maps automatically to a likely role), and interaction pattern (contacts with heavy engagement get surfaced as likely champions). The combination produces role suggestions that the AE confirms or corrects. The system does not force a role identification; the AE can leave a role as unknown. The committee map improves with AE investment but does not require it to produce a baseline score.
What is the model when the ICP is explicitly multi-segment?
Teams with multiple ICPs (horizontal product, multiple vertical motions, product-led plus sales-led) configure one scoring model per segment. The segment classification happens at account creation (firmographic rules), and each account routes to the appropriate scoring model. Each model has its own weights, bands, and thresholds. The admin surface shows the segmented models side by side with per-segment calibration metrics. The approach is more honest than a single universal model that compromises on every segment.
How do we handle accounts in the pipeline that we created before scoring was configured?
The engine backfills scoring on existing accounts on first configuration. The backfill process runs the current scoring model against the trailing 12 months of signal history for every account. Backfill typically takes an hour for a 50,000-account tenant and surfaces a progress indicator. After backfill, the full pipeline carries scores. The admin can filter the backfill to a creation date range if needed; some teams only want to score accounts created in the last 6 months to avoid noise from early data.
Does the model use Strkr AI or a transparent weighted sum?
The default B2B scoring model is a transparent weighted sum of the three components (ICP fit, intent, engagement) because transparency is more valuable than a 2-point accuracy lift from a black-box model. Teams on the Pro tier and above can enable the Strkr AI calibration layer, which runs a per-tenant model that suggests weight adjustments based on historical win data. The AI layer never changes weights silently; it suggests, the admin accepts or rejects. The underlying hover explanation remains readable regardless of whether the AI layer is active. We default to transparency because B2B reps need to defend scores to managers.
Can we export scored accounts to our ABM platform?
Yes. Account scores and component values export via the standard data warehouse connector to Snowflake, BigQuery, Redshift, or any supported destination. The export includes the three components, the top five contributing signals, the band, the committee shape, and the band entry time. ABM platforms like Demandbase and 6sense can consume the warehouse feed or receive scored accounts via webhook. The score becomes the input to downstream ABM orchestration rather than an isolated metric.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
We use cookies.
Essential cookies keep the site working. If you accept, we also enable Google Analytics
so we can see which pages help and which don't. Reject to opt out entirely. Details in our
Privacy Policy.
