A VP of Sales at a mid-market B2B SaaS asked a simple question in a QBR: "when the lead score says 72, what does that actually mean?" Nobody could answer. The score was a number from the marketing automation tool configured 3 years ago and never recalibrated. Strkr lead score runs on a transparent methodology (separated fit and intent, calibrated weights, exposed drivers, action thresholds) so the number has meaning.
Lead scoring is the layer where many CRM programs lose credibility. The score was configured once, never recalibrated, and now reports a number that reps distrust. The six pains below describe the common failure modes.
Many lead scoring models produce a number without a shared understanding of what the number means. Is 72 the top 10 percent of leads? Does 72 correlate with 40 percent conversion? Nobody knows. Strkr lead score ties thresholds to action and conversion rate so a score of 72 means something specific.
Collapsing firmographic fit and behavioral intent into one score loses nuance. A high-fit low-intent lead needs nurture; a low-fit high-intent lead needs disqualification. Strkr separates fit score (ICP match) and intent score (engagement) so actions can distinguish.
Many lead scoring models re-score weekly or daily. A pricing-page view at 10am should bump the score by 10am-plus-5-minutes, not Monday morning. Strkr re-scores on every signal change in real time so the queue reflects current state.
A score without visibility into drivers is a black box. The sales manager cannot calibrate, the rep cannot act, and model drift goes undetected. Strkr exposes the top 3 drivers behind every score inline so the number is explainable.
Weights set once rarely get recalibrated. The business grew, the ICP shifted, but the model stayed static. The score correlates weakly with closed-won. Strkr recalibrates monthly against actual closed-won so the model stays current.
Score tiers often exist in reports without driving action. "MQL means score 50-plus" exists as a definition but routing and SLA do not respond to the tier. Strkr thresholds tie to specific routing, SLA, and nurture actions so the score drives behavior.
The nine cards below describe the methodology that makes a lead score meaningful. Strkr ships each as a configurable primitive.
The scoring model produces two numbers: a fit score (firmographic ICP match) and an intent score (behavioral engagement and intent-data signals). Each score ranges 0 to 100. The combined view separates "right prospect, not ready" from "wrong prospect, actively engaged" so sales actions differentiate.
Weights calibrate against actual closed-won data rather than intuition. Signals that correlate strongly with closed-won get higher weights; signals that do not discriminate get low or zero weight; signals that correlate with closed-lost get negative weights. The calibration runs monthly as new deals close.
The score refreshes on every signal change: form submission, page view, email engagement, call activity, intent-data update. The lead queue reflects current state at all times. Hot-lead windows (pricing page view just now) catch at the moment they open.
Every scored record shows the top 3 drivers behind the score inline. "Pricing page viewed in last 24 hours (plus 30 points), ICP firmographic match (plus 25 points), demo-request form submitted (plus 20 points)." The rep sees why the score is high; the manager sees how to calibrate.
Negative weights deduct for signals that predict non-conversion: personal email domain in B2B motion, invalid or missing phone, matched to no-serve geography, consent-withdrawn status, prior closed-lost with no-interest reason. The negative weights prevent score inflation on bad leads.
Signal recency decay reduces the weight of older signals. A pricing-page view from 3 months ago is informative but less predictive than one from this week. Decay functions prevent stale signals from inflating current scores. Decay curves calibrate per signal type (behavioral decays fast, firmographic stays stable).
Threshold tiers map to specific actions. 75-plus: inbound AE pod, 5-minute SLA. 50 to 75: BDR, 2-hour SLA. 25 to 50: nurture cadence. 0 to 25: minimal cadence. Negative: spam filter. The score drives behavior rather than existing as advisory.
Account-level score rolls up from the contacts under the account. Buying committee size amplifies: multiple contacts from the same account showing engagement raises the account score. The account view surfaces hot accounts that no single contact flags on their own.
A score jump above a threshold (plus 20 in one session) triggers a real-time alert to the account owner or inbound rep. The jump signals active evaluation (pricing-page view, demo-request submit). The alert captures the window of attention. Batch-scored systems miss this.
The six cards below describe how teams rebuild lead scoring on Strkr. The pattern replaces a mistrusted number with a transparent, calibrated score.
Audit the current scoring model: every signal, its weight, its last-fired date. Signals that have not fired in 90 days flag as dead. Weights set without calibration flag for review. The audit becomes the baseline for rebuild.
Split the single collapsed score into fit and intent. Firmographic signals (company size, industry, revenue, geography) go to fit score. Behavioral and intent-data signals go to intent score. Qualification data joins intent after SDR touch. The two scores show side by side.
Pull 12 months of closed-won and closed-lost deals. Run correlation analysis on each signal. Assign weights proportional to correlation strength. The intuition weights replace with data-driven weights. Expect 20 to 40 percent of signals to retire or re-weight significantly.
Configure driver display so the top 3 drivers show inline on every scored record. Reps see why the score is high. Managers see how the model is distributing weight across signal types. Trust in the score builds because the number is explainable.
Define threshold tiers tied to specific actions: 75-plus to inbound AE pod with 5-minute SLA, 50 to 75 to BDR with 2-hour SLA, 25 to 50 to nurture, 0 to 25 to minimal cadence, negative to spam filter. The thresholds drive behavior rather than existing as report categories.
Launch the rebuilt model with explicit documentation: which signals matter, which weights apply, which thresholds trigger action. Reps and managers understand the model. Trust builds through transparency rather than through authority. Future recalibrations run on the same documented methodology.
The six cards below describe outcomes that follow a well-calibrated lead score. Each outcome measures against pre-rebuild baseline.
MQL-to-SQL conversion rate typically rises 20 to 40 percent after a scoring rebuild because the MQL threshold filters for leads that actually convert. Sales rejection rate drops. The marketing-sales QBR fight about MQL quality turns into a model refinement conversation.
Real-time score refresh plus score-jump alerts catch hot-lead windows at the moment they open. Pricing-page spike alerts route to reps within minutes rather than days. Connect rate on alerted spikes typically rises 30 to 60 percent.
A trustworthy score lets reps work the queue top-down. Previously, reps cherry-picked across the queue based on rep gut. The data-driven order typically converts 15 to 25 percent better than rep-selected order because the signals aggregate across more data than any one rep sees.
Negative-signal handling and spam filtering remove 15 to 25 percent of inbound before rep touch. The rep queue cleans up. Rep time redirects to leads that have real conversion probability rather than noise.
Separating fit score from intent score tells nurture programs who to invest in. High-fit low-intent leads warrant content investment because they are the right prospects at the wrong moment. Low-fit high-intent leads do not warrant investment because they are not ICP. Nurture budget redirects to high-fit segments.
Attribution reports show which campaigns produce leads at which score ranges. A webinar that produces high-score leads at 2x the rate of content syndication is clearly worth more. Marketing budget rebalances toward channels that produce higher-score leads.
Lead scoring in Marketo, Pardot, or HubSpot is where most B2B teams start. Here is the head-to-head when the question is a transparent, calibrated, real-time score.
| Feature | Strkr | Legacy marketing automation scoring |
|---|---|---|
| Fit vs intent separation | Two separate scores | Usually one collapsed score |
| Weight calibration source | Your closed-won data | Intuition or vendor template |
| Recalibration cadence | Monthly | Set once, rarely updated |
| Real-time refresh | Every signal, within seconds | Daily or weekly batch |
| Driver transparency | Top 3 drivers inline | Black box |
| Negative signal weights | Native | Often positive-only |
| Action-tied thresholds | Routing, SLA, cadence per tier | Report categories only |
| Score-jump alerts | Native real-time | Not typically available |
| Account-level rollup | Native with committee amplification | Varies |
| Spam filter via negative scores | Native | Separate configuration |
Five rebuilds of lead scoring on Strkr. Each one replaced a mistrusted number with a transparent, calibrated system.
A B2B SaaS separated the single collapsed score into fit and intent. The analysis revealed 40 percent of inbound was high-intent low-fit (SMB prospects of an enterprise product) that should have been disqualified rather than routed. Rep time redirected to high-fit high-intent produced 32 percent more SQLs at identical headcount.
A mid-market B2B team turned on monthly recalibration against closed-won. The analysis found 7 signals with weights needing adjustment and 3 signals with reversed polarity (previously positive, actually negative). The refreshed model raised MQL-to-SQL conversion rate from 18 percent to 22 percent in the following quarter.
An enterprise B2B team ran a rep survey on trust in the lead score before and after enabling driver transparency. Pre-rebuild: 32 percent of reps said they trusted the score. Post-rebuild: 78 percent. Reps worked the queue top-down rather than cherry-picking because they could see why the score was high.
A B2B services firm turned on real-time refresh plus score-jump alerts. Prospects whose scores jumped by 20-plus points in a session triggered real-time alerts to account owners. First-touch time on these spikes dropped from 2 hours to 4 minutes median. Connect rate rose 42 percent on alerted spikes.
A B2B SaaS defined threshold tiers tied to specific routing and SLA. 75-plus to inbound AE pod with 5-minute SLA. 50 to 75 to BDR with 2-hour SLA. 25 to 50 to nurture. Routing automation replaced rep cherry-picking. Pipeline velocity on high-score leads improved 28 percent because the SLA actually enforced.
Start a 14-day trial with separated fit and intent scoring, correlation-calibrated weights, driver transparency, and action-tied thresholds enabled. Rebuild the lead score reps trust in the first 30 days.
A lead score is a numeric indicator of a lead's likelihood to convert, based on firmographic fit, behavioral engagement, intent data, and qualification data. The score drives prioritization (which leads reps work first), routing (which pool receives which lead), SLA (how fast rep must touch), and nurture (which cadence applies). When the score is calibrated and trusted, the entire inbound motion runs on it. When the score is miscalibrated, the whole motion fails.
Two scores. Fit (firmographic ICP match) and intent (behavioral engagement) measure different dimensions and warrant different actions. A high-fit low-intent lead needs patient nurture because the right prospect is just not ready. A low-fit high-intent lead needs disqualification because the wrong prospect is not worth sales investment regardless of engagement. Collapsing into one score loses the ability to distinguish, which leads to misrouted sales effort and wasted nurture spend.
Monthly recalibration is the standard cadence. Pull the last 12 months of closed-won and closed-lost deals, re-run correlation analysis against the signals, update weights proportional to current correlation. Monthly cadence catches signal drift before it degrades the score meaningfully. Static models (set once, never recalibrated) typically become unreliable within 12 to 18 months as the business, ICP, and sales motion evolve.
A score without visible drivers is a black box. Reps distrust numbers they cannot explain. Managers cannot calibrate without seeing the components. Model drift goes undetected because nobody is looking at the inputs. Exposing the top 3 drivers inline on every scored record makes the model auditable, builds rep trust, and surfaces calibration issues (e.g., "this signal is driving 60 percent of all scores, is that right?") before they become operational problems.
Batch scoring recomputes scores on a schedule (daily, hourly). A prospect who views the pricing page at 10am sees their score update at 11am (hourly batch) or Tuesday morning (daily batch). The window of active engagement passes. Real-time scoring recomputes on every signal change within seconds. The pricing-page view bumps the score immediately, triggers a score-jump alert, and routes to a rep while the prospect is still in the browser window. Real-time scoring is the difference between catching the moment and missing it.
Trust builds through transparency, consistency, and measurable outcomes. Transparency: expose the top 3 drivers inline on every record, document the methodology, publish the recalibration cadence. Consistency: ensure the same signal pattern produces the same score across reps and over time. Measurable outcomes: tie thresholds to specific actions and measure whether high-score leads actually convert at higher rates in production. When reps see the score predict their deal outcomes, trust builds within 30 to 60 days of rebuild.
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.
