A lead scoring model that nobody can read becomes a black box the sales team eventually ignores. A model that the operations lead can read, tune, and audit stays useful for years. Strkr runs lead scoring as a transparent, weighted composite over a named signal catalog with tunable weights, nightly recompute, trigger-driven intra-day re-scores, and cited drivers on every record. The model is portable through the open API and auditable through the change log. Fifty dollars per seat flat, no per-score meter.
A model that lives for more than two quarters needs structure. Strkr groups the signal catalog into seven named branches with a named owner per branch. The structure is visible on the admin page and the branches are the vocabulary the operations lead and the sales team use to talk about the model.
The firmographic branch reads the industry taxonomy, company size bands, geography scope, and tech stack signals. The ICP definition lives here. The common failure of a model that cannot distinguish ICP fit from behavior gets avoided because the branch is named and owned.
The behavioral branch reads product usage events, campaign touches, documentation reads, pricing page visits, and demo request forms. Each event is a cited field with a timestamp. The sales team sees the events behind the score, not a mood number.
The engagement branch reads reply rate on outbound touches, meeting booking rate, and response latency to rep touches. Each metric runs on a rolling window. The common failure of over-rewarding stale engagement gets avoided.
The champion branch reads the role profile of the engaging contacts. VP-plus titles weight differently from manager titles. The role mapping is tunable per segment and the model cites the engaging contacts on the record page.
The timing branch reads trigger events (funding announcement, exec hire, product launch, M&A) and known budget cycle timing. Each trigger fires an intra-day re-score. The operations lead configures the trigger list on the admin page.
The custom branch reads the per-tenant custom fields the standard model does not know about by default (industry subcategory, regulatory tier, strategic account flag). Weights are configured per tenant without a data science engagement.
The historical branch reads the pattern of your closed-won deals at the same stage: firmographic distribution, engagement velocity, champion role mix. The model weights signals that correlate with your actual wins rather than a generic benchmark.
The weights are the operations lead's primary control over the model. Strkr starts the weights from the historical closed-won pattern and lets the operations lead tune from there. The weights are visible on the admin page, the counterfactual shows the impact of a tune before committing, and the audit log records every change.
The initial weight set is pulled from the historical closed-won pattern across the pipeline. The weights are not vendor defaults; they are the data pattern for your business. The operations lead reviews and sets the initial model.
Tenants with multiple ideal customer profiles run multiple weight sets. The admin page handles the splits. The common failure of forcing one weight set across segments that behave differently gets avoided.
Before committing a weight change, the operations lead sees the counterfactual impact: how many leads move up, how many move down, how the hand-off threshold would shift. The tune is a decision, not a guess.
The admin page runs the new weight set against a shadow set of historical leads before promoting to live. The operations lead sees the shadow-set score distribution and can roll back before the sales team sees the live change.
Every weight change is recorded in the audit log with the user, the timestamp, the before-and-after values, and the business rationale the operations lead typed. The common failure of a weight set that nobody can explain gets avoided.
Any weight set can roll back to the previous version with one click. The common failure of a weight tune that produced an unexpected pipeline shift and then became hard to reverse gets avoided.
The model produces a composite score from zero to one hundred within the segment. The score is a weighted sum of the branch contributions. The top five drivers behind the score and the counterfactual that shows what would move it up live on the record page. The number without the drivers is useless; the drivers without the number flatten prioritization. Both together drive the queue.
The composite score is the sum of each branch weight times its branch hit value. The math is visible on the admin page. The operations lead sees exactly how the number is produced. The common failure of a scoring tool that returns a number nobody can derive gets avoided.
The composite score is scoped to the segment, not an unscoped global ranking. A lead scored eighty-five sits in the top band for its segment, not against unrelated verticals. Account executives can compare like-for-like within their book.
The record page shows the top five drivers behind the current score with the cited records. The sales executive sees the signals that drove the ranking without clicking into a separate tool. The number shows its work.
The counterfactual panel shows what would move the score up. Example: an engaged VP of engineering would move the lead from 72 to 85. The sales executive works with context rather than a bare number.
The record page shows the score trajectory over the last ninety days with the events that moved it. A record that climbed from forty to seventy looks different from a record flat at sixty for a year.
Every score shows the model version and the last calibration date. Tenants on an older calibration can see it. The sales operations lead schedules recalibrations as the market shifts.
A model that recomputes once a quarter is a model that drifts out of alignment with the pipeline. Strkr runs the composite on a nightly recompute across every lead and every open opportunity, and fires intra-day re-scores on explicit triggers so records stay current. The last recompute timestamp lives on the record page.
The model recomputes nightly across every lead and every open opportunity. The score lands on the record by Monday morning for the rep to work from. The record is current when the sales team opens it.
Intra-day re-scores fire on explicit triggers: new behavioral event, new campaign touch, new trigger event (funding, M&A, exec hire). Recently-moved records show the updated score within minutes rather than waiting for the nightly run.
When a trigger event fires on a dormant record, the resurface workflow lifts the record out of the cold list and into the sales development queue with the resurface context cited on the record page. The sales team sees why the lead is back.
The last recompute timestamp lives on the record page. The sales team sees whether the score is fresh. The common failure of acting on a stale score because nobody noticed the model had not run gets avoided.
Every quarter, the recalibration pulls the latest closed-won pattern and recommends weight tunes. The operations lead reviews, approves, and promotes. The model stays current with the actual pipeline shape rather than drifting for a year.
Every week, the drift report shows the score distribution across the pipeline versus the segment benchmark. Shifts outside the expected range flag for review. The operations lead sees drift before the sales team notices.
The model does not live in a dashboard the sales team never opens. It lives on the record page, the daily working list, the sales development queue, and the manager rollup. The common failure of a scoring tool whose output never reaches the working surface gets closed off.
The record page shows the current score, the top drivers, the counterfactual, the score trajectory, and the model version. The sales executive reads the full context on one page.
The daily working list ranks records by composite score within the segment. Filters narrow the list to the active window. The rep opens one list and works top-down.
When a lead crosses the hand-off threshold, it lands in the sales development queue with the drivers cited and the counterfactual visible. The opener draft lives on the same page.
The sales manager sees the score distribution across the team: how many top-band accounts each rep is working, how many are stale, and how the mix compares to the segment benchmark. Coaching runs on the mix.
Trigger events on records the rep owns land in the daily queue rather than in a separate email digest. The rep sees the alert where they work.
Strkr AI drafts the first-touch opener against the top drivers. The opener references the signals that drove the score. The rep reviews, edits, and sends. The opener is a draft, not a sent email.
The model configuration lives in a documented place. The compliance review runs against documentation rather than tribal memory. The data is portable through the open API if the organization changes platforms. The common failure of a model nobody can audit or move gets closed off.
Every change to the signal taxonomy, the weight set, the ICP definition, the trigger list, and the hand-off threshold is recorded with the user, the timestamp, the before-and-after values, and the business rationale. The compliance review runs against the log.
The Strkr SOC 2 Type II audit is in progress through 2026-Q4 with the controls catalog available on request. Vendor security review teams see the posture, the controls in scope, and the audit timeline before signing.
Strkr is HIPAA compliant with a Business Associate Agreement available. Models in healthcare, life sciences, and health-adjacent lines run on the HIPAA posture.
The score, the drivers, the signal taxonomy, and the weight set are queryable through the open API. The data team can audit the model and pull the records into the warehouse without clicking through the product UI.
The signal inventory, the weight set, the ICP definition, the trigger list, the hand-off threshold, and the score history are exportable in a structured format on demand. The organization is never locked into a product-specific format.
Access to the model configuration is scoped by role. Only the sales operations lead and designated administrators can tune the weights, the ICP, and the hand-off threshold. The sales team reads the score and the drivers without editing.
Strkr runs lead scoring as a transparent, tunable composite with named signal branches, cited drivers, counterfactual panels, and nightly plus trigger-driven recompute. Fifty dollars per seat flat.
Four properties: a named signal taxonomy with named owners, tunable weights calibrated against your closed-won pattern, cited drivers on every record (not a bare number), and a recompute cadence that keeps the model current. A model missing any of those four drifts out of alignment with the pipeline within two quarters.
The initial weight set is pulled from the historical closed-won pattern across your pipeline. The operations lead reviews the data-driven starting point on the admin page and tunes from there. The weights are not vendor defaults; they are grounded in the real pipeline shape.
The model runs nightly across every lead and every open opportunity. Intra-day re-scores fire on explicit triggers so recently-moved records show the updated score within minutes. The last recompute timestamp lives on the record page.
Yes. The signal taxonomy, the weights, the ICP definition, the trigger list, and the hand-off threshold are tunable on the admin page. The counterfactual preview shows the impact of a change before committing. The shadow set runs the new configuration against historical records. The common failure of a model that requires a Python notebook for every tune gets avoided.
Strkr is HIPAA compliant with a Business Associate Agreement available, and the SOC 2 Type II audit is in progress through 2026-Q4. The audit log records every configuration change. Role-based access scopes model editing to the operations lead and designated administrators.
Yes. The signal inventory, the weight set, the ICP definition, the trigger list, the hand-off threshold, and the score history are exportable through the open API in a structured format on demand. The configuration is portable. Vendor change becomes a workflow rather than a reconstruction project.
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.
