What role should propensity models play in account selection?

Propensity models should prioritize and sequence accounts, not replace human judgment. Use them to rank-fit opportunities, size the next-best action, and time outreach—then prove lift with controlled tests. Calibrate quarterly and combine with fit + intent + whitespace + relationship for tiering.

Principles For Using Propensity In Selection

Anchor to ICP — Models work best on a well-defined ICP; exclude poor-fit accounts even with high scores.

Blend multiple signals — Combine model score with buying intent, product whitespace, and relationship strength.

Set clear thresholds — Define route/tier rules (e.g., P≥0.7 → Tier A; 0.4–0.69 → Tier B) and revisit monthly.

Guard against bias & drift — Monitor calibration, fairness, and data drift; retrain with recent wins and losses.

Tie to actions — Each score band should trigger a play: 1:1 ABM, 1:few programs, or automated 1:many motions.

Measure incremental value — Use holdouts to verify that propensity-led prioritization increases pipeline and win rate.

The Propensity-Driven Selection Playbook

A practical sequence to design, deploy, and govern propensity for account selection and tiering.

Step-by-Step

Define the decision — Prioritization, routing, tiering, timing, or next best product? Choose the outcome up front.
Assemble data — Wins/losses, firmographics, technographics, engagement, product usage, intent, and CRM notes.
Build & validate — Start with interpretable models (e.g., logistic/GBM). Validate with AUC, calibration, and stability.
Operationalize bands — Map score bands to plays, SLAs, cadences, and owners; publish in your GTM guide.
Add business rules — Apply ICP exclusions, strategic accounts, partner conflicts, and coverage caps.
Run lift tests — Compare “model-first” prioritized lists vs. control for pipeline coverage, win rate, and payback.
Monitor & retrain — Monthly drift checks; quarterly retrains; annual feature review and fairness audit.

Where Propensity Helps Most

Use Case	Signals Used	Primary Decision	Recommended Motion	Success Metrics	Notes
Net-New Prioritization	Fit + intent + lookalike wins	Prospect order & timing	SDR outreach, 1:few ABM	Meetings set, SQOs, win rate	Throttle by capacity; refresh weekly
Cross-Sell / Upsell	Usage gaps, product fit, seat growth	Next best product	CSM-led plays, in-product prompts	Attach rate, ARR expansion, NRR	Coordinate with renewal calendar
Churn Risk & Saves	Health, support, executive access	Save vs. expand readiness	Success plans, exec check-ins	GRR, time-to-intervene	Route red accounts to save squad
Territory & Coverage	Aggregate propensity by region/segment	Headcount and quota mix	Pod design, program allocation	Coverage ratio, CAC/payback	Great for annual planning

Client Snapshot: From Scores To Revenue

A data-rich B2B platform layered a propensity model onto fit+intent tiering and assigned score bands to distinct plays. Over two quarters, meetings per SDR rose 22%, win rate increased 4.8 points, and expansion attach rate lifted 15% with verified causal impact from holdouts.

Integrate propensity with account-based programs and a clear revenue operating model so analytics turn into action across GTM.

FAQ: Using Propensity Models In Account Selection

Concise answers for leaders making selection and tiering decisions.

Should propensity determine tiers by itself?

No. Use it to prioritize within ICP-qualified accounts. Tiering should blend fit, intent, whitespace, and relationship signals.

How do we validate the model?

Run A/B or holdout tests that compare propensity-led queues vs. business-as-usual; track meetings, SQOs, win rate, and payback.

How often should we retrain?

Quarterly for fast-moving markets; semiannually if cycles are long. Monitor drift monthly and recalibrate thresholds as needed.

What about fairness and bias?

Audit features, remove proxies for protected classes, evaluate disparate impact, and document the governance process.

Where do these scores live?

Surface scores and bands in CRM/CS tools alongside ICP, intent, health, and whitespace so reps can trigger the right plays.

Put Propensity To Work, Responsibly

We’ll help you operationalize scores, govern thresholds, and tie actions to measurable lift across new and expansion.

Reimagine Revenue Engine Get the AI Revenue Guide

Explore More

Account-Based Marketing Services Revenue Marketing eGuide Revenue Marketing Transformation AI Revenue Enablement Guide

Account Selection & Tiering:
What Role Should Propensity Models Play In Account Selection?