Account scoring has become a critical lever for revenue operations teams, yet implementation remains messy in most B2B organizations. This guide walks through the mechanics of building a scoring model from first principles, avoiding the common pitfalls that derail most pilots.
Before diving into implementation details, let’s ground why this matters. Account scoring answers a specific question: of all the accounts in your territory, which ones should your sales team focus on right now?
Most B2B teams operate reactively, responding to inbound inquiry volume rather than pursuing accounts that match their ideal customer profile. This leads to wasted activity on poor-fit prospects and missed pipeline acceleration on high-value accounts showing signals of buying intent.
Account scoring bridges that gap by creating a systematic way to rank accounts across two dimensions:
The combination drives prioritization. An account with perfect fit but zero intent stays lower-priority than a good-fit account with strong buying signals. Conversely, high-intent without fit wastes sales cycles chasing accounts unlikely to close.
Fit scoring starts with your ICP. If you haven’t formalized your ideal customer profile, do that before building any scoring model.
A fit score typically weighs three categories of firmographic data:
Company Size: Revenue, headcount, funding stage, geographic location. For most B2B SaaS, company size correlates strongly with purchase budget and deal complexity. A Series B SaaS company should weight headcount and revenue more heavily than, say, an individual freelancer using your tool.
Industry and Vertical: Do they operate in your target verticals? Companies in fintech, healthcare, and cybersecurity often have distinct GTM requirements. Make sure your scoring reflects whether an account’s industry is part of your addressable market.
Technographic Fit: What tech stack are they using? If you’re selling a Salesforce integration, companies running Salesforce are higher-fit than those on a homegrown CRM. If you’re selling a DevOps tool, companies using Kubernetes are higher-fit than those on legacy infrastructure.
Most RevOps teams implement fit scoring with a simple weighted sum:
Fit Score = (0.4 × Size Score) + (0.35 × Industry Score) + (0.25 × Tech Score)
The exact weights depend on your business. A startup founder tool might weight founder profile heavily; an enterprise software vendor might weight company size almost exclusively.
Intent scoring is harder to systematize but matters more. A prospect in-market buying intent changes within weeks or months. A prospect with perfect fit but no buying signals might take 18 months to convert, if ever.
Intent signals fall into these buckets:
First-Party Signals (website behavior, email engagement): - Website visits and time-on-site - Content downloads (especially high-intent collateral like pricing pages, use-case guides, ROI calculators) - Demo request submissions - Email opens, clicks, and replies to outbound campaigns - Participation in webinars or virtual events
Third-Party Signals (intent data platforms): - Account appears in G2, Capterra, or industry review sites (researching solutions) - Keyword search volume in your solution category at that company’s IP address or domain - Tech stack changes indicating active buying process - Job postings for roles that suggest product deployment (e.g., hiring for a “data engineer” role might suggest they’re building a data stack)
Sales Engagement Signals: - Inbound inquiry from the account - SDR notes indicating interest - Meeting scheduled or attended - Proposal stage in CRM
Each signal gets a point value. Recent signals (past 30 days) are weighted more heavily than older signals (past 90 days). Here’s a simplified framework:
Intent Score = Sum of (Signal Points × Recency Weight)
Recent (0-30 days): 1.0x multiplier
Medium (30-90 days): 0.5x multiplier
Older (90+ days): 0.1x multiplier
Example point values:
Website demo page visit: 5 points
Content download: 3 points
Sales call attended: 15 points
G2 review page view: 2 points
Pricing page visit: 8 points
Now combine your fit and intent scores into a single composite score:
Account Score = (0.4 × Fit Score) + (0.6 × Intent Score)
This weighting assumes intent matters more than fit (which is typically true for short sales cycles), but adjust based on your data. A 24-month enterprise sales cycle might reverse these weights.
Account scoring only works if your input data stays clean.
Assign one team owner (usually your MarOps or RevOps lead) to audit the model monthly:
Check for data gaps: Which accounts have missing firmographic data? Which accounts haven’t received any signal tracking in the past 90 days?
Monitor score distribution: Are most accounts clustered in the middle (score 40-60)? That means your model has low signal strength. Adjust weights until you see more bimodal distribution with clear peaks at high-fit and low-fit.
Validate against closed-won deals: Pull your last 10 closed-won deals. What was their average account score at the time of sales engagement? Use this to set thresholds. If your average closed deal had a score of 65+, set that as your minimum threshold for active sales outreach.
Test weight adjustments: Every quarter, run a hypothesis test. Take your highest-scoring accounts (top 10%) and track their conversion rates. Compare to mid-range and low-range cohorts. If the top 10% aren’t outconverting, your weights may be miscalibrated.
You have three options for implementing account scoring:
Native CRM: Salesforce, HubSpot, and Pipedrive all support workflow automation and formula fields. A RevOps team can build scoring models using native fields + roll-up calculations. This is the simplest approach for teams under 500 accounts and with stable ICP criteria.
RevOps Stack: Solutions like Terminus, 6sense, and Rollworks provide pre-built intent data + account scoring modules. These work well if you want to leverage third-party intent signals without building your own signal ingestion layer.
Custom Data Stack: Larger teams (500+ accounts, complex weighting rules) often build custom Python/SQL scoring jobs that run daily/weekly, normalizing data from multiple sources (Salesforce, web analytics, intent platforms, enrichment APIs) into a single scoring table.
Mistake 1: Over-weighting single signals. Teams often obsess over website traffic because it’s easy to measure. But a single website visit from an anonymous account proves little. Weight first-party signals (email engagement, meeting attendance) more heavily than passive signals (page views).
Mistake 2: Ignoring negative signals. An account downloading your pricing page is a positive signal. An account unsubscribing from your nurture emails is a negative signal. Your model should reflect this.
Mistake 3: Scoring static accounts. Accounts that scored 75 last month should re-score this month based on new signal activity. Intent scores should decay if an account goes quiet. If an account hasn’t engaged in 120 days, lower their intent score even if it was previously high.
Mistake 4: Not documenting decision rules. Write down why you weight each signal the way you do. Document thresholds for moving accounts between tiers (hot, warm, cold). Your successor or new team member needs to understand the logic.
Mistake 5: Ignoring the sales feedback loop. Your sales team will judge your model based on their lived experience, not your Excel sheet. If they consistently say, “This account scored high but it’s a waste of time,” trust them. Adjust the model.
Once your model is built, operationalize it:
Surface scores in Salesforce/HubSpot: Add an Account Score field that updates daily or weekly. Sales reps need this visible in their dashboard.
Build a waterfall report: List all accounts by score. Identify the “move to active outreach” threshold. Track how many accounts move into/out of each tier weekly.
Alert on score changes: When an account jumps from 30 to 70+ (high-intent signal detected), alert the assigned sales rep. This is signal-to-action time.
Align with account assignment: Use account scores to inform territory and account assignment. Higher-scoring accounts should go to top performers.
Review monthly: Set a calendar block for monthly model review. Pull recent closed deals, validate thresholds, adjust weights if needed.
Many teams pair account scoring with tiered GTM motions:
This creates resource allocation clarity. You can’t run 1:1 ABM at every account. Scoring ensures your most intensive efforts target your highest-probability deals.
Set success metrics for your scoring model itself:
Conversion rate by score band: Do higher-scoring accounts close at higher rates? Track close rate for accounts in each score tier.
Sales cycle length: Do higher-scoring accounts close faster? A good model should show shorter average sales cycles for higher-scoring cohorts.
Pipeline velocity: Do higher-scoring accounts move through sales stages faster?
Track these metrics quarterly. If your model isn’t predictive, the weights are wrong or your signal sources are noisy.
Account scoring is a leverage point in revenue operations. A well-built, maintained scoring model creates clear prioritization for your sales team, increases win rates by focusing effort on high-probability accounts, and shortens sales cycles by alerting teams to accounts showing buying intent.
Start simple (fit + basic intent signals). Validate your model against closed-won deals. Iterate quarterly. Build with your sales team, not for them.
Account scoring has become a critical lever for revenue operations teams, yet implementation remains messy in most B2B organizations. This guide walks through the mechanics of building a scoring model from first principles, avoiding the...
Account scoring has become a critical lever for revenue operations teams, yet implementation remains messy in most B2B organizations. This guide walks through the mechanics of building a scoring model from first principles, avoiding the...