Account Scoring: Prioritizing Accounts by Fit and Intent

Most B2B companies have a lead scoring model. Fewer have a well-functioning account scoring model. The distinction matters because in B2B, you're not selling to a person — you're selling to an organisation. A single enthusiastic contact in a company that doesn't fit your ICP is a trap. A company that fits your ICP perfectly but shows no current buying signals is a waiting game. Account scoring is how you find the accounts that are both a good fit and showing up right now.

Done well, account scoring transforms how your revenue team allocates its most limited resource: attention. This guide covers the mechanics of building a model that actually works — one your sales team trusts enough to act on.

Why Account Scoring Beats Lead Scoring in B2B

Lead scoring was designed for transactional, single-buyer sales cycles. You assign points to individual prospect behaviours — opened an email, visited the pricing page, downloaded a guide — and trigger a sales follow-up when the score crosses a threshold. It was a major improvement over the previous approach, which was "call everyone."

But in B2B, especially mid-market and enterprise, lead scoring has a fundamental problem: it ignores account context. A marketing manager who downloaded your whitepaper is not a buying signal on its own. That same download, from a company in your ICP, whose VP of Revenue also visited your pricing page last week, from a 400-person SaaS company in a vertical you serve well — that's a buying signal.

Research from Gartner consistently shows B2B buying committees involve between 6 and 10 stakeholders. Lead scoring only sees one of them at a time. Account scoring aggregates signals across the entire account so you can see the full picture.

The practical result: account scoring reduces wasted outbound effort dramatically. Teams using account scoring models typically find that the top 20% of scored accounts generate over 60% of new logo pipeline. That's not because the other 80% are unqualified — it's because timing and fit haven't aligned yet.

The Two Pillars: ICP Fit Score and Intent Score

Every sound account scoring model rests on two independent dimensions that are then combined into a single composite score.

ICP Fit Score measures how closely an account matches your Ideal Customer Profile. It's a static-ish quality score based on firmographic and technographic data. A high fit score means this type of company has historically bought from you, expanded with you, and churned less. The inputs are things like: industry vertical, company size (headcount and revenue), geography, tech stack, business model (SaaS, services, manufacturing), and growth signals like hiring rate.

Intent Score measures how active an account is right now around the problem your product solves. It's a dynamic, time-decaying signal based on behavioural data. Inputs include: first-party signals (visits to your website, engagement with your content, responses to outreach), second-party signals (activity data from your partner network), and third-party intent data from platforms like Bombora, G2, TrustRadius, or 6sense.

The reason you need both is that each pillar fails on its own. A high-fit account with no intent is a long-play prospect — it should be in a nurture sequence, not an outbound sequence. A high-intent account with poor fit is a waste of sales time. The accounts that deserve immediate sales attention score high on both.

Building Your ICP Fit Score

Start by analysing your closed-won customers from the last 18–24 months. Look for the firmographic and technographic patterns that the best accounts share — not just that they bought, but that they expanded, renewed, and didn't require disproportionate support. These are the accounts you want to clone.

Common ICP fit attributes and how to weight them:

• Industry vertical (20–25% of fit score): If you close 70% of deals in SaaS and professional services, those verticals should get a high score. Weight verticals based on your actual win-rate data, not assumptions.
• Company size (20%): Typically headcount bands or revenue ranges. Most companies have a "sweet spot" where they win most consistently. Companies significantly above or below that range score lower.
• Technology stack (15–20%): Which CRMs, MAPs, ad platforms, and data tools an account uses can be a strong fit signal. A company running Salesforce + HubSpot + 6sense is probably more ready for a sophisticated B2B marketing solution than one running spreadsheets.
• Growth signals (10–15%): Is the company growing headcount in marketing or revenue functions? Recent funding? Expanding to new markets? These are forward-looking fit signals that suggest budget and urgency may follow.
• Geography (10%): If you serve specific regions well, geography matters.

Providers like Clearbit, ZoomInfo, and Apollo can populate most of these attributes automatically when an account enters your CRM. The score should update when firmographic data refreshes, typically monthly or quarterly.

Capturing and Interpreting Intent Signals

Intent signals fall into three tiers based on how directly they indicate purchase readiness.

Tier 1 — High-intent first-party signals: Visited your pricing page, started a free trial, engaged with a bottom-of-funnel piece like a comparison guide or ROI calculator, booked a demo but didn't show, or had multiple visits from different people at the same company within a short window. These signals are gold because they're direct — the account is actively researching you or your category.

Tier 2 — Mid-intent first-party signals: Downloaded a top-of-funnel guide, attended a webinar, opened multiple emails, or visited solution pages repeatedly. These indicate awareness and some interest but not active buying mode.

Tier 3 — Third-party intent signals: Accounts researching competitor reviews on G2, spiking on category-relevant keyword clusters tracked by Bombora, or showing increased activity on content related to your problem category. These signals are useful for identifying accounts entering the market before they've found you, but they're noisier and require validation.

6sense and Demandbase add a predictive layer on top of these signals, using AI models trained on historical deal data to estimate where an account is in their buying journey. These predictions are most useful for large ICP-fit accounts where the signals are abundant enough for the model to be meaningful.

When scoring intent, weight recency heavily. An account that visited your pricing page three days ago scores far higher than one that did the same thing eight months ago. Most intent scoring models apply exponential decay — signals lose half their value every 30 days.

Combining Fit and Intent Into a Composite Score

Once you have separate fit and intent scores, combine them into a composite that drives prioritisation. The most common approach is a simple matrix: plot accounts on a 2x2 with fit on one axis and intent on the other, then define routing logic for each quadrant.

High Fit + High Intent (Tier 1 — Act Now): Immediate SDR or AE outreach. These accounts should enter an active sequence within 24 hours of hitting the threshold. For large enterprise accounts, trigger an account-based play that involves marketing and sales coordination.

High Fit + Low Intent (Tier 2 — Develop): These accounts belong in targeted nurture programs — personalised ad campaigns, account-specific content, executive outreach from field marketing. The goal is to create intent, not to chase it prematurely.

Low Fit + High Intent (Tier 3 — Qualify Carefully): Someone is interested, but it's unclear whether they're actually a good customer. Route to an SDR for light qualification before investing sales time. Many of these will be poor fit even if the individual buyer is enthusiastic.

Low Fit + Low Intent (Tier 4 — Monitor Only): Not worth active effort. These accounts should be in a broad awareness sequence at most — no personalised outreach, no dedicated budget.

A weighted composite score — for example, 50% fit + 50% intent — allows you to rank accounts numerically within each quadrant for further prioritisation.

Score Decay and Model Maintenance

Account scoring models go stale. The signals that predicted conversion 18 months ago may not be the strongest predictors today, especially if your product, pricing, or ICP has evolved. Plan to review your model every six months and retrain it annually against recent closed-won data.

Score decay is especially important for intent scores. An account that spiked in interest six months ago and hasn't engaged since has likely made a decision — either they bought a competitor, put the project on hold, or are in a blackout period. Continuing to flag them as high-priority burns sales goodwill in the scoring system.

Most teams implement automated score decay in their CRM or ABM platform: intent-based scores drop by a defined percentage each week without refreshed signals. HubSpot calls this "score decay" natively; in Salesforce, it requires a scheduled Apex job or a third-party tool like LeanData.

Also watch for model drift. If your closed-won customer profile shifts — for example, you start winning more deals in financial services and fewer in retail — your fit score needs to be updated to reflect that. A quarterly comparison of your top-scored accounts against your actual new logos is the simplest health check.

Integrating Account Scores Into Your CRM and Workflows

A score that lives in a spreadsheet or a separate analytics tool doesn't drive behaviour. For account scoring to change what your team does every day, it needs to be surfaced in the tools your salespeople actually use.

In Salesforce, store the composite score and the individual fit/intent components as custom fields on the Account object. This allows sales reps to sort and filter their named accounts by score, and lets you build views like "My Tier 1 accounts with recent activity." Sync the score daily so reps see current data, not week-old numbers.

In HubSpot, account scores work similarly via the company record. HubSpot's predictive lead scoring can be extended with custom scoring properties. For teams running 6sense or Demandbase, these platforms push scores directly into the CRM via native integration.

Build automated alerts for score spikes. When an account jumps from Tier 2 to Tier 1 — typically triggered by a burst of high-intent signals — the assigned rep or SDR should receive a Slack notification or a task in their CRM within hours. Speed-to-engagement matters when intent is hot.

For marketing, expose scores in your MAP (Marketo, HubSpot, Pardot) so you can trigger account-based advertising, direct mail, or personalised email sequences based on score tier changes. This ensures that high-fit accounts get the right content while they're in-market without waiting for sales to manually initiate a campaign.

Common Account Scoring Mistakes and How to Avoid Them

The most common failure mode is building a model that scores accounts by activity volume rather than buying intent. Opening ten emails is not the same as visiting your pricing page. If your model overweights engagement with top-of-funnel content, you'll surface a lot of engaged non-buyers and miss genuine opportunities that are doing research quietly.

Another frequent mistake is not involving sales in model design. If the reps who receive Tier 1 accounts think the scoring is noise, they won't act on it — and then the model gets blamed for poor results. Run regular win/loss reviews with your sales team specifically looking at whether high-scored accounts were being worked effectively and whether deals that were lost showed earlier score signals you missed.

Third, many teams underinvest in data quality. A fit score built on stale CRM firmographic data — companies that have changed size, vertical, or tech stack — will misffire regularly. Budget for data hygiene alongside the scoring model itself. Tools like Clearbit Enrich, ZoomInfo, or Cognism can automate firmographic refreshes on a schedule.

Finally, don't make the model too complex. A five-factor model with well-calibrated weights that your team understands and trusts will outperform a 40-factor black box. Sales adoption drives outcome, and adoption requires comprehension.