Account scoring helps B2B teams stop guessing and start focusing on the accounts that actually matter. Whether you're building your first model or fixing one that isn't working, the questions below cover everything — from basic definitions to implementation details. For a full walkthrough, see our in-depth guide to account scoring.

What is account scoring?

Account scoring is a data-driven method of assigning numerical values to potential customer accounts so your sales and marketing teams know which companies to prioritize. Instead of ranking individual people, you rank entire organizations based on how well they match your ideal customer profile, how engaged they are with your brand, and whether they're showing buying intent.

Think of it like a credit score — but for companies in your pipeline. A high score tells your team that a company is worth pursuing right now. A low score says "not yet" or "not a fit."

The output is simple: a ranked list of accounts, ordered by their likelihood to convert and generate revenue. Your best reps work the highest-scoring accounts. Marketing targets campaigns at the warmest tiers. Everyone stops wasting time on companies that were never going to buy.

How does account scoring work?

Account scoring works by collecting signals from multiple data sources, weighting those signals based on their predictive value, and calculating a composite score for each account.

Three categories of signals feed the model:

• Fit signals — firmographic data like industry, company size, revenue, geography, and tech stack. These tell you whether an account matches your ICP.

• Intent signals — third-party research activity (topic searches, competitor comparisons, review site visits) and first-party behavior (pricing page visits, whitepaper downloads). These tell you whether an account is actively in-market.

• Engagement signals — direct interactions with your brand like email opens, webinar attendance, ad clicks, and sales conversations. These tell you whether an account knows who you are and is leaning in.

Each signal gets a point value. The signals are aggregated at the account level — not the individual level — so that activity from multiple stakeholders at the same company rolls up into one score. When a score crosses a predefined threshold, it triggers an action: immediate sales outreach, a nurture sequence, or continued monitoring.

What's the difference between account scoring and lead scoring?

Account scoring evaluates entire companies; lead scoring evaluates individual contacts. That's the core difference, and it changes everything about what you measure and how you act on it.

Lead scoring assigns points to a single person based on their job title, behavior (email clicks, form fills), and demographic fit. When they cross a threshold, they become an MQL and get routed to sales.

Account scoring aggregates signals across the entire buying committee — every person at a company who visits your site, attends a webinar, or researches your category. This matters because modern B2B deals involve 6–10 stakeholders. One junior analyst downloading a whitepaper shouldn't trigger a sales call. But three directors from the same company researching your product category? That's a real signal.

The smartest teams use both together. Account scoring identifies which companies to focus on. Lead scoring identifies which people within those companies to contact first.

What data do I need for account scoring?

You need three types of data: firmographic, behavioral, and intent. The more complete and accurate each layer is, the better your model performs.

Firmographic data includes company size, industry, annual revenue, headquarters location, and technology stack. This is the foundation — it tells you whether an account even fits your ideal customer profile.

Behavioral data comes from your own channels: CRM records, marketing automation (email opens, clicks), website analytics (page visits, time on site), event attendance, and sales activity logs.

Intent data comes from third-party providers that track research activity across the web — which companies are searching for topics in your category, reading competitor reviews, or consuming relevant content. Tools like Bombora, G2, and 6sense provide this layer.

One thing that's often overlooked: data quality matters more than data volume. A scoring model built on outdated company sizes, wrong industries, or stale contact records will produce garbage scores. If your CRM hasn't been enriched recently, fix that before building a scoring model. Platforms like FullEnrich can help — waterfall enrichment across 20+ data providers fills in missing firmographic fields and validates contact data so your scores are built on accurate foundations.

What types of account scoring models exist?

There are four main model types, each with different tradeoffs between simplicity and accuracy.

Point-based (additive) scoring is the simplest. You assign fixed point values to each attribute — +10 for matching industry, +15 for having 200+ employees, +5 for each content download — and sum them up. Easy to build and explain, but it can't capture how signals interact.

Weighted formula scoring applies multipliers to different scoring dimensions. For example: Total Score = (Fit × 0.4) + (Engagement × 0.3) + (Intent × 0.3). This lets you emphasize what matters most for your business without overcomplicating things.

Tiered scoring groups accounts into buckets — Tier A, B, C, D — based on combined scores. Each tier maps to a specific action. This is less about the exact number and more about the category: "hot," "warm," "nurture," or "monitor."

Predictive (ML-based) scoring uses machine learning trained on your historical win/loss data to identify patterns humans might miss. It's the most accurate approach, but it requires a large enough dataset and ongoing maintenance.

Most B2B teams start with a weighted formula model and graduate to predictive scoring as their data matures. For a deeper look at model design, see our complete account scoring guide.

What criteria should I use to score accounts?

The criteria that matter most are the ones that actually predict closed-won deals in your business — not generic best practices.

Start by analyzing your last 50–100 closed-won customers. Look for patterns in:

• Industry — Do certain verticals convert at higher rates?

• Company size — Is there a sweet spot (e.g., 100–500 employees)?

• Revenue range — Do accounts above a certain revenue threshold close faster?

• Tech stack — Do companies using complementary tools convert more often?

• Engagement depth — How many stakeholders engaged before deals closed?

• Intent signals — Were closed-won accounts researching your category before they booked a demo?

Then assign weights based on correlation strength. If industry is the strongest predictor, it should carry the highest weight. If company size barely matters, don't give it 20% of the score.

Don't forget negative scoring. Visiting your careers page, having a competitor domain, or unsubscribing from emails should reduce an account's score — not leave it unchanged.

What role does intent data play in account scoring?

Intent data tells you which accounts are actively researching solutions like yours — often before they ever visit your website or fill out a form. It transforms account scoring from "who fits our profile" to "who fits our profile and is ready to buy."

There are two types:

• First-party intent — Signals from your own properties: pricing page visits, case study downloads, demo requests, repeated visits from the same company domain.

• Third-party intent — Signals from external sources: topic-level research activity detected across publisher networks, review sites like G2, and web-wide content consumption.

Without intent data, you're scoring on fit alone. An account might match your ICP perfectly but have zero interest in buying right now. Intent data separates "good fit, bad timing" from "good fit, actively evaluating" — and that distinction is what makes scoring actionable.

How is account scoring different from account tiering?

Account scoring assigns a numerical value to each account. Account tiering groups accounts into priority levels — usually Tier 1, 2, and 3 — based on how much attention and resources they deserve.

Scoring is the input. Tiering is the output.

In practice, most teams use scoring to inform tiering. Accounts that score above 80 go into Tier 1 and get personalized, 1:1 outreach. Accounts scoring 50–79 go into Tier 2 and get targeted campaigns. Accounts below 50 go into Tier 3 and get broad-reach marketing until their signals strengthen.

You can also tier accounts manually based on strategic importance (e.g., a logo you want regardless of score). But for most of your pipeline, letting the score drive the tier removes bias and ensures consistency.

Who should own account scoring on my team?

Revenue Operations (RevOps) should own the scoring model, with input from both sales and marketing. If you don't have a RevOps function, the person closest to your CRM and data infrastructure — typically a marketing ops or sales ops lead — should take ownership.

Here's why shared ownership matters: if marketing builds the model in isolation, sales won't trust it. If sales defines scoring criteria alone, they'll bias toward gut feel rather than data. The model works when both sides agree on what signals matter, how they're weighted, and what thresholds trigger action.

Run a joint workshop to define the initial model. Then review it quarterly with pipeline data to validate or recalibrate.

How do I set scoring thresholds that drive action?

Define clear tiers with specific actions attached to each one. A score means nothing if your team doesn't know what to do when an account hits 75.

A common framework:

• Tier A (80–100): High-fit, high-intent, engaged. Route to AE for personalized outreach within 24 hours.

• Tier B (50–79): Good fit with emerging signals. Enroll in ABM campaigns and SDR sequences.

• Tier C (25–49): Partial fit or low engagement. Add to nurture programs. Monitor for score changes.

• Tier D (0–24): Poor fit or no activity. Passive monitoring only.

The thresholds should be tight enough that Tier A feels genuinely hot — not diluted with mediocre accounts. If 40% of your accounts land in Tier A, your thresholds are too generous. Tighten them until Tier A represents the top 10–15%.

Wire these tiers into your CRM so score changes automatically trigger workflows — task creation, Slack notifications, campaign enrollment — without anyone manually checking dashboards.

What are the most common account scoring mistakes?

The biggest mistake is building a model and never validating it. Teams spend weeks defining criteria and weights, launch the model, and never check whether high-scoring accounts actually convert at higher rates. If Tier A and Tier C accounts close at the same rate, the model is broken.

Other common mistakes:

• Scoring on fit alone. A perfect ICP match that shows zero intent or engagement shouldn't be Tier A. Fit without interest is just a database entry.

• Overcomplicating the model. Fifty scoring attributes are impossible to maintain and impossible for sales to understand. Start with 8–12 high-impact signals and expand gradually.

• Ignoring negative signals. Career page visits, email unsubscribes, and competitor domains should subtract points. Not every interaction is a buying signal.

• Using stale data. If your firmographic data hasn't been updated in 6 months, your fit scores are unreliable. Company sizes change, industries shift, people leave. Regular data enrichment is non-negotiable.

• Not involving sales. If reps don't trust the scores, they'll ignore them entirely. Include sales leaders in model design and share closed-won data that proves the model works.

How often should I update my account scoring model?

Review your model at least once per quarter. Compare scoring predictions against actual outcomes — win rates, deal sizes, and sales cycle length by tier — and adjust criteria and weights based on what you find.

Scoring models decay over time. Your ICP may shift as you move upmarket or enter new verticals. Buyer behavior evolves. New competitors emerge. The criteria that predicted wins 6 months ago may not work today.

Signs your model needs recalibration:

• Tier A accounts aren't converting at meaningfully higher rates than Tier B.

• Sales reps are ignoring scores because they don't match reality.

• Win rates haven't improved since you implemented scoring.

• High-scoring accounts churn shortly after closing.

Beyond quarterly reviews, apply time-based decay to engagement and intent signals. A pricing page visit from last week should carry more weight than one from 90 days ago. Stale signals inflate scores and create false urgency.

How does data quality affect account scoring?

Data quality is the single biggest determinant of whether your scoring model works or fails. Bad data in, bad scores out — no amount of model sophistication can fix inaccurate inputs.

Common data quality problems that break scoring:

• Missing firmographic fields — If 30% of your accounts don't have industry or company size populated, your fit scores are incomplete.

• Outdated records — A company that was 50 people when you imported them may now be 500. Scoring them as a small business when they're mid-market undermines the entire model.

• Duplicate accounts — The same company appearing twice (or five times) in your CRM fragments engagement signals and produces artificially low scores.

• Incorrect contact data — If the people associated with an account have left the company, engagement signals from their replacements may not get attributed correctly.

Before launching a scoring model, audit your data. Run enrichment to fill gaps and validate existing records. Platforms that use waterfall enrichment — querying multiple data providers in sequence — deliver the highest coverage because no single vendor has complete data on every company.

Can small teams use account scoring?

Yes — and small teams often benefit the most because they have the least margin for wasted effort. When you only have two or three reps, every hour spent on the wrong account is an hour you can't afford.

You don't need enterprise software to start. A basic scoring model in a spreadsheet works if your account list is under 500. Define 5–8 criteria, assign simple point values, and rank your accounts. The sophistication can grow with your team.

What matters more than the tool is the discipline: defining your ICP clearly, choosing criteria based on actual closed-won data, and reviewing the model regularly. A simple model that's maintained beats a complex one that's abandoned after a month.

How does account scoring fit into an ABM strategy?

Account scoring is the prioritization engine inside an account-based marketing framework. ABM starts with choosing which accounts to target. Scoring tells you which of those accounts to engage first — and how aggressively.

In a typical ABM workflow:

1. Build your target account list based on ICP criteria.

2. Score each account using fit, intent, and engagement data.

3. Tier them into priority groups (Tier 1, 2, 3).

4. Design plays per tier — 1:1 personalization for Tier 1, 1:few campaigns for Tier 2, programmatic for Tier 3.

5. Monitor scores over time and move accounts between tiers as signals change.

Without scoring, ABM devolves into "spray and pray with a smaller list." Scoring ensures your ABM budget is concentrated where the buying signals are strongest.

How do I measure whether account scoring is working?

Track these metrics, segmented by score tier:

• Win rate by tier — Tier A accounts should close at 2x or higher the rate of Tier C. If the gap is small, your model isn't differentiating well enough.

• Sales cycle length — High-scoring accounts should move through the pipeline faster because you're engaging them when intent is highest.

• Average contract value — If your ICP definition is accurate, top-tier accounts should also produce larger deals.

• Pipeline contribution by tier — Ideally, Tier A and B accounts generate 70–80% of your qualified pipeline.

• Sales adoption — Are reps actually using scores to prioritize? Low adoption means the model doesn't match their reality.

Start measuring within the first month of implementation. Give the model a full quarter before making major changes — but monitor continuously so you catch obvious problems early.

How do I get started with account scoring?

Start simple and iterate. Here's a practical path:

1. Define your ICP. Analyze your best 30–50 customers. What do they have in common? Industry, size, revenue, tech stack, buying process. Write it down.

2. Pick 8–12 scoring criteria across fit (firmographics), engagement (behavioral data), and intent (research signals). Don't overcomplicate it.

3. Assign point values based on how strongly each criterion correlates with closed-won deals. Give more weight to criteria that show a clear pattern.

4. Set thresholds that map to actions. Define what happens at each tier — who gets contacted, how fast, and through which channel.

5. Clean your data. Enrich missing firmographic fields, deduplicate accounts, and validate contacts. Your model is only as good as the data feeding it.

6. Launch, measure, and adjust. Run the model for 90 days, then compare predicted scores against actual outcomes. Recalibrate weights and thresholds based on results.

You don't need a perfect model on day one. You need a directionally accurate model that you improve every quarter. The companies that win at account prioritization are the ones that treat scoring as a living system — not a one-time project.