Account Scoring: Rank Companies, Not Just Leads

B2B purchases are rarely made by one person. A finance director reads a pricing page, an IT manager downloads a security document and an operations lead attends a webinar, all from the same company, in the same month. A lead-by-lead model sees three moderately interested people. Account scoring sees one company showing strong, coordinated interest. Scoring at the account level helps sales focus on organisations, not just individuals, and is a natural fit for account-based selling. This guide explains how to score whole accounts by fit and engagement, roll up activity from contacts, avoid the common traps and put account scores to work.

Why lead scores miss buying groups

Lead scoring was designed for a world where one person fills in a form and becomes a customer. In B2B sales, several people usually influence a purchase, and each may engage in small ways. Lead scores split that interest across records. Worse, the contact who scores highest may not be the decision maker, while the account as a whole is close to buying.

Account scoring adds a second view. Leads and contacts keep their individual scores for follow-up, while the account score answers a different question: which companies should we prioritise right now?

The two parts of an account score

Like a good lead score, an account score separates fit from engagement:

• Account fit: how closely the company matches the customers you win and keep, based on industry, size, location, technology used and other firmographic data
• Account engagement: how much the people at the company are interacting with you, combined across all known contacts and, where available, anonymous visits matched to the company

Keep these as two numbers, or a grid, rather than one total. A perfect-fit account with no engagement needs outreach; a highly engaged account with poor fit may need a lighter touch.

Build account fit from evidence

Fit criteria should come from your own history. Compare won accounts, and accounts that renewed or expanded, with those that churned or never converted. Look at:

• Industry and sub-industry
• Employee count or revenue band
• Region and languages served
• Technologies they already use
• Business model, such as subscription, project or retail

Technology signals can be powerful when your product depends on or replaces specific tools. Our guide to technographic data covers sources and how to use them carefully.

Roll up engagement from contacts

Account engagement combines activity from everyone at the company. Design decisions:

• Matching: link contacts and leads to accounts reliably, usually by company domain plus manual review for edge cases such as shared email providers or subsidiaries
• Breadth: give extra weight when several different people engage, not just one very active person
• Role weighting: count activity from target roles, such as decision makers and evaluators, more heavily
• Recency: let engagement decay so last quarter's activity does not keep an account at the top
• Caps: limit how much one person or one activity type can contribute, so a single enthusiastic researcher cannot dominate

The individual lead scoring criteria you already use are a good starting point for the activity weights, adjusted for the account view.

Worked example: an account scoring model

The table below shows a hypothetical account scoring model for a company selling to mid-sized manufacturers. Points, caps and thresholds are illustrative.

ComponentCriterionPointsNotesFitTarget industry+25Based on won accountsFitEmployee band in target range+20From enrichmentFitUses a compatible ERP+15From technographic dataFitRegion not served-50Effectively excludesEngagementEach engaged contact in a target role+10Capped at 4 contactsEngagementDemo or quote request from any contact+30Strongest signalEngagementPricing page visits across contacts+5 eachCapped at +15, expire after 30 daysEngagementWebinar attendance across contacts+5 eachCapped at +10, expire after 60 daysEngagementThree or more engaged roles+15Breadth bonus

In this example, an account is flagged for sales when fit is at least 45 and engagement is at least 40, or immediately after a demo request from an account with acceptable fit.

Handle existing customers separately

Accounts that already buy from you behave differently. Their staff log in, read help articles and attend training, which can look like strong engagement in a new-business model. Decide early whether existing customers are excluded from the new-business account score or scored in a separate model aimed at expansion and renewal risk. Mixing them creates noisy priority lists, and new-business reps end up chasing accounts that the account management team already owns.

Data you need before you start

Account scoring depends on account data more than lead scoring does. Check that:

• Most accounts have industry and size values from a consistent source
• Contacts are linked to the correct account, including people who joined from a personal email
• Website and email activity can be tied to known contacts, and through them to accounts
• Duplicate accounts are merged, so engagement is not split across two records

If these basics are weak, fix them first. A sophisticated model on poorly matched data will rank the wrong companies with confidence.

Put account scoring to work

Scores only matter if they change what people do. Common uses:

• Prioritised account views for each rep, sorted by fit and engagement
• Alerts when an account crosses the engagement threshold
• Routing new accounts to the right team or territory
• Coordinated outreach to several contacts at a high-scoring account
• Marketing campaigns aimed at high-fit accounts with low engagement

In most CRMs, account scores can live in custom fields on the account record, calculated by built-in rules where supported, by workflows, or by an external service that writes them back.

Common account scoring mistakes

• Poor matching of contacts to accounts, which splits or misattributes engagement
• Counting raw activity totals, so one very active person outweighs a whole buying group
• No decay, so old engagement keeps accounts at the top
• Mixing fit and engagement into one number
• Ignoring existing customers, whose engagement may signal expansion or risk instead of new business
• Building the model without input from the reps who work the accounts

Check that the model works

Review account scoring each quarter:

• Do accounts above the thresholds create opportunities at higher rates?
• Which won accounts never crossed the threshold, and why?
• Are some fit criteria true for nearly every account, making them meaningless?
• Do reps find the account views useful, or do they work from their own lists?

Change one or two criteria at a time and watch the effect over the next period. Keep a short record of each change, the reason and the result.

Frequently Asked Questions

What is account scoring?

It is a way of ranking companies, rather than individual leads, by how well they fit your ideal customer and how much their people are engaging with you.

How is account scoring different from lead scoring?

Lead scoring rates individuals. Account scoring combines the activity of everyone at a company and adds company-level fit, which suits B2B purchases made by groups.

How do you link leads to accounts?

Usually by matching company email domains to account records, with rules for shared email providers, subsidiaries and multiple domains, plus manual review for unclear cases.

Should account scores replace lead scores?

No. Use both. Lead scores guide follow-up with individuals, while account scores help sales decide which companies to prioritise.

Prioritise the companies most likely to buy

We design account scoring models, connect contact activity to accounts and set up prioritised views and alerts in your CRM. See our AI automation solutions, or talk to us about account scoring.

About the Author

Manav Sharma is the founder of Insyrge and a CRM solutions architect. He works across Zoho, Salesforce, HubSpot, Odoo and more, designing the CRM setups, integrations and automations that keep sales and operations teams working from the same data.