Lead Grading
Lead grading rates how closely a lead matches your ideal customer profile using firmographic and demographic fit, often expressed as a letter grade like A through F.
Key takeaways
- Grading measures fit against the ideal customer profile; scoring measures behaviour and engagement.
- Weights differ per attribute, and one exclusion can force the lowest grade outright.
- Four or five bands match the precision that firmographic data actually supports.
- Size the top band to the capacity of the sales team working it.
- A perfect grade says nothing about when the company is ready to buy.
In depth
A grade is produced by comparing a lead's attributes against a written ideal customer profile and converting the match into a letter. Each attribute carries a weight: industry might count more than headcount, and an explicit exclusion such as a competitor domain can force the bottom grade regardless of everything else. The output is deliberately coarse. Four or five bands are enough for routing decisions, and finer resolution implies a precision the underlying firmographic data does not have.
Grade distribution moves for two reasons: the mix of traffic changes, or the criteria change. If most leads suddenly grade A, either a campaign found an unusually good audience or the profile has been drawn too wide to be useful. A grading model that produces very few A grades is precise but leaves sales idle. The practical tuning target is a top band small enough to be worked properly by the reps you actually have.
Most teams grade at the point of capture, because the attributes needed are the ones a form can ask for. A Pivix scorecard makes this explicit: company size, industry, role and current setup are quiz questions, and the resulting band becomes the grade before the lead reaches the CRM. Grades then drive concrete rules such as who gets a same-day call, who receives a self-serve onboarding email, and which leads never enter a paid retargeting audience.
A grade describes fit and says nothing about readiness, so a company that matches the profile perfectly may still be two years from buying. It also depends on data that is often self-reported and sometimes wrong, and on a profile built from the customers you already won, which quietly excludes segments you have never sold to. For horizontal products with no firmographic pattern, grading adds ceremony without adding signal. Grades also age, because nobody refreshes the attributes after capture.
Example in practice
How to measure it
The test of a grading model is monotonicity: win rate should fall steadily from A to F. If B outperforms A, the weights are wrong. Run the same check on average contract value and on first-year retention, because a grade that predicts closing but not keeping customers is optimising for the wrong outcome. Compare distributions across acquisition channels as well.
Two operational readings matter alongside that. First, the share of leads that cannot be graded because attributes are missing; if it is large, the capture form is asking too little or the wrong things. Second, how often reps override a grade manually. Frequent overrides in one direction are a specification, not a complaint: they tell you which attribute the model is missing.
Common mistakes
The commonest failure is building the profile from a wishlist rather than from closed-won data. Someone names the logos they would like, the criteria encode that ambition, and the model grades real buyers as C because they are smaller than the aspiration. Derive weights from the accounts that actually bought and stayed, then check the grade distribution of last year's customers; if many of them would grade poorly, the model is wrong, not the market.
The second is treating the grade as a permanent property of the record. Companies hire, get acquired, change industry classification and switch tools, but the letter assigned at capture sits in the CRM untouched for years. Re-grade on a schedule and on trigger events, and store the grade with the date it was set so a rep can see whether they are acting on a judgement made this month or three years ago.
Frequently asked questions
What criteria are used for lead grading?
Common criteria include company size, industry, job title or role, geography, and use-case fit against your ideal customer profile. Some teams also factor in technographics like the tools a prospect already uses. The criteria should mirror the traits of your best existing customers.
How can a quiz funnel improve lead grading?
A scorecard quiz can ask the exact firmographic questions that drive a grade and assign a tier instantly from the answers. This means leads are graded at the moment of capture rather than enriched later. High-grade leads can then be routed to sales while low-grade leads go to self-serve.
What is the difference between lead grading and lead scoring?
Grading answers whether the lead is the right kind of company; scoring answers how engaged they are. Grading uses attributes that rarely change, such as industry and size. Scoring uses behaviour that changes weekly, such as page visits and email replies. Most teams keep them as separate fields and combine them into a routing rule, for example A-grade plus high score goes to sales.
How many grades should I use?
Four or five, including an explicit disqualified band. Fewer than four cannot separate 'worth a call' from 'worth an email', and more than five forces distinctions your data cannot support. What matters more than the count is that each band maps to a different action; a grade with no consequence attached is a label nobody will maintain.
Which attributes should carry the most weight?
Whichever ones separate your won deals from your lost ones. Compare closed-won and closed-lost accounts attribute by attribute and keep the ones where the distributions genuinely differ. Company size and industry are common winners, but for some products the deciding attribute is the existing tool stack or a regulatory obligation. Weights should be evidence, not opinion.
How often should grading criteria be reviewed?
Quarterly is a workable rhythm for most teams, with an immediate review whenever the product enters a new segment or pricing changes. The review does not have to be elaborate: pull the last quarter's won and lost deals, check whether grades predicted the outcome, and adjust one or two weights. Rewriting the whole model every quarter makes historical comparison impossible.
Can I grade leads without buying a data enrichment service?
Yes, by asking. A short quiz or form can collect industry, company size, role and current tooling directly from the respondent, which covers most grading models. Self-reported data has its own error rate, mainly optimistic rounding of company size, but it is available at capture and costs nothing per record. Enrichment is worth adding once volume makes manual verification impractical.
What should happen to an F-grade lead?
Give it a defined destination rather than deleting it. Most F grades belong in self-serve onboarding, a low-frequency newsletter, or a suppression list if they are competitors. Keeping the record is useful because attributes change: a two-person company that grades F today may be a thirty-person company in eighteen months, and a scheduled re-grade will catch it.