Lead Grading vs Lead Scoring: How to Use Both

A rep opens a lead with an 85 out of 100 score and assumes it's hot. Turns out it's a college student who filled out every form on the site for a class project. The number was high. The lead was worthless. That gap is exactly what lead grading exists to close.

Most teams struggling with lead prioritization aren't missing a scoring model. They're missing the other half of the equation. Scoring tells you how engaged someone is. It says nothing about whether that person, or their company, could ever become a real customer. Grading answers that second question. Skip it, and your "hottest" leads list will keep mixing students, competitors, and job seekers in with your real buyers.

This guide covers what lead grading is, how to build an A-D model from your own closed-won data instead of opinion, how negative criteria keep bad-fit leads out of your pipeline, and how grade and score combine into a single priority call. If you haven't read our companion piece on lead scoring systems yet, start there for the engagement side of this equation. This article is the fit side, and the two are meant to be read together.

Key Facts: Lead Grading

• Sales reps spend just 40% of their time actually selling, with the rest lost to admin work, data entry, and chasing prospects that were never a fit in the first place, according to Salesforce's State of Sales report (Seventh Edition). Source
• Only 27% of B2B leads are sales-ready the moment they first engage with a company; the other 73% aren't ready for a sales conversation yet, per MarketingSherpa's 2012 B2B Benchmark Report. Source
• Reps waste an estimated 27% of potential selling time chasing bad or incomplete contact data instead of working real opportunities, according to ZoomInfo/Validity research. Source
• Only 44% of B2B companies use any formal lead scoring or grading system to separate good-fit prospects from bad ones, per Landbase's 2025 research. Source

What is lead grading?

Lead grading is the practice of assigning a letter grade, usually A through D, to a lead based on how closely it matches your ideal customer profile. It answers one question: if this person or company were ready to buy today, would we actually want them as a customer?

That's a completely different question from "how interested are they right now," which is what lead scoring measures. Grading looks at firmographic data (company size, revenue, industry), demographic data (job title, seniority, department), technographic data (what tools they already use), and geographic fit (do you actually serve their region). None of that changes because someone opened three emails this week. A 200-person fintech company is a 200-person fintech company whether its VP of Sales is actively browsing your pricing page or hasn't looked at your site in six months.

Scoring, by contrast, is built entirely from behavior: page visits, downloads, email clicks, demo requests. It moves constantly and typically decays over time. Grading is closer to a credit check than a mood ring. It's slow-moving, based on who someone is rather than what they're doing this week, and it rarely needs to be recalculated unless the company itself changes: a new funding round, an acquisition, a pivot into a new market.

Lead grading vs lead scoring: the core difference

The confusion between these two terms is understandable. Both produce a number or letter attached to a lead record, both feed into routing decisions, and both get lumped under "lead qualification" in casual conversation. But they measure orthogonal things, and collapsing them into one number is where most qualification systems go wrong.

Notice that neither column is "the qualification score." They're inputs to qualification, and qualification only gets accurate once you look at both together, a point we'll come back to in the priority matrix below. If you're still deciding where grading and scoring fit relative to your broader lead qualification frameworks, think of grading and scoring as the two raw inputs those frameworks (like MQL/SQL definitions) are built on top of.

The A-D grading model

Most B2B teams that grade leads use a four-tier letter scale, A through D, because it's simple enough for sales to trust at a glance and granular enough to be useful for routing.

Grade A leads are your textbook ICP match. Right company size, right industry, right title, right region, ideally with a technographic or growth signal layered on top. These are the accounts your best reps would draft first if they got to pick their own territory.

Grade B leads are a strong fit with one or two gaps. Maybe the company is slightly smaller than your sweet spot, or the contact is a Director instead of a VP. Still worth pursuing, just with a bit more qualification work up front.

Grade C leads have a partial fit with real gaps. They might sit in an adjacent industry, or hold a title that's an influencer rather than a decision-maker. These are usually better served by marketing-led nurture than an immediate sales call, and a good candidate for sharpening through buyer persona work.

Grade D leads are a poor fit, or they trip a disqualifier outright. Wrong geography, a personal email address, a competitor's domain, a company far outside your serviceable size range. Grade D doesn't mean "ignore them forever." It means don't let sales time go anywhere near them until something material about the account changes.

The exact point thresholds you use to draw the lines between A, B, C, and D should come from your own data, not a template. Here's how to build that.

Building your grading criteria from closed-won data, not opinion

The single biggest mistake in lead grading is designing the model around what sales assumes matters instead of what the data actually shows. Reps have strong, confidently held opinions about what a "good lead" looks like, and those opinions are frequently wrong, or at least incomplete.

Start by pulling every closed-won deal from the last 12 to 24 months. For each one, capture the firmographic and demographic attributes at the time the lead first entered your pipeline: company size, industry, title, region, tech stack, funding stage. Then pull the same attributes for a sample of closed-lost and disqualified leads, and look for the delta between the two groups.

Ask specific questions. What percentage of won deals came from companies in a certain employee-count band, versus your overall lead pool? If 70% of your revenue comes from companies with 100 to 1,000 employees but that band is only 25% of your total leads, company size is a strong predictor and should carry real weight. If industry shows almost no difference between won and lost deals, don't waste points on it just because it feels intuitively important.

This is the same discipline behind building a defensible ideal customer profile: let the pattern in your actual customer base define the criteria, then translate that pattern into points.

Sample grading criteria and weights

Once weighted, map the resulting point total onto your letter scale, simple enough that a rep can look at a record and understand why it landed where it did.

Point-to-grade mapping

Negative grading criteria: your disqualifiers

Positive attributes tell you who to chase. Negative attributes, sometimes called disqualifiers, tell you who to stop chasing regardless of how engaged they look. This is the part of grading most homegrown models skip, and it's usually the reason a "high scoring" lead turns out to be a dead end.

Common disqualifiers include competitor email domains, personal or free email addresses on a form that should only be used by business buyers, company sizes far outside your serviceable range, geographies you don't sell into, and job titles that indicate no purchasing authority (students, job seekers, recruiters browsing a careers page). Each of these should subtract enough points to guarantee a Grade D outcome, or set the grade to D directly regardless of what else is on the record. Don't let a strong positive score offset a hard disqualifier. If someone is a competitor, no amount of pricing page visits changes that.

Negative criteria also protect your model's credibility with sales. Reps stop trusting a grading system the first time it hands them an obviously wrong lead, and a clean disqualifier layer removes those embarrassing misses before a human ever sees them.

Combining grade and score: the priority matrix

Here's where grading earns its keep. On its own, a grade tells you who's worth pursuing eventually. On its own, a score tells you who's active right now. Neither one, by itself, tells you what to do today. Put them on two axes and you get a matrix that does.

Plot grade (A through D) on one axis and engagement tier on the other, moving from hot (actively engaging now) to cold (no recent activity). Every lead lands in exactly one cell, and that cell should map to a specific action.

Two cells in this matrix matter more than the rest, because they're where most homegrown systems fail.

High fit, low engagement (top-right of the A and B rows) is your nurture zone. These are accounts you genuinely want as customers, but they aren't showing active buying signals yet. Often they are simply early in the buyer journey, researching a problem they have not yet committed to solving. Don't hand these to sales for a cold call; that burns goodwill on an account you'll want later. Route them into a longer-cycle nurture track and revisit as engagement builds through your lead nurturing programs.

Low fit, high engagement (top-left of the D row) is the classic trap. This is the lead that looks amazing in a single-dimension score, because they've filled out every form and clicked every link, but they're a student, a competitor doing research, or simply a company you can't service. If your qualification process only looks at engagement, this lead sails straight to a rep's queue and wastes real selling time. If it only looks at fit, you might miss that this cell occasionally does contain a legitimate account that outgrew your assumptions, worth a quick manual glance rather than an automatic suppress. That's why "flag for review" beats "auto-route" for this cell: a human check costs a minute, a wasted sales call costs a lot more.

Routing and SLAs by grade-score combination

Once every lead has a position in the matrix, you can turn that position into a concrete operational rule: response time and owner. This is where grading stops being an analytical exercise and starts changing what happens in the first five minutes after a lead comes in, which is exactly the moment lead response time has the biggest effect on conversion.

Feed this table straight into your lead routing logic and your lead status management workflow so a lead's grade and engagement tier automatically set its stage, owner, and clock. That's also where MQL vs SQL definitions become concrete instead of debatable: an MQL, in practice, is usually a lead that's cleared a minimum grade and engagement threshold, and an SQL is one a rep has personally validated on top of that.

Manual vs rule-based vs predictive grading

Just like lead scoring, grading can be built three different ways, and the right choice depends on how much historical data you have and how much you need the model to be explainable.

Rule-based grading is the right starting point for almost every team here. It's transparent, fast to build once you've done the closed-won analysis above, and reps can see exactly why a lead landed where it did, which matters for adoption. Predictive grading is worth the investment once you have enough volume that a model can find patterns a spreadsheet can't, but it should layer on top of a rule-based foundation, not replace it on day one.

Recalibrating as your ICP shifts

A grading model built today is a snapshot of who your best customers were over the last year or two. That snapshot goes stale the moment your product, market, or go-to-market motion changes, and most teams don't notice until sales starts complaining that "A-grade" leads have stopped converting.

Set a quarterly review as the default cadence, and treat these as triggers for an earlier look: a new product line that opens up a different buyer, a pricing change that shifts your company-size sweet spot, a shift in go-to-market motion (moving upmarket, launching self-serve), or two or three consecutive quarters where Grade A leads convert at a noticeably lower rate.

The recalibration process is the same closed-won analysis you ran the first time, just re-run on the most recent cohort of deals. If company size stopped predicting outcomes but title seniority became more predictive, shift the points accordingly. Don't be afraid to move grade thresholds too. If 90% of your leads are suddenly landing in Grade A because the market shifted toward your ICP, your thresholds have gotten too loose to be useful for prioritization.

Common lead grading mistakes to avoid

Grading on gut feel instead of data. "We know a good lead when we see one" is exactly the assumption that closed-won analysis is meant to test, and it's wrong more often than sales teams expect. A gut-feel grade also can't be audited, explained to a new hire, or improved systematically.

Never revisiting thresholds. A grading model isn't a one-time project. Markets shift, products evolve, and a model nobody has touched in eighteen months is almost certainly misclassifying leads it would have graded correctly a year ago.

Letting engagement override fit. This is the low-fit, high-engagement trap from the priority matrix, and it's the single most common way bad-fit leads end up wasting sales time. A high score should never be enough, on its own, to route around a hard disqualifier.

Grading on data you don't reliably collect. If your form only captures email and company name, don't build a grading model that depends on accurate job title and employee count for half its weight. Either invest in lead data enrichment to fill those gaps reliably, or build the model around the attributes you actually have clean, consistent data for. A model built on spotty data produces spotty grades, and spotty grades are worse than no grades at all because they create false confidence.

Where to go from here

Grading and scoring are two halves of the same qualification system, not competing methods. Once both are in place, the rest of your lead management stack gets sharper: routing decisions stop relying on a single ambiguous number, types of leads become easier to define with fit and intent as separate variables, and even product-led motions benefit, since a product qualified lead still needs a fit check before it's handed to a rep, not just usage data.

Start with the closed-won analysis, build a simple rule-based grade, layer it against your existing engagement score, and route from the matrix rather than a single blended number. The model won't be perfect on day one. It just needs to be better than a rep's gut, and a data-built A-D grade almost always is.