B2B account scoring: how to prioritize the accounts most likely to buy
A complete, implementable guide to scoring target companies on fit and readiness — the criteria, the method, the mistakes, and how it differs from lead scoring.
Last updated · Written by the RevSyt team at Meta-Insyt LLC
What is B2B account scoring?
B2B account scoring is the practice of rating each target company on two things — how well it fits your ideal customer profile, and how ready it looks to buy right now — and expressing that as a single comparable number with the evidence behind it.
A score is only useful if it changes what a rep does next. The point is not the number; it is the ordering it produces and the reasoning it exposes. A good account score answers three questions at once: should we sell to this company at all, should we sell to them this quarter, and what would we say if we called them today.
Why account scoring matters
Outbound capacity is finite. A rep who can hold roughly a hundred live conversations a quarter cannot work a five-thousand-row list evenly, so some ordering happens whether or not anyone designs it — usually alphabetical, by list upload date, or by whoever answered the phone last. Account scoring replaces an accidental ordering with a deliberate one, and makes the reasoning reviewable: when a rep disagrees with a score, the argument is about the evidence rather than about a hunch.
Scoring also makes territory and campaign decisions comparable across people. Two SDRs working the same segment with the same criteria produce lists you can merge, audit and hand over.
Account scoring criteria
Most working models combine six dimensions. Each is scored independently so a strong signal in one place cannot silently hide a disqualifier in another.
| Dimension | What it measures | Typical evidence |
|---|---|---|
| Industry fit | Whether the company operates in a sector where your product solves a real, recurring problem | Company description, SIC/NAICS classification, product lines, regulatory context |
| Company size fit | Whether headcount, revenue or site count sits inside the band your product is built and priced for | Employee counts, filings, official "about" pages, location listings |
| Technology fit | Whether their current stack makes your product a natural addition, a replacement, or a poor match | Job posts naming tools, engineering blogs, partner and integration pages, public case studies |
| Growth signals | Whether the company is expanding in a way that creates budget and new requirements | Hiring volume and role mix, new locations, funding or expansion announcements |
| Buying triggers | Discrete recent events that create a reason to act now | Leadership changes, M&A, migrations, incidents, compliance deadlines, product launches |
| Urgency | How time-bound the trigger is — whether a window is closing | Announced deadlines, contract or renewal timing, stated roadmap dates |
Industry fit
Score the problem, not the label. "Manufacturing" is too broad to act on; "discrete manufacturers running multi-site operations with their own IT team" is a criterion a researcher or a model can actually verify. Write each industry criterion so that two people looking at the same company would score it the same way.
Company size fit
Size bands should reflect how you sell, not how the market is usually segmented. If your product needs an owner with budget authority and no procurement committee, a 300-person company may score higher than a 30,000-person one even though the larger firm has more seats.
Technology fit
Technographic evidence is the most misused input in scoring because it is easy to over-read. A job post mentioning a tool is evidence that the tool is in use; it is not evidence of contract size, satisfaction, or renewal date. Score what the evidence supports and record the rest as an open question for discovery.
Growth signals
Growth matters because it creates new requirements and unspent budget. The most reliable public growth signals are sustained hiring in the function you sell into, new locations or entities, and announced expansion. A single job post is weak; a pattern over a quarter is meaningful.
Buying triggers
Triggers are events, not attributes. A new CIO, a migration announcement, an acquisition, a publicized outage, or a compliance deadline all change who is willing to take a meeting this month. Triggers should carry a date, and a trigger with no date should be treated as weaker than one you can place in time.
Urgency and account momentum
Urgency asks whether the window is closing. Momentum asks whether the signals are accumulating or going quiet: three related events in a quarter say something different from one event eighteen months ago. Momentum is the dimension most often left out and the one that most changes sequencing.
Account scoring vs lead scoring
| Account scoring | Lead scoring | |
|---|---|---|
| Unit | The company | The individual person |
| Main inputs | Firmographics, technographics, public events | Form fills, email opens, page views, demo requests |
| Works before contact? | Yes — it is built for cold lists | No — it needs engagement to exist first |
| Best use | Choosing which companies to work at all | Routing and timing follow-up on inbound interest |
| Failure mode | Fit criteria written too loosely to verify | Rewarding engagement from people who cannot buy |
The two are complements. Account scoring decides where to spend outbound effort; lead scoring decides how fast to respond to a hand raised inside an account. A team that only does lead scoring will over-invest in whoever happened to visit the pricing page.
A scoring methodology you can implement
- Write the ICP as testable criteria. Each line should be checkable against public evidence, with a clear pass, partial or fail.
- Choose your dimensions and weights. Start with equal weights across fit dimensions and slightly lower weights on trigger and urgency until you can validate them.
- Define the evidence bar per dimension. State what counts as verified evidence, what counts as weak evidence, and what counts as inference.
- Score a calibration set. Take ten closed-won, ten closed-lost and ten never-responded accounts and score them blind. If won and lost accounts score the same, your criteria are not discriminating.
- Set thresholds, not just numbers. Decide the score at which an account enters a sequence, gets research, or is parked. A score with no action attached is a report, not a system.
- Re-score on a cadence. Refresh fit quarterly and triggers on your prospecting cadence.
- Review disagreements. When a rep overrides a score, log why. Repeated overrides in one direction are a weighting problem, not a discipline problem.
Worked example
Suppose you sell backup and disaster recovery software to mid-market managed service providers. An account with 250 employees (size fit: strong), classified as an MSP (industry fit: strong), publicly listing a competing backup vendor on its services page (technology fit: replacement opportunity, medium), hiring three cloud engineers this quarter (growth: medium), and announcing a new regional datacenter last month (trigger: strong, dated) scores far above an equally sized MSP with no dated events. The second company is not disqualified — it is simply not this month's call.
Common mistakes
- Scoring on availability of data rather than relevance. Fields are easy to collect and often irrelevant; the fact that you have a data point is not a reason to weight it.
- Presenting inference as fact. "Likely evaluating a replacement" is a hypothesis. Label it as one.
- Stale evidence. An eighteen-month-old trigger is history, not intent.
- One number with no breakdown. If a rep cannot see why an account scored 82, they will not trust the 82.
- Never re-calibrating. Markets and ICPs move; a model no one has revisited in a year is describing last year.
How RevSyt applies this
RevSyt builds an intelligence profile for the technology you sell, then researches each account against it and scores the six dimensions above. Every score comes with the evidence found, the buying signals detected, a target persona drawn from real roles at the company, and a specific outreach angle. Facts that could be verified in a source are kept visually separate from AI inference, and research recency is tracked rather than assumed. The RevSyt account scoring framework documents the model in detail, and the research freshness framework explains how source dates are handled.
Frequently asked questions
What is B2B account scoring?
B2B account scoring is the practice of rating each target company on how well it fits your ideal customer profile and how ready it appears to buy, so a sales team can rank a list instead of working it in arbitrary order.
What is the difference between account scoring and lead scoring?
Lead scoring rates an individual person, usually on their engagement with your marketing. Account scoring rates the whole company on fit and readiness, using firmographic, technographic and event-based evidence rather than one person's clicks.
What criteria should an account score use?
Most working models combine industry fit, company size fit, technology fit, growth signals, buying triggers and urgency. Each dimension should be evidence-based and independently explainable.
Do you need historical closed-won data to score accounts?
No. Historical data improves weighting, but a first model can be built from a clearly written ideal customer profile and public evidence about each account. Weights can be tuned once you have outcomes to compare against.
How often should account scores be refreshed?
Fit dimensions change slowly and can be refreshed quarterly. Trigger and urgency dimensions depend on recent events and should be refreshed on the cadence your team prospects — weekly or monthly for active territories.
How does RevSyt score accounts?
RevSyt researches each company against the technology you sell, scores six dimensions, and returns a 0–100 total with the evidence behind each dimension, a target persona and an outreach angle. Verified facts are kept separate from AI inference.