Account Tiering
Account tiering is the practice of grouping target companies into ranked segments (such as Tier 1, 2, and 3) based on their fit and revenue potential, so teams invest effort proportionally.
Key takeaways
- Tiering turns a continuous score into groups so each one triggers a different motion.
- Run as many tiers as you have genuinely different treatments, usually two or three.
- Place each boundary where capacity for that motion runs out, not where statistics suggest.
- Weight signals by how well they predicted closed revenue, not by how available they are.
- Wide deal-size variation inside one tier means value is missing from the boundary rule.
In depth
Tiering converts a continuous fit score into a small number of discrete groups, and the value comes from the discreteness. A score of 71 and a score of 68 are indistinguishable in practice, but a boundary between them forces a decision about which motion each account receives. Every tier is a bundle rather than a rank: a contact cadence, a content type, an ownership model and a response commitment. The score orders accounts, while the tier determines what actually happens to them.
The number of tiers should match the number of genuinely different motions you can run, which for most teams is two or three. A fourth tier without a fourth distinct treatment produces a label with no behaviour attached. Boundary placement is a resourcing question rather than a statistical one: draw each line where the capacity for that motion runs out. Signals feeding the score deserve weighting by how well they have predicted closed revenue, not by how easily they can be collected.
In practice tiering runs as a scheduled job over the account table, combining firmographic fields, product usage where it exists, and recent intent or engagement signals into a score, then applying the boundaries and writing a tier field the routing rules read. A scorecard quiz supplies the inputs no database holds, such as the current process, the timeline or who owns the budget, so a declared answer can move an account between tiers immediately after submission.
Tiering assumes the signals that predicted past wins still predict future ones, which fails after a pricing change, a product shift or entry into a new segment. It also compresses information, because two accounts in the same tier can need entirely different conversations and a rep who treats the tier as a script loses that distinction. Where deal sizes vary by an order of magnitude inside one tier, the boundaries are probably drawn on fit alone and should include expected value.
Example in practice
How to measure it
The basic test is monotonicity: win rate, average deal size and opportunity rate should improve as you move up the tiers. If Tier 2 outperforms Tier 1 on any of them, either the boundary or the signals behind it are wrong. Read each metric separately, because a tier can win on deal size while losing on win rate, and that pattern argues for different treatment rather than a different tier.
Also track movement between tiers at each refresh: how many accounts changed and what triggered the change. Almost no movement means the model is reading static fields and ignoring behaviour entirely. Very high movement means the signals are noisy and reps cannot plan against them. Where a quiz feeds the score, check how often a submitted answer alone changes an account's tier.
Common mistakes
A common failure is scoring on data that is merely available. Headcount and industry code dominate the model because every record carries them, while the signals that actually separated past wins from losses, such as a specific role being present or a process already in place, get left out because they need enrichment or a form question. The result ranks accounts confidently and predicts nothing. Test each signal against closed revenue before weighting it.
The second is tier inflation. Every account someone cares about drifts upward until most of the list sits in the top two tiers and the lower ones are empty, which restores exactly the undifferentiated queue tiering existed to break. Cap each tier by capacity so promoting one account requires demoting another. That constraint forces the conversation about relative priority that a scoring model alone will never produce.
Frequently asked questions
How many tiers should a tiering model have?
As many as you have distinct motions to apply, which for most teams means two or three. Each tier must differ in something a rep or a campaign actually does: cadence, ownership, content type or response commitment. A fourth tier receiving the same treatment as the third is only a label, and labels with no behaviour attached quietly stop being maintained.
Which signals should feed the tier score?
Whichever ones separated past wins from past losses in your own data, weighted by how strongly they did so. That usually means a mix of firmographics, a technology or process indicator, and a recent engagement signal. Resist adding a field simply because a data provider supplies it, since every unpredictive signal dilutes the weight of the ones that work.
How often should tiers be reviewed?
Rescore on a regular cadence, commonly monthly or quarterly, and let trigger events move an account immediately. The cadence should be slower than campaign planning so audiences stay stable, but faster than the sales cycle so no account sits in the wrong tier through a full evaluation. Record what changed at each refresh so drift is visible.
Should tiers be based on fit or on intent?
Both, but keep them visible separately. Fit says whether the account is worth winning; intent says whether now is the moment. Collapsing them into one number hides which is driving a placement, so a high-fit account with no activity looks identical to a poor-fit account showing interest, and those two situations call for opposite responses.
How does account tiering relate to lead scoring?
Tiering scores the company while lead scoring scores the person and their behaviour. A senior contact at a Tier 3 account and a junior contact at a Tier 1 account call for very different routing, and you need both dimensions to decide. Most routing rules read the tier first for ownership, then the lead score for urgency.
Can a quiz change an account's tier?
Yes, and that is one of the stronger reasons to run one. Quiz answers supply process, timeline and budget-ownership facts no database holds, so a submission can move an account up when it reveals fit the firmographics missed, or down when the workflow your product depends on turns out to be absent. Apply the change before routing evaluates.