Retention, Churn Mitigation & Lifecycle Email AutomationPlaybook4 min readUpdated September 2026

How to Build an Account Tiering Model That Matches Your ICP

Most account-based marketing programs fail quietly: every account gets the same email cadence, the same ad budget, and the same sales attention, regardless of whether it looks anything like your best customers. Tiering fixes that by forcing a decision, upfront, about which accounts deserve a custom play and which ones get a lighter, automated touch.

This is a framework for building that tiering model from scratch, not a theory of account-based marketing. It covers how to pick fit criteria, how to score intent without over-engineering it, and how to keep the model from rotting six months after you build it.

Vendors Covered in this Article

Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.

How do you build an account tiering model from closed-won deals?

Before you touch a spreadsheet, pull your last 20 to 30 closed-won deals and list what they had in common: employee count, industry, tech stack, the trigger event that got them looking, and who championed the deal internally. Don't guess at your ideal customer profile from a positioning document. Read the deals. Patterns that show up in the data (a certain headcount band, a specific job title as champion, a shared vendor they were replacing) are worth more than any assumption your team has been repeating in meetings.

Rank the traits you find by how consistently they show up, not by how interesting they sound. A trait present in nearly all of your closed-won deals belongs in Tier 1 criteria. A trait present in only a handful of them is a nice-to-have, not a filter.

How many tiers should an account tiering model have?

Three tiers is enough for almost every team this size: Tier 1 gets one-to-one outreach and custom content, Tier 2 gets one-to-few plays grouped by segment, and Tier 3 runs on automation alone. Five or six tiers sound rigorous but usually just mean nobody can tell which accounts belong where, and the extra tiers get ignored within a quarter.

Define each tier with a short, testable rule your whole team can apply without asking you. If a rep can't look at an account and tell you its tier in ten seconds, the rule is too complicated.

Score intent separately from fit

Fit tells you whether an account looks like your best customers. Intent tells you whether they're actually in a buying window right now. Conflating the two puts a perfect-fit account that's dormant in the same bucket as a perfect-fit account that just visited your pricing page three times this week, and that's a mistake: the second one needs a call today, not next month.

Track a short list of intent signals: website visits to high-intent pages, contact with your AI advisor or chat tool, job postings for roles your product supports, and, where you have the data, third-party intent signals from your outbound tools. A B2B contact database like Apollo can surface company-level intent and technographic signals alongside contact data, which is useful for confirming a tier assignment before your team spends real time on an account.

Match outreach intensity to tier, not to who asks loudest

Tier 1 accounts get named contacts, personalized subject lines, and a human check-in before any automated sequence fires. Tier 2 accounts run through templated but segmented sequences, grouped by industry or use case, with a real person reviewing replies. Tier 3 accounts run on full automation: a nurture sequence, retargeting ads, and no manual review unless they reply.

A cold-email tool built for deliverability and personalization at scale, such as lemlist, fits the middle and bottom tiers well: it can vary subject lines and first lines by segment without a person hand-writing each one. Save your team's actual writing time for the accounts where a generic email would read as an insult.

Re-tier accounts on a schedule, not when someone remembers

Tiering models decay. An account that was Tier 3 last quarter might have hired a new VP who fits your ICP exactly, or a Tier 1 account might have gone quiet for six months and no longer deserves the attention it's getting. Put a recurring review on the calendar (monthly for Tier 1, quarterly for the rest) and use it to move accounts up or down based on new signals, not gut feel.

Steps to build the tiering model:

  1. Pull your recent closed-won deals and list what they share, including company size, industry, tech stack, trigger event, and internal champion.
  2. Set three tiers: one-to-one outreach for Tier 1, segmented one-to-few plays for Tier 2, and automation alone for Tier 3.
  3. Score intent separately from fit, so a dormant perfect-fit account does not share a bucket with one actively in a buying window.
  4. Match outreach intensity to the tier, with named contacts and human check-ins for Tier 1 and automation for Tier 3.
  5. Put a recurring re-tiering review on the calendar so the model does not decay or bend to the loudest internal opinion.

Mistakes that quietly break a tiering model

The most common failure is letting the loudest internal opinion override the criteria: a rep insists an account is strategic, it gets bumped to Tier 1 without meeting the fit rules, and within a few cycles the tier has lost its meaning because half the accounts in it don't actually match the pattern. Hold the line on the documented criteria, and if an account genuinely deserves an exception, note why in writing rather than quietly stretching the rule.

The second common mistake is building the model once and never revisiting the underlying fit criteria themselves, not just the account list. If your product, pricing, or ICP shifts, the traits that predicted a good account six months ago may no longer apply. Revisit the criteria themselves, not just which accounts sit where, at least once a year.

Executive Capability Standard

What Good Looks Like

Good tiering means every account in your pipeline has a documented tier, a clear rule for why it's there, and an outreach plan that matches that tier, reviewed on a fixed schedule rather than left to drift.

Building The Capability (5-Stage Skill Ladder)

1. Learn:Read through 20 to 30 closed-won deals and write down the traits they share before building any tiers.
2. Do Manually:Build the tier list in a spreadsheet, score each account by hand against your fit and intent criteria, and update it during a weekly pipeline review.
3. Delegate:Hand the scoring and review cadence to a marketing operations hire or an outsourced RevOps contractor, with the criteria and review schedule documented so it doesn't rely on one person's memory.
4. Automate:Push firmographic and intent data into your CRM so tier assignment updates automatically as new signals come in, with a human check before an account moves into Tier 1.
5. Buy:Bring in a B2B data and intent platform that scores accounts against your defined criteria and syncs the result into your CRM and outbound tools.

How to Get Started

Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.

Frequently Asked Questions

How many accounts should be in Tier 1?

Keep it small enough that your team can genuinely personalize outreach for every account on the list. If your reps can't name the champion and the trigger event for each Tier 1 account off the top of their head, the tier has grown past what one-to-one attention can cover, and it's time to either add headcount or move some accounts down.

Should marketing or sales own the tiering model?

Both need to agree on it, but one team should hold the pen. Marketing usually owns the model because it drives content and ad spend decisions, while sales validates the criteria against what they see in live deals. If the two teams keep separate lists, accounts fall through the gap between them.

What if we don't have enough closed-won data yet?

Use your best current customers as a proxy, even if the sample is small, and lean more heavily on qualitative signals: who champions your product internally, what triggered their search, and what almost stopped the deal. Revisit the model after your next ten closed deals rather than waiting for a large dataset that may take a year to build.

About the numbers

This guide doesn't quote a sourced benchmark. Figures in it are estimates or general guidance, so check them against your own numbers.

Related Guides