Account tiering in Customer Success is the practice of segmenting your customer base into bands, then deliberately spending different amounts of CSM time on each band. It is not the same as ABM tiering on the acquisition side — there you tier prospects by buying intent; here you tier existing customers by the value of keeping them and the effort it takes. The output of the framework is a touch model and a CSM ratio per tier, so every account has a defined service level and every CSM has a defined book.
Tier on two axes: value and effort
Tier each account on a value axis and an effort axis, then let the combination decide the touch model.
- Value is what you lose if the account churns and what you gain if it expands. Use ARR as the base, but weight it: a logo with high strategic value (reference account, marquee brand, expansion-into-parent-company potential) ranks above its raw ARR. NRR potential belongs here too — a $40K account growing 40% a year outranks an $80K flat one.
- Effort is how much CSM time the account consumes to hit its target outcome. A technically complex deployment, a low-maturity buyer, a multi-stakeholder org, or a contract up for a risky renewal all raise effort. High value plus high effort is your scarce-attention quadrant; high value plus low effort is where managed automation pays off most.
Plot the two axes and the tier falls out. Do not tier on ARR alone — that puts a self-sufficient $200K account in the same bucket as a $200K account three weeks from churning, and they need opposite things.
The standard three tiers, with calibrated ratios
| Tier | Touch model | CSM ratio (accounts : 1 CSM) | Book of business |
|---|---|---|---|
| Tier 1 — high touch | Named CSM, scheduled QBRs, custom success plan, exec sponsor | 8–20 | $1.5M–$4M ARR |
| Tier 2 — mid / low touch | Named CSM, pooled for some tasks, quarterly check-ins, templated plays | 40–80 | $2M–$5M ARR |
| Tier 3 — tech touch | No named CSM, pooled queue, automated lifecycle + in-app | 200–500+ | digital, not headcount-bound |
These ratios are B2B SaaS norms for a $20K+ average ACV; they compress for complex enterprise products (Tier 1 can run 5–8 accounts per CSM) and stretch for self-serve-adjacent ones. The book-of-business column is the sanity check: if your Tier 1 ratio implies a $9M book per CSM, the ratio is wrong, not ambitious.
Map the touch model to the tier
The tier is meaningless until it changes what the CSM actually does. Make each tier a different motion, not the same motion at a different cadence:
- Tier 1 (high touch): human-led, account-specific. A mutual success plan, scheduled QBRs, a named exec sponsor, proactive risk reviews. The CSM knows the org chart. This is where you run book-of-business planning by hand.
- Tier 2 (low touch): human-led but templated and partly pooled. Lifecycle plays trigger on health-score changes; the CSM intervenes on exception, not on schedule. One-to-few webinars replace one-to-one QBRs.
- Tier 3 (tech touch): digital-led. Automated onboarding, in-app guidance, email nurtures, a pooled CSM queue that catches only the accounts a customer health score flags. No account gets a calendar invite by default.
A platform like Gainsight, Vitally, Planhat, or Totango exists to run exactly this: tier as an account attribute, health scores per tier, and playbooks that fire automatically for Tier 3 while surfacing a prioritized queue for Tier 1 and 2. The tool does not decide the tiers — you do — but it enforces them.
Worked example
A 600-customer base, $24M ARR, six CSMs. Tier 1 = top 60 accounts by value-and-effort ($12M ARR, 10 per CSM = 6 CSMs’ worth at high touch). That already consumes the whole team, so Tiers 2 and 3 must be digital-led or you need more headcount. The framework just told you a hiring decision: to give Tier 2 a named CSM, you need roughly three more heads, or you move Tier 2 to tech touch and accept the lower retention it implies. That trade-off is the entire point of tiering.
Common pitfalls
- Tiering on ARR only. Pairs the wrong service level with the wrong account. Guard: always include the effort axis and a strategic-value weight; review any account where ARR-rank and tier disagree by more than one band.
- Static tiers. A once-a-year tiering wastes high touch on stabilized accounts and starves newly-risky ones. Guard: recompute tier eligibility on a schedule (monthly) driven by the health score, with explicit promotion/demotion rules, not a planning offsite.
- Tier 1 inflation. Every AE wants their logo in Tier 1; the ratio quietly creeps to 30:1 and high touch becomes medium touch for everyone. Guard: cap Tier 1 headcount as a hard percentage of the base (commonly 10–15%) and force a demotion for every promotion.
- Same playbook, different cadence. If Tier 3 is just Tier 1 with fewer meetings, you have not built a tech-touch motion — you have under-served Tier 1’s playbook. Guard: each tier must name at least one play the other tiers do not run.
Related
- Tech touch vs high touch — the touch models each tier maps to
- Customer health score — the signal that drives promotion and demotion
- Customer segmentation — the broader segmentation the tiers sit inside
- Book of business — how a tier’s ratio becomes a CSM’s portfolio
- NRR vs GRR — the retention metrics each tier is optimized against