Most SaaS companies segment customers by ARR band and stop there. It's the easiest axis to compute and the weakest one to act on — a $50k account that's barely using the product and a $50k account that's expanding fast look identical on a revenue-sorted list, but they need completely different treatment. Real segmentation combines several axes and uses the result to decide where CS time actually goes.
The axes that actually predict what an account needs
ACV/tier is the obvious starting axis and a legitimate one — it sets the ceiling on how much white-glove attention an account can justify economically. But it's a ceiling, not the whole answer. Usage maturity — how deeply and how long an account has used the product, whether they're still in initial rollout or fully embedded into daily workflow — tells you whether an account is at risk of never reaching value or comfortably past that point. Industry vertical matters because a healthcare or financial-services customer has compliance and workflow needs a generic playbook won't address, and often needs different onboarding content entirely. Growth potential / expansion propensity — seat growth trajectory, usage against plan limits, org headcount trends — identifies accounts worth investing CS time in not because they're at risk, but because the return on that time is a bigger contract, not just a saved one.
ACV is convenient because it's already in your billing system. Usage maturity and expansion propensity require pulling product telemetry into the CS tooling. The second kind is more work to set up and does most of the actual predictive work.
Why a single CS motion fails across a real customer base
A single CS motion applied uniformly means one of two things happens: either every account gets the high-touch treatment, which doesn't scale past a few hundred customers and buries the team, or every account gets the same light-touch treatment, which starves the accounts that would justify — and need — deeper investment. Neither serves the business. A 200-seat enterprise account and a 3-seat self-serve account should never be running through the identical cadence of check-ins, QBRs, and renewal conversations; the cost structure and the expected value don't match.
Tiered CS models: high-touch, tech-touch, low-touch
High-touch accounts get a named CSM, scheduled QBRs, proactive outreach ahead of renewal, and involvement in implementation decisions — reserved for accounts where ACV and/or expansion potential justify the headcount cost. Tech-touch accounts get automated check-ins, in-app guidance, and a CSM who steps in only when a health signal fires (usage drop, support escalation, renewal window) rather than on a fixed cadence — this is where most of the "middle" of a customer base should live. Low-touch / self-serve accounts run almost entirely on product-led automation: in-app onboarding, help docs, triggered emails, with human involvement reserved for inbound support tickets. The tier isn't fixed at signup — an account can move from tech-touch to high-touch when its expansion signals justify the reassignment, or from high-touch down when a champion leaves and usage flatlines.
Accounts get segmented once at onboarding and rarely reassessed. A tech-touch account that's tripled its seat count in a year should have moved to high-touch six months ago. Review tier assignment on a regular cycle, not just at signup.
How segmentation should actually drive resource allocation
The segmentation model earns its value at the point of assigning CSM headcount and building playbooks — not as a reporting exercise. Book-of-business sizing (how many accounts per CSM) should differ by tier: a high-touch CSM might carry 15-30 accounts, a tech-touch CSM might own several hundred with automation covering the routine cadence. Playbooks should be tier-specific too — a high-touch onboarding playbook involves a kickoff call and a named implementation contact; a tech-touch playbook is a sequence of in-app milestones with a CSM alert only on deviation.
{
"segment": "tech_touch",
"criteria": {
"acv_usd": { "min": 5000, "max": 50000 },
"usage_maturity": ["ramping", "established"],
"expansion_score": { "min": 40 }
},
"csm_model": "pooled",
"cadence": "triggered_only",
"escalate_to_high_touch_if": { "expansion_score": { "min": 75 } }
}The mistake of segmenting on revenue alone
Revenue-only segmentation misses the accounts most worth acting on in both directions. A high-ACV account with declining usage looks fine on a revenue-sorted list right up until it churns — usage maturity would have flagged it months earlier. A modest-ACV account with fast-growing usage and headcount looks unremarkable on revenue alone but is exactly the account worth a proactive expansion conversation. If usage and growth signals aren't part of the segmentation model, the CS team is triaging on a lagging indicator and finding out about risk and opportunity at the same time — when the renewal or the expansion conversation is already happening.
| Tier | CS model | Typical book size |
|---|---|---|
| High-touch | Named CSM, scheduled cadence | 15-30 accounts |
| Tech-touch | Pooled CSM, triggered outreach | 100-300 accounts |
| Low-touch / self-serve | Automation + reactive support | Unbounded, no dedicated CSM |
Segmentation is only useful if it changes what your CS team does differently per group. Build it from ACV plus usage maturity plus growth signal, review tier assignment on a cycle rather than freezing it at signup, and treat the tiers as the basis for headcount and playbook design — not as a label that sits unused in a CRM field.
Independent software engineer in Nairobi specialising in Acumatica customisations, Laravel backends, and tax fiscalisation integrations across East and Southern Africa.