Tech-touch and high-touch are the two ends of the CS coverage spectrum. High-touch means a named CSM owns the account, runs QBRs, and builds a success plan; tech-touch means the account is served by automation — in-app guides, lifecycle emails, and a pooled CSM queue — with no named owner. The decision rule is simple to state and hard to hold: an account gets high-touch only when the ARR it carries (or the expansion it can unlock) covers the fully loaded cost of the CSM time it consumes.
The coverage spectrum
Most teams run three or four bands, not two. The cleanest framework is a spectrum from fully automated to fully bespoke:
- Tech-touch (no named CSM). Served by product: onboarding flows, in-app walkthroughs, lifecycle email, a help center, and a shared inbox staffed by a pooled team. One pooled CSM might “cover” 500-2,000 accounts. Cost-to-serve per account runs in the tens of dollars per year.
- Low-touch / scaled (digital-led, human-assisted). A named CSM owns 150-400 accounts but engages reactively and at milestones, not on a fixed cadence. Most of the journey is automated; the human steps in on risk signals and renewals. Cost-to-serve in the low hundreds.
- High-touch (named CSM, fixed cadence). A CSM owns 20-80 accounts, runs scheduled QBRs, builds success plans, and is the named relationship owner. Cost-to-serve runs into the thousands per account per year.
- Strategic / white-glove. The top handful of accounts get a CSM managing 5-15 logos, often with a solutions architect or exec sponsor attached. Cost-to-serve in the tens of thousands.
The cost-to-serve math
The decision turns on one ratio: ARR-per-CSM relative to fully loaded CSM cost. A fully loaded CSM in North America costs roughly $150K-$220K per year (base, benefits, tooling, overhead). A common target is that each CSM manages $2M-$5M of book in high-touch, and $5M-$15M in low-touch. That sets the floors.
Worked example. A CSM costs $180K fully loaded, and you want CS to run at roughly a 1:10 cost ratio — CS spend no more than 10% of the ARR it touches. That means each CSM must touch at least $1.8M of book. Split it by segment:
- A high-touch CSM with 40 accounts needs each account to average $45K ARR to clear the $1.8M floor.
- A low-touch CSM with 250 accounts clears the same floor at $7.2K ARR per account.
- Below ~$5K ARR, no human cadence pays back at a 1:10 ratio — these accounts belong in tech-touch, full stop.
The bands fall out of the math. Set your cost ratio, divide by the book size each model supports, and the ARR thresholds that separate the bands compute themselves. Run this per segment, because a $5K account in a self-serve PLG motion and a $5K landing-deal in an enterprise account behave nothing alike.
Hybrid is the real answer
Pure tech-touch and pure high-touch are both traps. Pure tech-touch leaves expansion on the table in accounts that would grow with one human conversation a quarter. Pure high-touch burns CSM time on accounts that a well-built onboarding flow would serve better and faster.
The model that works is digital-led across the whole base, with human time allocated by signal. Every account — including the $2M strategic logo — gets the automated layer: lifecycle emails, in-app guides, health scoring, and self-serve help. High-touch accounts get that plus a named CSM. The automation is the floor for everyone, not a consolation prize for small accounts. This is why teams running Vitally, Totango, or Gainsight wire the same health model into both motions and only change who gets paged on a red account.
Graduation and demotion rules
Bands are not permanent, and the graduation trigger is the single rule that moves the most in the whole system. Build it on signal, not on a quarterly account-review meeting:
- Graduate up when an account crosses an ARR threshold, shows a buying-committee expansion signal, or files a strategic use case. A tech-touch account that doubles seats and opens an API integration ticket has told you it wants more contact.
- Demote down when a high-touch account goes quiet for two cadence cycles, when the champion leaves and isn’t replaced, or when the renewal is locked multi-year and de-risked. Holding a healthy, locked account in high-touch is wasted CSM capacity that a growing account needs.
Automate the trigger. A weekly recompute that moves accounts between pooled and named ownership keeps the book aligned with reality instead of with last quarter’s org chart.
Common pitfalls
- Segmenting by logo size instead of cost-to-serve. A famous brand name on a $4K contract is still a tech-touch account. The guard: tier on ARR-plus-expansion-potential, not on how the logo looks on a slide.
- No automated layer under high-touch. When CSMs hand-build onboarding decks for every account, the model doesn’t scale and the $40K accounts subsidize the CSM’s busywork. The guard: every account, every band, sits on the same automated onboarding and health-scoring layer; the CSM adds judgment on top, not data entry.
- Static bands. Tiering once at the annual planning offsite means you carry dead high-touch accounts for a year. The guard: a weekly scoring recompute with hard graduation and demotion triggers wired into the workflow.
- Tech-touch as a dumping ground. Treating the pooled segment as accounts you’ve given up on guarantees they churn. The guard: hold tech-touch to a measurable bar — onboarding-completion rate, time-to-first-value, and GRR by band, watched the same way you watch the named book.
Related
- NRR vs GRR — the retention metrics you watch per coverage band
- Gainsight — health scoring and playbooks across both motions
- Vitally — scaled CS for product-led books
- Intercom — the in-app and lifecycle layer under tech-touch