Case study · Pragma · 2024–present
From two CSMs to the P&L.
How I built Pragma's post-sales organization from a two-person account team into a 15-person, three-function global group — while the customer base grew 7x, a second product launched, and the games industry went through its hardest stretch in a decade.
Figures from Pragma's internal measurement systems, mid-2026. Some details generalized for confidentiality.
01
The setup
Two CSMs, one product, no systems.
When I took over customer success at Pragma in February 2024, the function was two CSMs supporting roughly fifteen game studios on a single product — Pragma Engine, a backend platform for live-service games. There was no technical services function, no professional services motion, no support tiers or SLAs, no EMEA presence, and no systematic view of account health or churn risk.
Over the next two and a half years, the business got dramatically harder and bigger at the same time. The company launched a second product — FirstLook, a playtesting and community platform — that needed a scaled, high-volume support motion completely unlike Engine's high-touch enterprise model. The customer base grew roughly 7x. And the games industry went through its most brutal funding contraction in a decade, with studios shutting down mid-contract.
The organization had to be built for two opposite motions at once — and pay for itself along the way.
02
Building the org
From 2 to 15, across three functions.
- Grew the CSM team from 2 to 5 — including a Principal CSM track so senior ICs could own strategic, cross-account programs — launch readiness, live-ops process — without being forced into management.
- Created the Game Services Engineer function from zero — three engineers embedded with customers for implementation, release support, and escalations. CS got its own technical capacity instead of borrowing from product engineering.
- Built Professional Services into a seven-person delivery team — turning ad hoc favors into a real revenue line: first contracted engagements in late 2024, a multi-million-dollar delivery business by its first full year — and a technical trust-builder for the platform itself.
- Expanded globally — EMEA coverage across England, Europe, and Romania alongside North American teams in Vancouver, LA, and Montreal — same-timezone support on both continents.
- Ran it with real operating discipline — per-person quarterly goals, weekly reviews of ticket queues and SLAs, and quarterly team reviews — through parental leaves, coverage gaps, and new-hire ramps.
03
Building the systems
From inbox support to an operating machine.
- A tiered support model — Free through Enterprise, each tier with its own SLAs, onboarding flow, channels, and cadence — the economics that let five CSMs credibly cover 100+ accounts.
- A unified studio support queue — one queue across both products with SLAs, categorization, ownership, and escalation paths — replacing scattered Slack-channel triage.
- A launch-readiness program — launch plans, risk assessment, and a live-ops process — because a studio's launch day is the highest-leverage moment in its lifecycle.
- Success planning and an EBR motion — standardized success plans, executive business reviews, and deep-research account briefs for the accounts that warrant them.
- Retention machinery — a churn playbook, account health monitoring, an NPS program, and renewal tracking backed by a contracts database that became the commercial source of truth.
- A weekly operating cadence — team and customer briefs every Friday, stack-ranked by revenue — surfacing renewals inside 60 days, launches inside 30, and open fires. Nothing waits for a QBR to be noticed.
04
A second product
Scaling a launch to 70+ studios.
FirstLook launched to customers in early 2025, and my org owned its support and success motion from day one. By mid-2026 it served more than 70 studios — including some of the most recognized names in games — with revenue growing 2.5x over the first half of 2026 alone.
The design that made it work: dedicated CSMs for enterprise accounts, pooled coverage for standard tiers — high-volume support that stayed economical. We also drove multi-product adoption, cross-selling FirstLook into the Engine base so that a studio's relationship with Pragma deepened rather than fragmented.
05
When it counted
Launches, saves, and the score that moved.
The payoff of the launch-readiness program: five clean game launches supported in the first half of 2026, every one with two or fewer critical issues — engagements that ranged from a live backend migration on a shipped game to launch-day panic-room support. Engine NPS rose from 29 to 42 over the same six months as the launch program and support machinery matured.
Through the industry's hardest stretch, the team ran dedicated firefighter workstreams for at-risk accounts and found third options where a renewal conversation looked binary — saves, downgrades instead of churns, and product conversions that kept the relationship when a contract couldn't survive.
And the learning loop ran the whole time: release-quality feedback from embedded engineers to engineering leadership, and a customer-asks pipeline aggregating requests from support, Slack, and meetings into product planning.
The outcome
In July 2026, Pragma asked me to run the Engine business.
General Manager of Pragma Engine — P&L, go-to-market, and pricing — while continuing to lead the 15-person Studio Operations organization. The GM role was a direct consequence of the CS work: the person who built the retention machinery, the services revenue line, and the commercial source of truth was the person trusted with the business.
Working on a similar build?
Let's talk.
If you're building or rebuilding a post-sales organization — or trying to make two support motions work under one roof — I'd be glad to compare notes.