Game-studio version control · 2026
What Version-Control Migration Work Will 6–50 Person Game Studios Pay For?
Active 6–20 person asset-heavy repositories are the narrow entry; tiny-team setup and unbounded hosting are not.
The report follows the whole cutover—history, permissions, binary assets, artist workflow, engine integrations, build machines, training and rollback—not merely file transfer.
- Solution sample
- 5 current hosting and tool routes.
- Demand sample
- Two survey frames with 1,176+ responses.
- Migration evidence
- Named studios, workflows and current prices.
- Market
- Established adoption and managed-service spending.
- Best bounded job
- A rehearsed 6–20 person production-repository migration.
- No-Go
- Tiny-team setup or unbounded solo 24/7 hosting.
Brief mode keeps the findings, prices, segment comparison and final answer.
Research context
- Background: Game repositories combine code with large binary assets, and a live studio can outgrow Git, SVN or a self-managed server without being able to pause production safely.
- Answer sought: Which studio sizes and repository conditions create paid migration work, what does a safe cutover include, and where is a small supplier's responsibility bounded?
- Main sources: Two current workflow surveys, five hosting/tool routes and prices, named studio migrations, and current cutover/service documentation.
In plain language: what does the customer buy?
Game projects contain code plus large binary files such as scenes, textures, audio, animation and 3D assets. When a studio outgrows Git, Git LFS or Subversion, it may pay someone to inspect the live repository, move history and permissions, configure file locking and branches, reconnect Unity or Unreal and the build system, train artists and programmers, rehearse rollback, and keep production running during the change.
Copying files is the small part. The purchased outcome is a controlled cutover in which people, assets, builds and automation still work.
Scope and sample
- Five current solution routes: Perforce P4 Cloud, Assembla P4 Cloud, Diversion, Unity Version Control, and self-managed P4/Git LFS.
- Six overlapping jobs: setup for five or fewer users, 6–20 person migration, self-hosted P4 to managed cloud, multi-site performance, permissions/backup/disaster recovery, and continuing administration/training.
- Demand denominators: a 2026 cross-industry Perforce/AWS survey of 600+ practitioners and a 2024 Perforce/JetBrains survey of 576 respondents in 64 countries.
- Case evidence: Assembla reports 150+ game studios and multiple named migrations; Perforce and Diversion publish concrete migration workflows and named game-studio cases.
Both surveys came through vendor ecosystems and cover industries beyond games. They are useful demand proxies, not a census of independent studios.
Findings at a glance
- [Observation] Version control is now near-universal in the surveyed real-time-workflow sample. The 2026 survey reports 94% adoption, up from 86% in 2025; hybrid cloud/on-premises deployment rose from 10% to 27%.
- [Observation] Asset handling is a specific game-production problem. The 2024 sample reports 69% using version control to store/share source and art assets, while 38% call large files a leading collaboration challenge.
- [Observation] Managed operation is a recurring paid product. P4 Cloud is US$39/user/month; Assembla's P4 route is US$52.25/user/month on annual commitment; P4 also sells yearly Remote Administration.
- [Observation] The target buyer range is explicit. P4 Cloud supports teams under 100, Assembla positions its multi-tenant game-studio service below 50, and Diversion publishes pricing through 20 users.
- [Observation] Small-team alternatives are aggressive. P4/self-hosting and Diversion are free through five users; Unity removed cloud seat charges and includes 25GB storage plus 100GB monthly egress.
- [Inference] The viable small-supplier job begins after a real repository has become painful. A 6–20 person team with large binaries, active builds and mixed artist/programmer workflows has a bounded migration problem; an empty repository does not.
Who pays, and what triggers the project?
| Buyer | Trigger | Purchased result |
|---|---|---|
| Growing indie or AA studio | Git LFS/SVN conflicts, slow pulls, repository growth | Tested move to P4, Diversion or another suitable host |
| Remote Unity/Unreal team | Artists and contractors need safe locking and permissions | Role-based workflow, client setup and training |
| Studio already running P4 | Server updates, backup, cost or single-admin risk | Migration to managed cloud or bounded remote administration |
| Distributed production team | Latency, replicas and build systems fail across regions | Architecture review and staged multi-site rollout |
Public service pages name studios in North America, Europe and Asia, from indie teams to large distributed companies. They do not expose a complete country or team-size distribution.
Current prices show why tiny-team setup is weak
| Route | Current public entry | Included or qualifying boundary |
|---|---|---|
| Self-managed P4 | Free through five users | The studio owns infrastructure and administration |
| Diversion Indie | First five users and 100GB free; users 6–10 US$12 each; 11–20 US$25 each | Requires game revenue below US$100K and funding below US$1M |
| P4 Cloud | US$39/user/month | Single tenant, 64GiB, updates, backups and support; up to 100 users |
| Assembla P4 Cloud | US$52.25/user/month, annual | Multi-tenant, 1TB, managed service; positioned below 50 users |
| Unity Version Control | Unlimited cloud seats; 25GB and 100GB egress free; excess egress US$0.10/GB | Unity-oriented public cloud; storage/egress meters remain |
At ten users and before storage or extras, Diversion Indie calculates to US$60/month because the first five users are free; P4 Cloud is US$390 and Assembla US$522.50. The latter are 6.5× and 8.71× the Diversion figure, but capabilities, storage, tenancy and eligibility differ.
Assembla's per-seat rate is 34.0% above P4 Cloud's while including different storage and contract terms. The comparison shows strong substitute pressure, not which product is “better.”
Direct evidence that migration is more than copying files
| Source | Observed problem and work | Decision relevance | Caveat |
|---|---|---|---|
| Game Studio / Perforce | Concurrent SVN and Git caused merge failures and data inconsistencies requiring troubleshooting almost daily; new P4/Azure path included SSO and two projects | Mixed-tool debt and identity/build integration trigger migration | Vendor-selected case, no price |
| Tequila Works / Perforce | 400,000 files and 40GB; SVN to P4; Unreal, TeamCity, C++, Maya and 3ds Max integration; customer reports first-year ROI | Repository scale and toolchain acceptance matter | Older case and self-reported ROI |
| Assembla | Multiple-terabyte and growing-team cases; migration, optimisation and daily operations; most reported migrations under two days | Managed migration and recurrence are explicit products | Vendor claim; no denominator or failure rate |
| Diversion | Side-by-side bidirectional sync, real production validation, pilot groups, training and controlled cutover | A current migration acceptance pattern is documented | Vendor route, not independent-service demand |
The counterexample is decisive: P4 Cloud says a team can be running in an hour, while Diversion and Unity remove most seat cost for small teams. Therefore basic setup is not the same market as production migration.
Six jobs compared
| Segment | Demand evidence | Competition and substitutes | Entry judgment |
|---|---|---|---|
| New setup for five or fewer users | Version control is foundational | P4, Diversion and Unity have free/turnkey paths | No-Go: little independent pricing power |
| 6–20 person Git/SVN migration | Large-file denominator plus Game Studio and Tequila cases | Three managed vendors provide migration help | Best bounded segment: only with a live repository and acceptance plan |
| Self-hosted P4 to managed cloud | Assembla multi-TB cases and P4 Remote Administration | P4 and Assembla operate the service directly | Sell assessment/cutover, not an unmanaged server |
| Multi-site and replica performance | Distributed-team cases and hybrid growth | Mature enterprise P4/Assembla capabilities | Higher value, poor small-supplier risk fit |
| Permissions, backup, DR and SSO | Every managed route emphasises them | Standard hosting includes much of the work | Only complex legacy boundaries remain |
| Continuing administration and training | P4 annual administration and Assembla daily ops | Vendors and partners already compete | Recurring, but response liability is high |
The smallest sellable migration
The evidence supports this deliverable:
Repository and workflow audit; rehearsal on a production-sized copy; history and permission mapping; large-file, lock and branch rules; Unity/Unreal and build-machine validation; pilot-group training; scheduled cutover, rollback and a 30-day stability window.
Acceptance is concrete: agreed history is present, permissions match the matrix, representative binary assets lock and sync, a clean workstation opens and builds the project, automation runs, rollback is rehearsed, and named users complete the workflow.
A host or cloud partner is a more defensible route than independently operating mission-critical infrastructure. Ongoing administration must have explicit hours, monitoring, escalation and recovery ownership.
Risks, stop conditions and evidence gaps
| Issue | Effect | Boundary or stop condition |
|---|---|---|
| Free setup and automated hosting | Tiny teams can self-serve | Stop if the repository is new, below five users and has no toolchain debt |
| Production interruption | A bad migration blocks every discipline | No cutover without rehearsal, backup, freeze window and rollback |
| Large binary and egress cost | Storage and download patterns can dominate | Model actual depot growth and regional transfer before choosing a host |
| Support liability | Builds and source control can fail during crunch | No unbounded 24/7 solo SLA |
| Missing economics | Migration prices, win rates, hours and failure costs are private | Seat price and customer savings are not service profit |
Final answer
The market and the problem are established. Current surveys, five active routes, recurring managed-service prices, 150+ vendor-reported game-studio customers and multiple migrations show sustained demand.
The enterable segment is narrow: an active 6–20 person, asset-heavy repository. The paid work is risk reduction across history, permissions, artists, engine integrations, builds, training and rollback.
Basic tiny-team setup and independent always-on hosting are No-Go. Free and turnkey products remove the first, while the second creates reliability exposure that exceeds a bounded migration project.
Direct sources
- 2026 State of Real-Time Workflows press release (accessed 2026-08-27)
- 2024 State of Game Technology report (accessed 2026-08-27)
- P4 Cloud (accessed 2026-08-27)
- Perforce cloud-route comparison (accessed 2026-08-27)
- Assembla for game studios (accessed 2026-08-27)
- Diversion Indie pricing (accessed 2026-08-27)
- Diversion migration workflow (accessed 2026-08-27)
- Unity 2026 Version Control pricing changes (accessed 2026-08-27)
- Game Studio migration case (accessed 2026-08-27)
- Tequila Works migration case (accessed 2026-08-27)
Completeness note: The report covers five current solution routes, six paid jobs, two survey frames with 1,176+ combined responses across different years, current prices and four migration/service evidence groups. It supports a segment, delivery and risk-boundary decision—not the number of addressable studios, typical transaction price, acquisition cost or profit.