Lesson 4 of 6
Funding models
9 min
Listen to this lesson
How a TCoE is funded decides how people use it. Fund it badly and projects will bypass it, the regression pack will decay, or the TCoE will be cut in the first budget squeeze. There are three basic models, and most SAP customers end up combining them.
The three basic models
| Model | How it works | Behavior it encourages | Typical risk |
|---|---|---|---|
| Central budget | IT funds the TCoE as a run cost, like the service desk | Projects use TCoE services freely; standards spread | Seen as overhead and cut in a bad year; demand is unlimited |
| Project funding | Each project pays for the testing it consumes | Costs land where the change is; projects value what they pay for | Nobody funds shared assets between projects; the pack decays |
| Chargeback | Services have a price; consumers are charged per use | Demand is disciplined; the TCoE must be efficient | Projects skip services to save budget, weakening standards |
Showback is a useful middle step: the TCoE reports what each consumer used and what it would have cost, without moving money. It builds cost awareness without the bypass risk of real chargeback, and it gives you the usage data to design a chargeback later if you need one.
A showback report needs only four columns: consumer, service used, volume and the cost it would have carried at catalog rates. Contoso's first quarterly report, below, was produced from the TCoE's intake log and SAP Cloud ALM test plans. It moved no money, but it settled two arguments: order to cash was the heaviest consumer of the shared regression pack, and the Monterrey roll-out was already paying its own way.
| Consumer | Service used (quarter) | Volume | Cost at catalog rates | Funded by |
|---|---|---|---|---|
| Order to cash (Maria, key user) | Regression pack runs and maintenance after the stack update | 1 release event, 310 of 700 h | $29,450 | Run budget (core) |
| Finance (Dan) | Regression pack runs, close-week evidence pack | 1 release event, 220 of 700 h | $20,900 | Run budget (core) |
| Wave 2 Monterrey (1910) | Test planning, data loads, UAT facilitation | 640 h | $60,800 | Project budget (demand) |
| Enhancement backlog | Test design and execution support | 28 changes × 22 h | $58,520 | Run budget (core) |
| Security patch day | Targeted note and patch checks | 3 checks × 35 h | $9,975 | Run budget (core) |
Contoso's first quarterly showback. Illustrative course data at the agreed $95 blended rate; the other 170 regression hours served procurement and manufacturing.
The hybrid most SAP customers use
An SAP TCoE has two kinds of work. Core services never stop: maintaining the regression pack after every release, data recipes, the environment calendar, standards and reporting. Demand services rise and fall with projects: test planning and execution for a roll-out, UAT facilitation, a new automation build. Fund the first centrally, because nobody else will, and the second from the projects that create the demand.
Contoso's choice
- Run budget for the core team of five and the shared services, held by IT and reviewed annually against the scorecard.
- Project funding for demand services. The Wave 2 Monterrey roll-out (company code 1910) pays for its own test planning, data loads and UAT facilitation from the project budget, at rates published in the service catalog.
- Investment tranches for the roadmap: the $229,000 one-time investment was released in three tranches, each after a milestone such as "order-to-cash regression automated and stable for one release."
- Showback to every consumer each quarter, so the quality council can see who uses what.
- Partner capacity under a framework agreement with call-off rates, charged to whoever triggered the peak: a feature pack stack update to the run budget, a roll-out to its project.
Accounting questions to settle early
Ask finance two questions before the case is presented. First, can automation builds and data tooling be capitalized as part of a project asset, or are they operating expense? Rules differ by company and accounting standard, so never assume. Second, how are subscription costs for tools treated across years? The answers change how the case looks in each year's budget, even though the economics stay the same.
- Is the work needed every year regardless of projects?Yes → central run budget · No → next question
- Is it caused by one project or roll-out?Yes → project funding at catalog rates · No → next question
- Does it build a lasting asset, such as automation or data tooling?Yes → investment released by milestone · No → next question
- Is it a peak, such as an upgrade?Yes → partner call-off charged to the trigger · No → review whether it belongs in the TCoE at all
What would you do?
A year after launch, Contoso's CFO proposes moving the whole TCoE to chargeback "so that it pays for itself." What should Priya recommend?
- Central funding spreads standards but looks like overhead; project funding starves shared assets; chargeback invites bypassing.
- Fund never-ending core services centrally and project-driven demand services from projects.
- Release investment in tranches tied to milestones, and use showback to make usage visible.
- Settle capitalization and subscription treatment with finance before presenting the case.
Check your understanding · not graded
Lesson check
1. What is the main risk of funding a TCoE entirely by project budgets?
2. What does showback do that chargeback does not?
3. Which Contoso cost should be charged to the Wave 2 Monterrey project budget?
4. Finance says automation builds cannot be capitalized at Contoso. What changes in the business case?