The commercial model is part of the delivery architecture
A modernization contract allocates responsibilities, uncertainty and incentives. When those allocations do not match the work, both buyer and provider can spend more effort arguing over scope than resolving the transformation problem.
Henna83 recommends choosing a model after understanding the outcome, evidence, dependencies and controllable effort. A fixed fee can suit a bounded diagnostic. A variable model may suit uncertain engineering discovery. Milestones can support implementation when acceptance is clear.
Three decisions come first: which assumptions are sufficiently known, which risks each party can influence, and how completion will be demonstrated. The engagement should connect these decisions to staffing, acceptance and change control.
Commercial design also includes client capacity. A provider cannot complete reconciliation if no business owner can approve meaning or access the required evidence. The plan should name those dependencies, their dates and the consequences of delay.
Choose the fee structure that fits the uncertainty and accountability of the work.
Operational and economic responsibilities intersect
The interagency third-party guidance [1] addresses relationship risk management, including contract and monitoring considerations. Banks should align those considerations with the activity, rather than treat outsourcing as a transfer of institutional responsibility.
The 2026 State of FinOps report places governing AI value among future priorities [2]. That reinforces the need to understand cost and value across evolving technology services. It does not prescribe a consulting contract or a guaranteed outcome.
AWS readiness guidance spans six perspectives [3]. Henna83's commercial application is to identify the work and client contributions across those perspectives before pricing. Governance, training or operational handoff should not disappear from scope merely because the engineering estimate is more concrete.
Published research informs the decision, but the contract should be grounded in the actual engagement. Establish the accepted output, access assumptions, staffing model, specialist requirements and change mechanism. Avoid outcome language that depends primarily on factors outside either party's control.
A credible commercial model makes uncertainty explicit and assigns a workable response.
Match the model to the work
Fixed fee: useful for a bounded output with known assumptions and acceptance. Define exclusions, client inputs and revision limits. A fixed price for unclear scope can create friction rather than certainty.
Time and materials: useful when discovery or engineering effort is uncertain. Establish rates, authorized capacity, spending controls and expected outputs. Hours consumed are not a sufficient performance measure.
Milestone: useful when outputs can be accepted objectively. Tie payment to agreed deliverables and evidence, with a process for timely client review and correction.
Managed service: useful for an ongoing operating responsibility with defined service levels, demand boundary and change route. Specify ownership, volume assumptions, incident handling and exit arrangements.
Outcome-linked: appropriate only when baseline, measurement, attribution and controllable influence are established. A provider should not be rewarded for apparent savings that depend on unverified volume changes or costs the institution cannot retire.
Economics context: provider margin and client ROI are different calculations. A provider prices resources, delivery risk and support obligations. A client evaluates business benefit and total investment, including internal effort. Keep those views separate and use assumptions consistently. Where outcome-linked compensation is considered, test whether measurement is reproducible and whether volume, client decisions or market changes can dominate the result.
Exhibit: commercial model selection
This is an original Henna83 decision matrix. It supports discussion and does not replace negotiated legal terms.
| Model | Best-fit condition | Key control |
|---|---|---|
| Fixed fee | Bounded scope and known inputs | Assumptions and acceptance |
| Time and materials | Uncertain effort; flexible priorities | Rates, cap and output reviews |
| Milestone | Discrete demonstrable outputs | Acceptance and review timetable |
| Managed service | Stable operating boundary | Service levels, demand and exit |
| Hybrid | Different uncertainty by phase | Clear phase boundaries |
| Outcome-linked | Measurable, attributable influence | Baseline, attribution and auditability |
Using the matrix: identify uncertainty by phase. The discovery phase may establish facts that make a later fixed-scope milestone possible. For each model, record authorization limits, output reviews, acceptance timing and change triggers. Define what happens when a client dependency is delayed. Review the structure with delivery and legal advisers before the parties commit.
A hybrid can separate fixed-scope diagnosis from engineering work that is estimated after discovery.
Worked example: separate diagnosis from build
Illustrative scenario, not a negotiated client agreement. A servicer needs a new data feed but lacks an approved mapping and complete consumer list. Pricing the entire implementation as fixed fee would require substantial assumptions.
The parties first agree a bounded diagnostic with fixed outputs: mapping questions, dependency register, architecture recommendation and implementation estimate. Client owners supply source evidence and review the recommendations on agreed dates.
Implementation is then scoped from the findings. If effort remains uncertain, the parties may use time and materials with an authorization cap and frequent output reviews. A subsequent milestone can cover accepted integration once the criteria are stable.
Change control records the trigger, impact, options, price and authorized decision before additional work starts. Specialist resources are identified through rate cards and agreed commitments. The buyer understands which expertise is included and which requires a separate allocation.
The model creates evidence before making a larger commitment. It also preserves a clear decision point if the diagnostic shows that the proposed build should change or stop.
Questions for the buyer and provider: Which risks can each party influence? What baseline governs? Who authorizes change? What counts as acceptance? What happens when access is delayed? How are specialists approved?
Prepare the engagement for execution
Create a one-page commercial architecture: problem, outputs, model, staffing, client responsibilities, assumptions, acceptance and change control. Review it with the sponsor, procurement, finance and delivery leads.
Model provider effort and costs separately from client benefits. Keep sample economics clearly labeled. Agree which changes need authorization and how each party will respond to access or decision delays.
At mobilization, confirm named owners, dates, resources and evidence. Review progress against outputs and assumptions, not just invoices. Correct the scope when the facts change.
Henna83 connects commercial judgment with integrated delivery. Use the Vendor RFP Assessment Kit to compare scope and assumptions, then request executive advisory to discuss a consequential engagement structure. Full agreements should be finalized with appropriate legal review.
Measurement detail: review spending against authorized capacity and outputs against acceptance. Track assumption changes separately from defects. In a variable model, forecast remaining effort. In a milestone model, retain evidence supporting acceptance rather than relying on a billing event.
References
[1] Federal Reserve, FDIC and OCC. Interagency Guidance on Third-Party Relationships: Risk Management. June 6, 2023. https://www.federalreserve.gov/frrs/guidance/interagency-guidance-on-third-party-relationships.htm
[2] FinOps Foundation. State of FinOps 2026. 2026. https://data.finops.org/
[3] Amazon Web Services. Evaluating migration readiness. Living technical documentation; accessed October 10, 2026. https://docs.aws.amazon.com/prescriptive-guidance/latest/evaluating-migration-readiness/introduction.html