Loading...
Loading...
Profit split and TNMM are both transactional profit methods, but they solve different problems. This guide explains when to use one-sided TNMM, when profit split is more reliable, and how to document the choice.
Borys Ulanenko
CEO, ArmsLength AI
Continue exploring
Browse the full resource library or contact us if you want recommendations for your specific use case.
TNMM tests one party to a controlled transaction. It is best when that party is routine, does not own unique intangibles, and can be benchmarked against independent companies using a net profit indicator.
Profit split evaluates the combined profit from controlled transactions and divides that profit between the parties based on their relative contributions. It is best when both parties contribute unique and valuable intangibles, highly integrated functions, or shared economically significant risks.
Both are transactional profit methods, but they answer different questions:
For full background, see CPM vs TNMM, profit split method guide, and transfer pricing methods guide.
| Factor | TNMM | Profit split |
|---|---|---|
| Method family | Transactional profit method | Transactional profit method |
| Analytical perspective | One-sided | Two-sided or multi-sided |
| What it tests | Net profit of the tested party | Combined profit allocation |
| Typical question | Is the routine party earning an arm's length margin? | How should residual or combined profit be divided? |
| Best fact pattern | One party is routine and benchmarkable | Multiple parties make unique and valuable contributions |
| Comparable data | Comparable companies or transactions for tested party | Allocation keys, contribution evidence, sometimes routine return benchmarks |
| Intangibles | Tested party usually has no unique intangibles | Often used where more than one party contributes intangibles |
| Integration | Works when parties can be evaluated separately | Stronger when operations are highly integrated |
| Complexity | Moderate | High |
| Main audit risk | Misidentifying a non-routine party as routine | Unsupported allocation key or unreliable combined profit pool |
TNMM is usually stronger when one party can be isolated and tested reliably.
Good TNMM fact patterns include:
In these cases, the tested party earns a routine return, and residual profit stays with the party that owns the key intangibles, controls strategic risk, or performs entrepreneurial functions.
TNMM depends on a coherent tested party selection. If the selected party actually makes unique contributions or controls economically significant risks, a one-sided method may understate its arm's length return.
Profit split becomes more persuasive when one-sided testing cannot capture the economics.
Indicators include:
describes profit split as identifying the relevant profits from controlled transactions and splitting those profits on an economically valid basis. It is not merely a fallback when TNMM is inconvenient; it may be the most appropriate method when both sides create non-routine value.
If both parties own valuable intangibles or perform unique development, marketing, trading, or risk-control functions, a one-sided TNMM can allocate too much residual profit to one side by assumption.
Profit split is not required just because an intangible exists. If one entity owns and controls the valuable IP while the other performs routine services, TNMM may still be reliable.
Allocation keys should reflect value creation. Costs, headcount, assets, R&D spend, or weighted contribution scores can be relevant, but none is automatically correct.
Profit split requires the right combined profit pool. Including unrelated business lines or excluding key transaction costs can distort the allocation.
Many cases call for a residual analysis: allocate routine returns first, then split residual profit. Skipping the routine-return step can overstate the profit attributable to unique contributions.
Facts: ParentCo owns product IP, sets global pricing strategy, and bears inventory and product risk. LocalCo distributes finished goods locally under limited-risk terms.
TNMM fits because LocalCo performs routine distribution and can be benchmarked separately. A profit split would add complexity without improving reliability.
Facts: USCo builds core technology while EUCo develops market-specific modules and controls regional product strategy. Both teams contribute engineers, product managers, and proprietary know-how. The product is sold as one integrated platform.
Profit split fits because both parties contribute unique value and the platform profit cannot be reliably attributed to one routine tested party.
Profit split is usually applied in one of two ways.
| Approach | How it works | Best for |
|---|---|---|
| Contribution analysis | Splits total combined profit based on relative contributions | Integrated operations where all relevant contributions are non-routine |
| Residual analysis | Allocates routine returns first, then splits residual profit | Cases with both routine functions and unique contributions |
For detailed mechanics, see the profit split method guide.
For TNMM, auditors will focus on:
For profit split, auditors will focus on:
Profit split documentation often requires more operational evidence than TNMM: management accounts, project records, R&D histories, risk-control evidence, product roadmaps, and interviews with business owners. See the functional interviews guide and benchmarking study guide for supporting workstreams.
Not generally. Profit split is better when both parties make unique and valuable contributions or are highly integrated. TNMM is better when one party is routine and can be benchmarked reliably.
No. Under the OECD framework, the most appropriate method depends on the facts. Profit split may be the best method from the start when the transaction involves shared unique contributions or integrated value creation.
Yes, if the tested party does not make unique and valuable contributions to those intangibles. For example, a routine contract R&D service provider may be tested under TNMM while the principal earns residual profit.
The biggest risk is treating a non-routine entity as routine. If the tested party contributes valuable intangibles, controls key risks, or performs unique functions, TNMM may undercompensate it.
The biggest risk is an unsupported allocation key. A profit split must explain why the chosen split reflects relative contributions in a way independent parties might have agreed.
Yes. A residual profit split often uses TNMM-style benchmarking to assign routine returns first, then splits residual profit based on unique contributions.