A systems engagement is scoped around business outcomes rather than a deliverable list, and typically runs as a defined build phase followed by an operating phase. Outcome-linked compensation aligns incentives only when outcomes are measurable, attributable, and substantially within the agency's control. Where any of those three conditions fails — and in Indian mid-market marketing at least one usually does — performance pricing produces gaming rather than alignment.
This argument is commercially inconvenient for us to make. We are an agency arguing against the pricing model buyers most often ask for, which is self-serving on its face. So the case below rests on Nobel-recognised economics and on survey data from the world's largest advertisers rather than on our preference, and we have included the strongest counter-arguments.
If you've read growth partner versus vendor, this is the commercial structure that follows from that distinction.
Why does performance pricing feel obviously right?
Because the logic is clean. If the agency earns more when you earn more, interests are aligned, your downside is capped, and you only pay for results.
Three assumptions sit underneath that. The outcome can be measured accurately. The outcome can be attributed to the agency's work. The agency substantially controls whether it happens.
Each of those is doing far more work than it appears.
Truth line: Performance pricing does not align incentives. It aligns incentives with a metric — which is a different and much narrower thing.
What does the economics literature actually say?
This is the body of work marketing publishing ignores, and it addresses exactly this problem.
In their 1991 paper Multitask Principal-Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design, published in the Journal of Law, Economics, and Organization, Bengt Holmström and Paul Milgrom established the mathematics of what happens when an agent performs several tasks and only some are measurable. Holmström later received the Nobel Memorial Prize in Economic Sciences for work in this area.
Their finding is specific and it maps exactly onto marketing. When some tasks are easily measured and others are not, rewarding the measurable ones causes a predictable reallocation of effort away from the unmeasurable ones. The rational response to a measured incentive is to optimise the measure.
Marketing is unambiguously multi-task. Lead volume is measurable. Brand equity, channel incrementality, creative quality, measurement hygiene, and the judgement to tell a client to stop spending are not. Pay on the first and the literature predicts the rest degrade.
Holmström and Milgrom's conclusion follows: where performance measurement is poor across important dimensions, a fixed fee is superior to a piece rate.
Canice Prendergast's survey of the empirical evidence, The Provision of Incentives in Firms in the Journal of Economic Literature, adds the pricing consequence. An agent forced to bear risk outside their control demands a risk premium. For marketing that means the agency is underwriting your product-market fit, your pricing, your sales team's follow-up speed, and the economy. Rational agencies price that risk in — so clients on pure performance models frequently pay more for the same work, with the premium buying nothing but risk transfer.
Charles Goodhart's observation, formulated in 1975, states the problem in one line: when a measure becomes a target, it ceases to be a good measure. The documented side effects are catalogued in Goals Gone Wild by Ordóñez, Schweitzer, Galinsky and Bazerman — narrowed focus, distorted risk-taking, inhibited learning, and increased unethical behaviour.
What does gaming look like in practice?
Two specific, common, and entirely legal behaviours.
Brand search cannibalisation. An agency paid on ROAS or revenue share has strong incentive to bid on your own brand terms. Those clicks convert at high rates from people who were already coming to you. The agency's attributed revenue rises; your incremental revenue does not. The evidence on brand bidding incrementality is in why your SEO and paid media should share one strategy.
Retargeting weighting. Budget shifted toward audiences already close to purchasing produces excellent attributed returns and a shrinking top of funnel. Reported performance improves while the business's actual demand base contracts.
Neither is fraud. Both are rational responses to the contract.
The measurement precondition has quietly collapsed
Even setting incentives aside, performance pricing requires a number both parties trust. That number is less reliable than it was.
Google Ads retired its first-click, linear, time-decay and position-based attribution models in 2023, leaving last click and data-driven attribution. Data-driven attribution distributes credit across Google's own touchpoints using a machine-learning model whose workings are not externally auditable. Paying a performance fee on it means paying on a number produced by the company selling you the media — which has an interest in the answer.
Meta counts view-through conversions under certain attribution settings, crediting itself when a user saw an ad without clicking and later purchased.
The most revealing example is from LinkedIn, which markets its Conversions API by claiming advertisers see roughly a 31 percent lift in attributed conversions after switching from browser pixels to server-side tracking. That lift is a measurement change, not a business change. An agency on a performance contract could increase its own bonus by about a third by installing tracking infrastructure, having sold nothing additional.
The broader post-cookie measurement picture is in attribution after cookies.
What do the world's largest advertisers actually do?
The empirical answer contradicts the industry narrative.
The Association of National Advertisers, surveying client-side marketers for its Trends in Agency Compensation research, found that 82 percent of marketers use fee-based compensation in at least one agency relationship. The ANA's commentary is blunt: respondents indicated that incentives do not improve agency performance, and that structuring and managing effective incentive plans is complicated, time-consuming, and often ineffective.
Research by the World Federation of Advertisers with MediaSense, surveying multinational companies representing more than $60 billion in advertising spend, found that the large majority cap performance-based elements at under 20 percent of total agency compensation. Around three quarters intend to change their remuneration models — but the direction of travel is from labour-based billing toward output and deliverable-based structures, not toward pure performance.
These are the buyers with the most sophisticated procurement functions, the best measurement infrastructure, and the most leverage over their agencies. They have largely declined the model.
When does performance pricing genuinely work?
It does work, under narrow conditions. All three tests must pass.
Measurable. The outcome is recorded in a system both parties can inspect — typically your CRM, not an ad platform.
Attributable. The causal link is clean. This generally requires single-channel, short-cycle, on-platform conversion. Affiliate marketing is the clearest case: a tracked link, an immediate purchase, no ambiguity.
Controllable. The agency affects the outcome without depending on your sales team, your pricing, or your delivery capacity.
Realistic fits: affiliate and partnership programmes, single-channel ecommerce with clean on-site conversion, and lead-generation where the client verifies qualification in a shared CRM and pays per qualified lead rather than per form fill.
Realistic non-fits: anything multi-channel, anything with a sales cycle beyond a few weeks, anything closing over WhatsApp or by phone, and anything where brand investment is part of the scope. That describes most Indian mid-market B2B and high-ticket categories — the structural reasons are in performance marketing for considered purchases.
So what structure fits a systems engagement?
A phased shape borrowed from management consulting and systems integration, where build-then-operate arrangements are long established.
Phase 1 — Diagnostic and build. Fixed price, defined scope, defined end.
Measurement audit and correction, positioning, tracking and CRM integration, baseline establishment, foundational assets. Typically four to eight weeks. This phase exists partly to establish the baseline that any later discussion of performance would need — you cannot measure improvement against a number nobody recorded.
Phase 2 — Operating retainer. Fixed fee, scoped by capability rather than deliverable count.
Ongoing strategy, channel operation, creative production, and measurement. Fixed because the work is multi-dimensional and the unmeasurable parts matter.
Optional Phase 3 — A capped outcome component, once the baseline is real.
If both parties want shared upside, structure it as a modest bonus — well under 20 percent of total compensation, consistent with what sophisticated advertisers do — against a metric that passes all three tests, measured in your CRM rather than an ad platform, with the baseline agreed in writing before it starts.
The scoping mechanics sit in how to write an SOW that holds, and the contract terms in eleven contract clauses that cost Indian businesses lakhs.
What are the Indian tax mechanics?
Rarely discussed and genuinely consequential for cash flow.
GST. Integrated growth and marketing consulting falls under Service Accounting Code 9983, covering other professional, technical and business services, as classified by the CBIC. It attracts GST at 18 percent, which applies to retainers, build fees, and any performance bonus alike.
TDS. Section 194J of the Income Tax Act, 1961 governs tax deducted at source on fees for professional or technical services. Marketing strategy, creative, and management consulting are treated as professional services at 10 percent; certain narrowly specified technical services attract 2 percent. Threshold limits have been revised in recent Finance Acts, and sources differ on the current figure — confirm with your chartered accountant. At any plausible threshold, a ₹2 lakh monthly retainer breaches it in the first billing cycle.
Why this matters for structure. Performance and revenue-share arrangements complicate both. The taxable value of a bonus is uncertain until the period closes, which complicates invoicing, GST timing, and TDS deduction. Fixed fees are administratively simpler for both sides — a mundane point that affects real cash flow.
For Canadian clients. Regulation 105 of the Canadian Income Tax Regulations requires Canadian payers to withhold 15 percent from payments to non-residents for services rendered in Canada. Waivers are available through the Canada Revenue Agency, including under the Canada-India tax treaty, but without advance approval the withholding applies.
The three-test check
Before agreeing any outcome-linked component, run it:
- Is the outcome recorded in a system we both control? If it lives only in an ad platform, no.
- Can the agency's contribution be separated from everything else? If your sales team's response speed moves the number, no.
- Would a reasonable person agree on the figure without a dispute? If not, you have designed a disagreement.
Three yeses means a capped performance component is defensible. Fewer means a fixed fee is the honest structure, and an agency proposing otherwise is either not thinking carefully or planning to manage the metric.
What we could not establish
Three gaps.
No Indian mid-market compensation data exists. ANA and WFA data comes from global enterprise advertisers. Nothing published isolates how Indian ₹5–50 crore businesses pay their agencies.
No dispute statistics exist. There is no registry of contractual disputes arising from attribution disagreements in performance contracts, so the frequency of the failure mode is undocumented even though the mechanism is well understood.
Build-then-operate prevalence in Indian marketing is unmeasured. The structure is well established in consulting and IT services. Nobody has published data on its use in marketing here.
How Midgrow structures engagements
Stated plainly, including the part that costs us business.
- A fixed-price diagnostic and build phase, with a defined scope and a defined end, which establishes the baseline everything later is measured against
- A fixed-fee operating retainer, scoped by capability and named allocation rather than by deliverable count
- We decline pure performance and revenue-share arrangements. Not because we doubt our results, but because our scope includes work that cannot be measured monthly, and we would rather be paid to do that work than paid to avoid it
- We will agree a capped outcome component where all three tests pass, measured in your CRM, against a baseline written down before it starts
- Full transparency on tax treatment at proposal stage, since GST and TDS affect your actual cost
Full scope is on our digital marketing services page, including social media systems.
The proof is public rather than promised. We generated 10,890 leads at 11.3x ROI for a solar EPC client and delivered 585 percent organic growth for Autosys Solar — both fixed-fee engagements in manufacturing and energy, where the work that produced the result would not have survived a monthly performance metric.
Book a 45-minute growth diagnostic. Bring any performance-based proposal you're considering, ours or anyone's. We'll run the three tests against it. Start the conversation.
Frequently asked questions
Should I pay my marketing agency based on performance?
Only when three conditions hold: the outcome is recorded in a system you both control, the agency's contribution can be separated from other factors, and a reasonable person would agree on the figure without dispute. In most Indian mid-market marketing, attribution fails the second test, which turns a performance clause into a dispute mechanism rather than an alignment one.
Why do most large advertisers use fixed fees?
Because incentive plans have not delivered. ANA research found 82 percent of marketers use fee-based compensation in at least one agency relationship, with respondents reporting that incentives do not improve agency performance and that structuring them is complicated and often ineffective. WFA research found multinationals typically cap performance elements below 20 percent of total compensation.
What is the multitask problem in agency compensation?
Holmström and Milgrom's 1991 research established that when an agent performs several tasks and only some are measurable, rewarding the measurable ones causes effort to shift away from the unmeasurable ones. Marketing has highly measurable elements like lead volume and unmeasurable ones like brand equity and measurement hygiene, so paying on the former predictably degrades the latter.
Is revenue share a good model for marketing agencies?
Rarely, for two reasons. It requires clean attribution, which most multi-channel engagements lack. And it forces the agency to bear risk it does not control — your pricing, product, and sales follow-up — for which rational agencies charge a risk premium. Clients frequently end up paying more for identical work, with the premium purchasing only risk transfer.
How can an agency inflate performance metrics without improving results?
Several ways, all legal. Bidding on your brand name captures customers who were already coming. Weighting budget toward retargeting harvests demand created elsewhere. Installing server-side tracking raises attributed conversions substantially — LinkedIn claims around 31 percent for its Conversions API — without selling anything additional.
What is a build-then-operate engagement structure?
A fixed-price diagnostic and build phase lasting roughly four to eight weeks, covering measurement correction, positioning, tracking integration and baseline establishment, followed by an ongoing fixed-fee operating retainer. It is standard practice in management consulting and systems integration, and it solves the baseline problem — you cannot measure improvement against a number nobody recorded.
What GST and TDS apply to marketing agency fees in India?
Integrated marketing and growth consulting falls under SAC 9983 and attracts 18 percent GST. Section 194J of the Income Tax Act governs TDS, at 10 percent for professional services including marketing strategy and creative. Threshold limits have been revised recently and sources differ, so confirm the current figure with your chartered accountant.
Does performance pricing attract better agencies?
Often the opposite. Capable agencies decline contracts requiring them to underwrite your product-market fit, pricing, and sales execution. The agencies most willing to accept pure performance risk are frequently those planning to manage the metric through brand bidding and retargetting weighting rather than those building long-term demand.
