Managed marketing services solve your capability gap on someone else’s balance sheet of knowledge — and after three years, the knowledge is still on their balance sheet, not yours.
The offer is everywhere in India right now, and it is genuinely attractive. The global agency groups sell managed data and campaign operations out of Indian delivery centres; the India-native engagement platforms sell managed-campaign tiers to brands that bought software and discovered nobody in-house could fly it. The pitch has not changed in a decade: you supply the data and the strategy sign-off, we supply the platform, the data scientists, the campaign operations team and the reporting. For a mid-market Indian brand — a real marketing budget, a two-person analytics function, no campaign platform — it is often the only honest route to competence on any reasonable timeline. This piece is not an argument against buying it. It is an argument about what you must buy alongside it.
Across a decade of deployments, the entry condition is remarkably consistent. A global sportswear brand’s India business, at the point it signed a managed services contract in the email-and-text-message era, ran campaigns like this: lists pulled manually from point-of-sale and events data into spreadsheets, execution outsourced to agencies on commodity mailing and messaging tools, the overwhelming majority of volume on a single channel, responses shared back by hand, no segmentation, no propensity models — pure broadcast. That was not backwardness; it was the standard operating state of mid-tier Indian retail at the time, and versions of it are still common today, just with better-looking dashboards. The managed service replaced it within months with a proper marketing data mart, statistical and campaign-responsiveness segmentation, propensity models for cross-sell, an automated execution platform and a standing reporting layer — run by a compact specialist team, mostly offsite in the vendor’s delivery centre. On its own terms, it worked. That is the first half of the story, and it is the half the vendor’s case study tells.
The second half is structural, and no proposal we have ever reviewed volunteers it. Watch where the capability physically accumulates. The segmentation models are built, tuned and retuned by the vendor’s data miners. The data mart’s architecture, its cleansing rules, its quirks and workarounds live in the heads and wikis of the vendor’s engineers. The campaign workflows, the deliverability craft, the test-and-learn history — all of it compounds inside the delivery centre. Two or three years in, the brand owns its customer data in every legal sense and almost none of the operational knowledge required to use it. End the contract and you keep the data and lose the capability; you are rebuilding from scratch, with the additional indignity of paying exit-transition fees for an orientation tour of your own customer base. We call this the dependency asymmetry, and it is not a bug in any particular vendor’s model. It is the model. The recurring fee is priced on the continuing gap between what you can do and what they do for you, which means no line item in the standard proposal exists to close that gap. In the contracts of this kind we have seen — including ones we helped write, on the selling side — the knowledge-transfer workstream is conspicuous by its absence.
The dual lens here is brand-versus-evidence in miniature. The brand voice says: marketing capability is identity — a brand that cannot operate its own customer relationships has outsourced part of what it is. The evidence voice is more pragmatic: rented capability produces real, measurable output from quarter one, and an in-house build at low maturity produces eighteen months of hiring, tooling and nothing shipped. Both are right, which is why the answer is a sequenced one — rent, then transfer — rather than a religious one. Where managed services make commercial sense at all is itself a bounded question. As an illustrative order-of-magnitude guide for the Indian mid-market, drawn from practice rather than any single engagement:
| Annual marketing-operations spend (illustrative, order of magnitude) | Sensible default model |
|---|---|
| Below about ₹1 crore | Tech-enabled agency on a standard platform stack |
| Roughly ₹2–5 crore | Managed services — the commercial sweet spot |
| Above about ₹10 crore | Build in-house; managed services now cost you compounding capability |
The engagements we have seen in that middle band priced out at low single-digit crore per year all-in — enough to fund a small specialist team and a platform, far less than an equivalent in-house build-out. The band matters because the dependency trap tightens at its upper edge: the brand that grows through the band while the know-how stays in the delivery centre is the one that wakes up, at in-house scale, unable to insource.
There is an instructive precedent from an adjacent industry. When large Western retailers built their retail media businesses, the standard structure became agency-built with an explicit insourcing milestone — a two-to-three-year partnership scoped from day one to end in transfer, not a perpetual managed service. The operating knowledge was treated as a deliverable with a due date. There is no reason marketing analytics and campaign operations should be contracted any other way; the only reason they usually are is that one side of the table benefits from the omission and the other side does not think to ask.
On Monday, if you are in a managed services contract, ask your team one question: if this ended in ninety days, what walks out the door? Have someone who does not manage the vendor write the answer down — models, data mart design, workflow logic, deliverability history, test-and-learn record. If you are about to sign one, add the workstream the proposal left out and pay for it explicitly: documentation as a standing deliverable, your analysts shadowing model builds from year one, model code and data mart schemas escrowed in your environment, an insourcing milestone with a date and a named internal owner. Expect it to add meaningfully to the fee, and pay it anyway. A managed service priced without knowledge transfer is not cheaper. It is the same capability, rented in perpetuity, with the buy-out clause deleted.