In-house or agency: how a ₹1 to 10 Cr D2C brand should decide

We run an agency, so treat this with the suspicion it deserves. It is still worth writing, because the build versus buy decision is usually made on the wrong inputs: a salary compared against a retainer, with none of the surrounding costs on either side, and no honest read of what the brand needs next year.

What follows is the comparison as we would make it for a brand doing roughly ₹1 to 10 Cr a year, including the cases where the answer is not to hire one.

The cost comparison, done properly

Start with published numbers rather than assertions.

On the hiring side, Talent.com puts the average performance marketing manager salary in India at ₹10,50,000 a year, with entry level positions starting around ₹4,25,000 and the most experienced up to ₹22,40,000. For the creative half, the same source puts the average graphic designer in India at ₹6,00,000 and a video editor at ₹2,40,000, rising to ₹19,10,000 at the experienced end.

On the agency side, two Indian firms publish their view of market rates. PeakPilots, writing in June 2026, gives typical D2C retainer ranges of ₹40,000 to ₹80,000 a month at ₹1 to 3 lakh of monthly ad spend, ₹80,000 to ₹1.5 lakh at ₹3 to 8 lakh spend, and ₹1.5 lakh and above beyond that, and says most serious D2C agencies require ₹1 lakh or more in monthly ad spend before the maths works at all. upGrowth, writing in April 2026, describes three structures: a flat retainer of roughly ₹1.5 lakh to ₹6 lakh a month, a percentage of ad spend at typically 10 to 15 percent, or a hybrid of the two. Those are each firm's published figures, not universal rates, and they differ from each other by a wide margin. We do not publish our own pricing, so nothing here should be read as ours.

The comparison people get wrong is one salary against one retainer. A working in-house function at this stage is rarely one person: it is a media owner, someone producing creative at volume, and someone senior enough to decide strategy, often the founder. Add tooling, recruitment and the unproductive early months, and the two options land closer together than the headline numbers suggest.

When in-house is the right answer, plainly

There are brands that should not hire an agency, and pretending otherwise wastes everyone's quarter. Build in-house when:

• Your ad spend is small and stable. Below roughly ₹1 lakh a month, most of a retainer is paying for capability you cannot yet use. PeakPilots says the same thing from the other side of the table.
• The founder is genuinely good at this and has the time. A founder who understands the customer and can read a media report will outperform a junior account manager at any agency, and it is not close.
• The product needs a specific voice nobody outside can carry. Some brands live or die on a founder's voice, community tone or a category so niche that external teams would spend six months learning it.
• Creative is the constraint and you already have a studio. If you shoot in-house and the bottleneck is media execution, that is a hire, not an engagement.
• You want the capability permanently. Buying capability is fine for a phase. If the plan is for performance marketing to be a core competence in three years, start building it now and use outside help to accelerate rather than to replace.

When an agency is the right answer

• You need a diagnosis, not execution. Most stalled accounts are not badly executed, they are structurally wrong: wrong hero SKU, wrong offer, wrong measurement. An account manager will not fix that, and a full-time hire cannot be reversed cheaply if they are the wrong answer to it.
• You need creative volume you cannot staff for. Performance media consumes creative faster than most in-house teams can produce it. This is the single most common reason in-house programmes plateau.
• You need multiple channels at once. Meta, Google, Amazon and quick commerce are four different disciplines. One hire will be good at one of them.
• Speed matters more than ownership right now. Hiring, onboarding and getting to competence is a two to four month process, and it is longer if the first hire does not work out.
• The work is a fixed-length repair. On Kroslo the brief was momentum, not a rebuild: a 25-day sprint that doubled ROAS on the same budget by finding where spend was leaking, with no new positioning, range or site. That is a poor reason to hire and a good reason to buy a short engagement.

Hiring risk is the part founders under-price

A retainer that is not working can end at 30 days notice. A hire that is not working takes longer to recognise, longer to unwind, and costs a rebuild of the account when they leave with everything in their head. At ₹1 to 10 Cr the second failure is materially more expensive than the first, even where the monthly numbers look similar.

There is also an evaluation problem. Performance marketing is hard to interview for if you have not done it yourself: candidates present platform ROAS screenshots from accounts they did not own, and a founder without the vocabulary cannot separate someone who scaled an account from someone who was in the room while it scaled. Buy the diagnosis first, from anyone competent, so you know what you are hiring for.

Speed to competence

A useful way to frame the choice: how long until the function is producing decisions you trust?

• Hire: four to eight weeks to find someone, two to four weeks notice, then six to twelve weeks to learn the category and the account. Four to six months to steady state, if the first hire works.
• Agency: two to four weeks to onboard, and the first structural fixes typically land inside the first month, because they are removing waste rather than creating demand.
• Hybrid: the fastest in practice, because the external side works while the internal side is recruited and trained.

Speed is not automatically decisive. If this capability has to be yours in three years, buying four months of speed at the cost of never building it is a bad trade.

Creative volume is the real constraint

On Meta especially, the creative is the targeting. The real question is not who buys the media, it is who can ship eight to twelve genuinely different concepts a month, indefinitely, and kill the losers without sentiment.

upGrowth makes the same point from the pricing side: it describes creative production as the most expensive line item in a performance retainer, and notes that a mid-tier retainer shipping 30 variants a month needs scripting, editing, motion graphics, creator management and copywriting behind it. That is a production function, not a marketing hire, and it is where most in-house builds quietly stall.

On Parasbaagh, which came in with effectively no recall, five-plus winning creatives held click-through above 3.5%, and that library is the asset that mattered most: a set of proven hooks is what lets spend increase without cost per result climbing with it. Whoever produces that library, in-house or outside, is doing the highest-leverage work in the account.

The hybrid that usually wins between ₹1 and 10 Cr

The version we see work most often is not either extreme. It is one internal owner, often the founder or a marketing generalist, who owns brand, community, retention and merchandising, paired with an outside team for media strategy, creative volume and the analytical layer.

Aazol is a clean example of the shape. The engagement was strategic consultation, performance marketing consultation and a creative strategy revamp rather than pure media buying: the brand was leaning almost entirely on Meta, mostly on automated campaigns, with Google nearly untouched. The work rebalanced the engine across both platforms, identified the SKUs worth building on, and rebuilt the creative. Four months in, CAC was down nearly 70%, ROAS was close to doubled at 2.7×, and revenue had moved past ₹22.5 lakhs a month, leaving a balanced engine an internal team can run.

Two conditions make hybrids work. Decision rights are written down, so both sides know who calls the final shot on budget, offer and creative direction. And the internal owner is senior enough to say no. Hybrids fail when two parties are accountable for the same number.

What actually transfers when an agency leaves

This is the question to ask before you sign, not on the way out. Agree in writing that you keep:

• Ownership of the ad accounts, pixels, Conversions API setup and tag manager container, in your business manager and not the agency's.
• Raw creative files, not only exported MP4s and JPEGs, plus fonts, project files and shoot footage.
• The creative test log: what was tested, what won, what failed, and the hook that made the difference. This is the single most valuable artefact and the one most often lost.
• Audience lists, exclusion lists and the segmentation logic behind them.
• Landing pages, their variants and the CRO test history.
• The reporting model, including how blended CAC and contribution margin are calculated.

What does not transfer is judgement: why a particular angle worked for your customer, and the instinct for when to kill a creative. That is a fair argument against outsourcing the function permanently.

Buy capability while you are still deciding what you need. Build it once you know, and the role is execution rather than diagnosis.

The decision, in four questions

• Is the problem diagnosis or execution? Diagnosis favours buying, execution favours building.
• Can you produce eight to twelve new creative concepts a month, indefinitely? If not, that gap decides the answer.
• Is your spend stable enough to keep a full-time person usefully busy? If not, you are buying idle time.
• Do you want this capability permanently? If yes, start building now and use outside help to shorten the curve rather than to avoid it.

How we work is on the performance marketing, creative and content and strategy pages, the FMCG page, and the studio in Bangalore and Kolkata: 160+ brands, ₹150 Cr+ ad spend managed, ₹450 Cr+ revenue attributed, 3.8× average ROAS over six years.