विमर्श Vimarsh · Udyamnīti

The Argument for India Is a Supply Argument

Everyone arguing for India argues demand. Demand without capacity is a queue. The real argument is on the other side of the ledger, and it is the one I have spent a career inside.

Nikhil Sharma · उद्यमनीति Udyamnīti · 22 July 2026 · 19 min read

I run Fabulous.Media, the agency network this essay describes, and GoCommercially, the productised-services house that prices its work. Read it knowing that.

A queue is not a market

Every argument for India I have heard in twenty years has been an argument about demand.

A billion and a half consumers. The fastest-growing large economy. A median age under thirty. The middle class that is always about to arrive. I have sat through the deck in Mumbai, in Singapore and in Dubai. The slides change. The argument does not. India is a very large number of people who will one day buy things, and you should be there when they do.

The argument has been made since 1991 and it has made fewer people rich than it promised. In May 2007 the McKinsey Global Institute published a report called The Bird of Gold and said India’s middle class would grow more than tenfold, from 50 million people to 583 million, by 2025. It is 2026. In March 2025 Blume Ventures put the real consuming class, the people with discretionary money, at about 14 crore. Ten per cent of the country. Two-thirds of all the discretionary spending. The definitions differ and the gap is still the size of a continent.

I am not saying the demand is false. I am saying it was never the argument. Demand without the capacity to serve it is not a market. It is a queue. People standing in line for a thing that has not been made yet.

Demand without capacity is not a market. It is a queue.

The real argument for India is on the other side of the ledger. It is a supply argument. India now has more capable operators in more cities than any comparable market on earth, and almost no structure that lets them work as one thing. That gap is the one I have spent a career inside, and this essay is what I have learnt from standing in it.

The other side of the ledger

Look at what India actually sells to the world and the demand story falls over on its own.

In the year to March 2025 India exported $387.5 billion of services. That is up 13.6 per cent in a year and it is not far behind the $437.4 billion of goods that the whole of Make in India was built to grow. Nobody bought those services because of India’s median age. They bought them because somebody in Pune or Kochi or Coimbatore could do the work.

The capability centres tell the same story from inside the multinationals. By the Nasscom and Zinnov count there were more than 1,700 global capability centres in India by the 2024 financial year employing 19 lakh people and earning $64.6 billion. A capability centre is the purest supply argument there is. A foreign company looks at India and does not see a customer. It sees the people who will run its engineering, its finance and its risk, and it moves the work here rather than wait for the consumer to show up.

Then look at where the founders are. Of the DPIIT-recognised startups to the end of 2024, 51 per cent were from tier 2 and tier 3 cities. Half. Not Bengaluru, not Gurugram, not the Bandra Kurla Complex. Half the register is in cities the demand deck does not have a slide for.

And my own trade. Advertising spend in India was ₹1,15,291 crore in 2025 by the Pitch Madison count, and 46 per cent of it was digital, up from 15 per cent in 2016. Digital work does not care where the office is. The person who runs a performance campaign in Jaipur is running the same dashboard as the person in Lower Parel, at a third of the rent, and the client in Chennai cannot tell the difference and should not have to.

Four numbers, and not one of them counts a consumer. Each counts a person who can do the work and says where they are. That is the India I would put money on, and have.

What Smith and Marshall knew about where the work sits

None of this is new. The third chapter of The Wealth of Nations has a title that is the whole argument: That the Division of Labour is Limited by the Extent of the Market. Adam Smith meant that a specialist can only exist where there are enough buyers to keep him busy. A porter in a village does everything. A porter in a city does one thing well. The market makes the specialist possible.

A century later Alfred Marshall watched what happened when specialists sat near one another, in Sheffield’s cutlery and Lancashire’s cotton, and wrote the sentence every cluster theorist since has borrowed.

Read the two together and you have the shape of the Indian problem. Smith says the specialist needs a market large enough to feed him. Marshall says specialists get better when they sit together. India has done the second thing by accident, in forty cities, and has not done the first at all. Jaipur has a cluster of digital operators who learn from each other in the air. So does Indore. So does Kochi. Each cluster is limited by the extent of its own market, which is one city, and so each one stays small, under-priced and unable to take a brief that crosses a state line.

The extent of the market for a Jaipur operator should be India. It is Jaipur. Nothing joins the clusters, and so each one is a specialist starved of the market that would let it specialise further. That is the supply gap in one sentence and it is a structural problem, not a talent problem.

Why Fabulous.Media is a network and not an agency

I started Fabulous Media in 2008 as one office. I had been building companies since 2005 and I knew exactly how an agency grows. You win a client in another city, you fly there, you hire two people, you rent a floor, you send someone from the centre to run it for a year and you hope the culture survives the flight. It is a hiring problem. I have done it and it works until about the fourth city, where the hiring runs ahead of the people you can trust to do it.

Somewhere around then I noticed that every city I was flying to already had the people. Not people I could hire. People who already ran a shop, had their own clients, knew the local market better than I ever would, and were held back by exactly one thing. They could not take a national brief. No client in Mumbai was going to hand a pan-India launch to a twelve-person agency in Jaipur, however good, because the client would have to manage eight of them.

So I stopped hiring outward and started building the thing that was missing. Today Fabulous.Media is a network of agency brands across eight Indian cities with Fabulous Media India as the lead agency. Infabio in Jaipur and The Bad Company in Gurugram are the original city brands. Aarmbh Media, Express Digital, Frame Wizards, Infinitium and Marketing Ruler are the rest. Between them the network has worked with over 850 brands across 35 nations. That is our own count, made by pooling the partner firms’ client lists and removing duplicates, so a brand is counted once however many partners served it and a nation once however many brands were billed there. Nobody outside the network has audited it and I would not put it in a prospectus. I will name only the two cities that are confirmed in public today. The other six are real and I would rather under-claim than put a city in an essay before its partner has put it on a door.

Ronald Coase gave me the theory for it in 1937, long before I needed it. His question was why firms exist at all if markets are so efficient, and his answer was that using the market has a cost. Finding the supplier, agreeing the price, writing the contract, checking the work. A firm exists wherever doing the thing inside is cheaper than buying it outside. And then the line that decides the whole question of agency against network.

Note the last clause. Or the costs of organising in another firm. Coase saw the third option that most people miss. Not hire, not buy, but arrange it so that another firm does the work under your standard. The agency model says hire. The marketplace says buy. The network says the eighth city already has a firm, so organise it. Which one is right depends entirely on which of the three costs is lowest, and in India in 2026 the answer is not close. Hiring in eight cities is expensive and slow. Buying in eight cities on the open market is cheap and uncontrolled. Organising eight firms that already exist is the cheapest capacity in the country, if you can solve one problem.

If that sounds exotic, look at the four largest professional-services brands on earth. Deloitte, PwC, EY and KPMG are not companies. Each is a network of legally separate member firms, broadly one per country, held together by a central entity that does no client work. Deloitte says so in its own governance pages: DTTL and each of its member firms are legally separate and independent entities, and DTTL does not provide services to clients. PwC says the same of PwC International Limited, which does not practise accountancy or provide services to clients and exists so that the member firms abide by common policies and maintain the standards of the network. Nearly 500,000 people in Deloitte’s case, under one name, with no common payroll. The largest services brands in the world run the model I am describing and have for a century. They also carry its bill. An audit failure in one member firm lands on the brand in every other, which is the standards problem below at planetary scale.

The one problem is trust, and trust is the slower asset.

What each model costs and what each can win

I am going to be fair to the agency model because I ran one for years and the people who run them are not fools.

The agency wins on control. One payroll, one culture, one person who can be fired. A decision taken in Gurugram on Monday morning is being executed in the Bengaluru office by lunch, because the Bengaluru office works for you. Quality is consistent because the same people trained everybody. A confidential pitch stays confidential because everyone who has seen it is on your payroll. When a client calls at eleven at night with a crisis, one phone call moves the whole company.

The agency pays for that control in three ways. It carries fixed cost in every city whether the city has work or not. It can only ever be as good as the people it can hire, and the best operator in Kochi is not looking for a job, he is running a shop. And it hits a ceiling at the point where the founder can no longer personally know the person running each office. Past that point the agency becomes a holding company with a brand on it.

The network wins on supply. The talent is already there, already employed, already carrying its own cost. Local knowledge is real and not learnt from a deck. The fixed cost of eight cities is spread across eight balance sheets. And it can take the national brief, which is the one thing none of its members could do alone and the whole reason it exists.

The network pays in time and in margin. Trust is built in years and lost in an afternoon. Standards drift unless someone is paid to stop them drifting. Every rupee of a national brief is split, so the centre earns less than an agency would on the same work and has to earn it on volume instead. Decisions are slower because eight owners have a view and the centre has no right to overrule them, only to persuade.

And the network is worse at some things, and I will say which. A forty-eight-hour brief. A crisis. A confidential pitch against a client’s competitor who is also a partner’s client in another city. Anything that needs one hand on one wheel. For those the agency is the right instrument and I will tell a client so. The network is the right instrument for the thing that is actually most of Indian marketing. Sustained, multi-city, standards-driven work that has to look the same in Jaipur and in Pune and costs the client one conversation.

The spine

If the cities already have the people, what exactly does the centre build. Four things, and they are duller than any deck.

Standards. What a finished piece of work looks like, written down, so that a client in Chennai receives from Jaipur what she received from Gurugram. Not a brand book. A checklist with edges that a partner can be held to and the centre can be held to as well.

One published scope. This is the GoCommercially half of the argument and the half most agencies will not touch. Indian services are priced in the room. The price depends on who is asking and how badly they want it and whether the account director had a good quarter. A network cannot run that way because eight partners quoting eight prices for one brief is not a network, it is a bazaar. GoCommercially exists to put the price on the tin. Fixed scope, fixed price, same in every city. I should say plainly that the list is not yet public. It is being written and I have promised it in public more than once. Until it is on the site you should treat that promise as unkept, because it is.

Account discipline. One person who owns the client across all eight cities and answers for the whole, so the client never manages the network. The day the client has to call Jaipur herself the network has failed at the only thing it was for.

The national brief. The ability to say yes to a pan-India launch on a Tuesday and have eight cities moving by Thursday without a hiring plan. That is the product. Everything else is the machinery that makes it true.

The Hitopadeśa has the verse for this and it is better than anything I have written above.

Grass does not hold an elephant. A rope made of grass does. The verse is not about the grass and it is not about the elephant. It is about the twisting, which is the only part that has to be done on purpose. The twisting is the spine.

The grass already exists in every city. The rope is the only thing I have to make.

Networks fail more often than agencies

This is the section that keeps the essay honest and I would rather write it than have someone else write it for me.

Networks fail more often than agencies. The management literature on alliances has said so for decades. Prashant Kale and Harbir Singh, reviewing the field in the Academy of Management Perspectives in 2009, open with the number. Studies have shown that between 30% and 70% of alliances fail; they neither meet the goals of their parent companies nor deliver on the operational or strategic benefits they were set up for. Take the midpoint and half of these arrangements do not work. An agency with one owner fails too, but it fails for ordinary reasons. A network has three extra ways to die and all three are structural.

The local partner defects. A client comes through the network, the partner does the work well, and in year two the client and the partner quietly agree that the centre is a tax. The partner takes the client direct. Nothing illegal has happened. The network has simply been used as a sales channel by a member who no longer needs it. This has happened to us. It will happen again. The only defence is a centre that is visibly worth its share every quarter, and a contract that was written while everyone still liked each other.

Standards drift. Eight cities, eight habits. A partner under pressure cuts a corner nobody sees until the client does. The centre finds out from the client, which is the worst way to find out anything. This has also happened to us, and more than once, and every time the cause was the same. The standard existed and nobody was paid to check it. A standard that nobody is paid to check is a hope.

The centre over-charges for services nobody wanted. This is the quiet one and the one I watch hardest in myself. A centre that has to justify its share starts inventing things to charge for. A dashboard. A monthly review. A brand audit the partner did not ask for. The partners pay because they are told to, and the resentment finances the first failure mode. I have caught the network doing this at least once and I caught it late. The honest fix is the tin again. If the centre’s services are on a tin with a scope and a price, a partner can decline them, and a service that every partner declines was never a service.

There is a fourth failure and it belongs to me rather than to the model. A network is only as good as the man at the centre’s ability to stay out of the operator’s chair in eight cities at once. I hold one rule for the founders I back: I never sit in the operator’s chair. The network tests that rule eight times a week. Some weeks I fail it.

Why trust is the slower asset, and the better one

Hiring is fast and trust is slow, and every instinct a founder has says to do the fast thing. I want to explain why I keep doing the slow one.

Walter Powell wrote a paper in 1990 called Neither Market nor Hierarchy, and the argument is that networks are a third form of organisation and not a halfway house between the other two. His test for when you are looking at one is the best description of my working week I have read.

Obligation and reputation. No common ownership. Not easily measured. That is eight agency owners and one centre, written down by a sociologist thirty-six years ago, and it has held up for every one of them.

A hire is a contract. It costs the same to enforce in year five as in year one. Trust is different. It is expensive to build and almost free to use once built. The partner in Jaipur I have worked with for a decade does not need a brief. He needs a phone call. The one I signed last year needs a document, a review and a second review. By year five, if we both behave, the document gets shorter. Trust is the only asset on my books that gets cheaper to use every year I hold it, and it is the only one a competitor cannot buy from me with a cheque.

That is also why it compounds across the ventures. NS Transform will carry up to ₹1 crore of a founder’s operating cost in return for equity. The crore matters. But what the founder is actually plugged into is the network, the distribution across eight cities that took eighteen years to make trustworthy and could not be assembled in eighteen months at any price. The money is the fast asset. The network is the slow one. A founder can raise money anywhere. He cannot raise trust.

Money is the fast asset. Trust is the slow one. Only one of them cannot be bought.

My declared lens is Rāshtrahit, India’s interest, and I will state once where it sits in this argument before I bet on it.

If Indian services stay structurally one-office-per-brand, the national briefs go to the only organisations that can take them, which are the global holding companies, and the operator in Jaipur stays a subcontractor to a subcontractor. The margin leaves. The standard is set somewhere else. The 51 per cent of founders outside the metros stay limited by the extent of their own city and the division of labour stops where Smith said it would.

If, instead, Indian service businesses learn to run networks, the capacity that already exists in forty cities becomes one national supply, priced openly, held to one standard, able to take the brief. That is worth more to the country than the next consumer survey, because it converts the queue into a market. The demand is standing there. The supply is standing there. The only thing missing is the rope.

The wager

An argument about capacity should be scored on capacity, so here is the claim.

I am saying that the operator depth I have described is real, and that the proof will be networks, not agencies. Specifically.

By 31 March 2029 at least three Indian service businesses headquartered outside the metros will be operating a national network of five or more city partners under one published scope and one quality standard. Not franchises of a foreign brand. Not holding companies that bought the offices. Networks of independent firms, led from a city that is not Mumbai, Delhi, Bengaluru, Hyderabad, Chennai or Kolkata, with a price on the tin and a standard anyone can read.

Three markers for whether it is coming, all public. 1. Published scopes and prices for business services become normal, visible on the sites of Indian agencies and professional firms rather than hidden behind a form. 2. At least one national brief from a top-hundred Indian advertiser goes to a network led from a non-metro city, and the advertiser says so. 3. A sector body or a serious trade publication starts counting networks as a category distinct from agencies, because you only count a thing once it exists.

My call, first published 22 July 2026.

If by March 2029 Indian services are still structurally one office per brand, the operator depth I am betting on is not real yet, or it is real and cannot be organised, and either way the premise of this essay is weaker than I think. I will write that essay and link it from this one. And I will hold my own network to the same date. If GoCommercially’s scopes and prices are not public by then, I lost the wager at home before anyone else could score it.

The argument for India was never that people would buy. It is that people can already do the work, and nobody has joined them up.

Nikhil Sharma, writing as Nishkrant Nikhil. Gurugram, 22 July 2026.

Sources: McKinsey Global Institute, The Bird of Gold: The Rise of India’s Consumer Market, May 2007, for the 50 million to 583 million projection. Blume Ventures’ Indus Valley Report 2025 as reported by The News Minute, 1 March 2025, for the 14 crore consuming class and its two-thirds share of discretionary spending. RBI export figures for 2024-25 via News on AIR, 2 May 2025, for $387.5 billion of services and $437.4 billion of goods. Nasscom-Zinnov on global capability centres via ThePrint, 15 September 2024, for 1,700+ centres, 1.9 million employees and $64.6 billion. Startup India factbook to 31 December 2024 for the 51 per cent tier 2 and 3 share. Pitch Madison Advertising Report 2026 via exchange4media, 24 February 2026, for ₹1,15,291 crore and digital’s 46 per cent. Kale and Singh, Managing Strategic Alliances: What Do We Know Now, and Where Do We Go from Here?, Academy of Management Perspectives 23(3), August 2009, pages 45 to 62; the failure range is on page 45. Powell, Neither Market nor Hierarchy: Network Forms of Organization, Research in Organizational Behavior 12, 1990, pages 295 to 336; the passage quoted is on page 301. Deloitte on its network structure and PwC on its network structure, each firm’s own governance page. Smith, Wealth of Nations, Book I chapter III. Marshall, Principles of Economics, Book IV chapter X section 3. Coase, The Nature of the Firm, Economica 1937, page 395. Hitopadeśa 1.35 in the Wisdom Library text; verse numbering varies by edition. Fabulous.Media client and country counts are the network’s own and have not been independently audited.