Grokking India: High fixed cost per transaction


This blog post is really worthy only of a tweet since the core lesson is simple and either you grok it or you don’t; either you believe it or you don’t. There really is little middle ground.

I have previously written about hidden costs of starting up a business in India, I’ve written about accounting for cost of collections in calculating LTV and I have written about the unique challenges in building a distribution network in India. The common theme, the underlying economic fact of India, is as follows:


OK, that’s all I have to say. Thank you for reading.

Juuuust kidding! This idea is very important to understand and in the context of VC funded companies has a non-trivial impact on capital requirements, scalability and therefore returns. Let me re-hash a few examples I have given previously to explain this further:

  1. As you go from B2B to B2B2C (FMCG) to B2C, the revenue per transaction keeps decreasing, but after a point the fixed costs per transaction (cost of collections, returns, customer service, defaults, high friction courier etc) do not go down. Therefore unit profitability and/or scalability on B2C continues to be a challenge. B2B2C (FMCG companies) are big and hugely profitable and it is no surprise that their collections and deliveries are to consolidators of volume and therefore they are able to profitably pay for the fixed cost per transaction.
  2. VC investments in internet businesses have to be thought about like it is ’94 in the US. The early stage bets are big even to just prove out the concept because like in ’94 in the US, companies have to first build their operating eco-system in order to even validate their ideas. Borrowing the present-day methodology in Silicon Valley of doing $250K experiments in more cases than not will not tell you anything in India. In other words there is a high fixed investment amount for early stage transactions.
  3. Lack of trust: Consumers trusting businesses, citizens trusting laws and law enforcement, businesses trusting each other – due to systemic issues India is characterized by low trust and that almost always requires face-to-face contact and physical access to close all kinds of transactions. Even business transactions require much higher level of diligence due to worries of being cheated. This is again HIGH fixed cost for many different kinds of transactions.

And the list goes on.  Analyzing the sources of the fixed costs can tell you what needs to be done to reduce the fixed costs. And there you find all the usual suspects, for example:

  1. Electronic payment systems need to be in place (to completely “variablaize” cost of collections)
    • which in turn require consumer trust
      • which in turn require reliable credit agencies
        • which in turn require trustworthy laws and law enforcement
    • OR you can build a telecom company that does electronic payments. A telecom company has the consumer’s trust and obviates the need for an EXTERNAL credit agency or good law enforcement around payments (Put it another way: the fixed cost of building a national level electronic payment network is building a telecom company!).
  2. Better infrastructure: good roads (reduced fixed costs of transportation),  good power systems (reduce fixed cost of operations), etc
  3. Trustworthy law enforcement so that there is implicit trust in legal contracts

These fixed costs are worth reducing as they have non-trivial implications, including:

  1. Unleashing innovation as cost and risks of experimentation goes down; with high fixed costs experimentation becomes tougher
  2. Improving investor returns as the bets can be staged better
  3. Decreaing time to scale for ideas that show potential because lower fixed costs imply a “lower friction” operating environment

For all these reasons, India is, and will remain in the foreseeable future, an economy dominated by operational complexity not strategic complexity. The big question in India is HOW to get products and services to consumers not WHAT. In other words India is not demand constrained but supply constrained and keeping an eye on the fixed costs in the system can help navigate the complexities of doing business while staying true to needs of profitability, scalability and investor returns.

Apple vs Google: vertical integration vs crowd-sourcing


I am a big fan of both Apple and Google and from the distance see two very different DNAs that are good at solving a very distinct set of problems. Apple is the epitome of vertical integration owning the end to end supply-chain including the consumer touch point. Google is the epitome of crowd-sourced development: throwing stuff on the walls and letting users vote with their time & money on what sticks and what does not.

In this post I am going to make some predictions on what kinds of products and services each company will win at and what might be tough for them.


Let me start with Apple. The biggest reason I use Apple products is ease-of-use around consumption of media. From being able to buy content (music, TV shows, movies) easily on an a-la-carte basis, to being able to consume them without ads, or consume them in any place (on the iPhone, TV – through AirPlay or streaming, iPad), I really can just enjoy the media rather than having to stitch together my custom solution. I would go as far as to say that post iPod, Apple has not just given me a music player and an online music store, but what they have really enabled is allowing me to connect with music; something that is challenging in a hectic, internet enabled startup life. They have taken full ownership of the customer experience, simplified it and let anyone in my family (including my kids) easily access the content anywhere. To be fair, I have not tried Google Play or Amazon, but I know that out of the box Apple will work and it makes it that much harder for me to give the other solutions a try. Add to this apps that give access to specific content ( was my most recent purchase) and it becomes a killer solution. This requires an unprecedented level of vertical integration: from relationships with content owners, to knowing every detail of the manufacturing supply-chain (going to the extent of locking out certain components from the rest of the market by making investments in its suppliers), to the software and distribution to the consumer. The complexity of what Apple has been able to achieve is awe inspiring. I am personally looking forward to the disruptions Apple will bring to TV viewing.

A lot has been talked about the design aesthetics of Apple products and while there is no question about the elegance of each product, I don’t think without the benefit of the solution this would have mattered as much (anyone remember Vertu?). On the strength of the underlying benefit, IMO, the design takes the solution to another level and builds a strong competitive moat for the core money making activities.

Yet, this strength around vertical integration leaves it vulnerable is some critical areas where it will try and own the customer experience but IMO won’t be able to beat a crowd-sourced model that easily. Three key areas come to mind immediately: Maps, Voice & Browser. I am quite worried about Apple substituting Google Maps because the strength of Google Maps comes from crowd-sourcing content and it will continue to get better each time it is used. I don’t know if Apple’s DNA will allow that kind of continuous improvement by users. The same logic applies to Voice Search. I was never excited by Siri and recent reports that Google Voice Search is better don’t surprise me at all. And then the browser. I knew someday I would have to pay for Apple in a way I will not like. The new Chrome browser for the iPhone/iPad is much better than Safari and I think will continue to get better but Apple has not yet allowed to make it the default on the iPhone or iPad.


I really like GMail & Google Apps. The way GMail has incrementally improved over time has just been amazing. So many little usability issues and features keep getting fixed and with labs there is always something new to try out and vote on. I’d be loath to try a competitive email offering from Apple – I just don’t think they will be able to pull it off. Search of course is the ultimate product that has improved because of crowd sourcing. I am also extremely optimistic about G+. I am really hoping that they will do a much better job than Facebook  by first doing what is right for the user. Can you imagine Apple doing a social network – you know Ping?! Social network for music? Whatever. In some ways Apple was smart to just simply integrate tightly with Twitter and Facebook. I was once at a Google office and was surprised to see the person I was visiting  with an iPhone. This person was quick to point out that Google encourages use of all competitive products so that employees are aware of everything that is happening. Imagine if this was Apple – the employee would have been fired!

And yet Google’s approach has limitations. I am not surprised by the poor reviews Google TV has received – it is not hosted software that can be incrementally improved behind data. And while Android will most likely be THE most dominant mobile operating system, it will achieve this primarily because of price. It is again no surprise to me that revenue per user is much higher on the Apple App Store than it is on Google Play. Similar arguments apply to the Chromebook – conceptually an interesting model and idea, but tough to pull off until Google is willing to own every detail of the customer experience all the way to the manufacturing of components. In the case of Apple the ownership went all the way to the CEO not a product manager. That is powerful.


One thing that I really like about Google is how much it tried to create stickiness primarily by trying to do right by the user. It is commendable. But as the company faces competitive threats it is being forced to take a less idealized view leading to users complaining and in fact some employees leaving. It is a harsh reality of the environment. But yet, there is enough tension in their DNA that they will (hopefully) try to do the right thing.

Apple has always wanted to keep the user in its walled garden and it rubs a lot of people the wrong way. In the pre-iPod world it rubbed me the wrong way too but today the story is quite different because it has out-innovated the market and gave me something I could not refuse. But the tendency is still strong and worrisome.

Truth is that every company tries to create switching costs for the user – some the user is willing to live with, some not. But no company escapes the need to do this.  For the user the good news of course is that there are competing choices which do their part in keeping each company honest. As a user I am excited by the choices and I hope that I will continue to have the option to mix and match as I please without too many hindrances.

Professional Photography Is Dead. Long Live Professional Photography!


I did a post on about what I think about the future of pro-photography. Here is a small excerpt from it:

“… never before have average consumers been exposed to this much high quality imagery – web, mobile, magazines, hoardings, blogs – you name it. The quality of images on Pinterest or Tumblr is just mind blowing. Sub-consciously consumers have developed a greater appreciation for good images.  … [Consumers] are coming to appreciate how difficult the art and craft of photography is. So when it comes to important life events (operative words being “important life events”), my prediction is that over the next decade pros will be hired in record numbers. Consumers will be spending in record amounts. 

Check out the full post on here

Airtel Money is potentially revolutionary


This post also appeared on the VCCircle blog here

Airtel Money is the first time in 5 years that I have truly become excited about a payment solution for the India market. I think this can be revolutionary. At Canvera we have evaluated each and every payment type available in India and have looked at the end to end cost, scalability and reliability. We concluded that managing cash (not COD) and encouraging online payments through debit and credit cards was most suited for our business. Building our own payment network was by no stretch in our business plan so it was disappointing and costly to have to do it. But in the process we got insights into the unique challenges of a payments company and that is why as soon as Airtel Money came about, we realized the potential for it to completely change the landscape. Let me explain why.


Broadly speaking there are two specific use cases:

  1. FOR POST-PAID ACCOUNTS: Statements can potentially become like credit card statements where the telecom companies may be able to assign certain credit limits based on payment history, guarantees etc.
  2. PRE-LOADED ACCOUNTS: both pre-paid and post-paid accounts charge a certain amount for payments (this is what is being offered today).

Consumers would use their mobile phone to make payments just as credit/debit cards are used today and merchants such as Canvera sign up to receive payments.


In evaluating the different payment solutions in India we understood some of their limitations. These included:

  1. Lack of consumer trust in the instrument: as a merchant we felt that we would have to support a number of payment options none of which had any significant traction. In fact in some cases we felt that we would need to build that trust in our customers and got in to talks with a few of the payment companies but eventually dropped the idea
  2. Lack of distribution capabilities: distribution at scale is possibly the minimum requirement to building trust. But more than that the payment method needs to be easily accessible and all use cases around the instrument need to be simple. A lot of the alternate payments services failed on this front.
  3. Convenience: A lot of the payment solutions had just too many steps for the consumer that there was no way that average consumer was going to use these solutions regularly. Some of the issues stemmed from regulation and were out of control of the payment company but the net result was that they asked too much of the consumer.

The telecom companies are uniquely poised to overcome these challenges. They already are known brands and have trust of the consumer; and they have significant distribution capabilities to provide physical touch points for payments, re-charges, customer queries, issue resolution etc.  No startup can quite compete with the investments that telecom companies have made in establishing their reach and brand.  And finally the convenience of paying with a mobile phone – who can top that!


They are of course many hurdles. The two key ones are:

  1. NEW PRODUCT & SERVICE DEVELOPMENT: Can the telecom companies truly create an internal environment where a brand new product and service offering can be developed, improved and deployed at scale? As a merchant I sure hope so, but it is a challenge nevertheless. This new product /service development will require a completely new mindset and skills while leveraging the power of the present infrastructure of the telecom company.
  2. REGULATORY: Will the telecom companies be allowed to essentially become banks providing credit to consumers? Again, my hope is that this will be allowed, with appropriate but minimal regulation to protect consumers. Given the clout telecom companies have within the government, no one else is better poised to lobby the forces-that-be to make this happen.  Certainly no startup can develop that level of clout and reach within the government.

While I don’t want to trivialize the above two challenges, I do feel that compared to the challenges in front of a payment startup these are surmountable in a “reasonable” time frame and if overcome can very quickly change the payments landscape in India.

In addition to allowing the new product and service to be developed fully I do also hope that telecom companies learn the lessons from the mVas debacles that by taking too much away from the innovators they actually kill the innovation. I hope that they will learn to play the role of a neutral and cost effective platform and not compete or interfere with the free market competition between the merchants.  Done correctly this can potentially even correct the ARPU issues that telecom companies face in India.


India is a country of great diversity and yet, due to number of structural issues, it is virtually impossible for a single service provider to PROFITABLY service the very long tail of consumer preference. A key enabling piece (amongst many others) to serving the long-tail needs is an electronic payment network that allows merchants to profitably collect small amounts of money. The low penetration of credit cards and lack of growth, lack of trust in the infrastructure and law enforcement , the challenges to create a national credit rating system etc meant that solutions that have worked in developed markets will take decades to reach scale in India. A telecom company also becoming a payments company and being able to provide credit can dramatically shorten the timeframe to achieve the same result.  A unique solution to a unique operating environment that could potentially unleash long tail creativity.

I am rooting for this one!

Distribution Is King. Much More So In India


This post first appeared on the VCCircle blog here


The easiest way to tell if a PE/VC investor has earned their chops in India or in the west is when within 15 mins of your conversation the former will ask you: “Now that you have created your distribution network in India (or plan to create it) what else beyond your present offering can you pump through it?” This test has never failed me! There is a lot of wisdom behind this bias because the toughest thing to do in India is to build out the distribution network. A US trained VC’s bias will be towards vertical growth within the same product suite, but an India trained VC will have a bias towards horizontal growth across multiple product suites. I would argue that the challenges of distribution are so huge in India that product innovation is secondary and it is more capital efficient to “import” products and figure out the distribution network rather than spend time and energy on innovation.

Today we are seeing an angel investing boom in the US. There is a lot of infrastructure in place that allows this to happen. Amazon Web Services and the ilk are often mentioned as the key enablers but what does not get mentioned as often are: variabalized sales/marketing costs through online marketing; reliable physical infrastructure; reliable courier services; an electronic payments network; trustworthy law enforcement; overall high level of trust in society etc. All these are different elements of distribution that have on a relative scale commoditized distribution for an entrepreneur so that she and her team can focus on the core product. In essence, it has freed up a lot of energy to focus on pure innovation.

In this post, through a comparison to the US market, I will simply highlight the unique distribution challenges of India. The key to success in India is to be aware of these challenges and develop plans to work within the constraints they force upon you. With present-day low levels of per-capita consumption, domestic consumption in India will continue to grow for decades and companies that address the India market at its own terms rather than following copy-cat business models are the ones that are going to capitalize on this growth.

Let me start with listing down the various elements of a successful distribution channel. I would say that the distribution channel for a brand has been setup if ALL of the following have been accomplished:

  1. TRUST: Consumers and channels have basic trust in the brand to try it out repeatedly and recommend it to others
  2. RELIABLE DELIVERY: The company has the ability to deliver the product/service reliably day in and day out
  3. ABILITY TO COLLECT: The company has the ability to collect money, which also includes ability to enforce a financial obligation (hopefully through legal means!)
  4. PROFITABILITY: At its simplest, this measures the firm’s ability to charge the right amount and the operational efficiency of the channels to keep costs down.
  5. SCALE: The business needs to have reasonable scale – exact revenue numbers will depend on the type of business, but north of $5-10M is a good first order estimate in the context of a VC- funded company

So let’s compare the challenges on all of these fronts between India and US.


Any time a new company is formed there is resistance to accept the new product or service and overcoming this through great delivery, marketing, sales etc costs a lot of money in every country. But a major issue in India is the low level of trust that exists in businesses due to daily poor experiences. Think of your favourite home service (plumbing, electrician, car repair) or banking services (“I will call you back tomorrow” – yeah right!) or horror stories getting medical care – the default starting point for the consumer is to assume “it will not work” or “s/he is lying or stretching what’s actually possible”. A consumer starts with a position of mistrust and then trust has to be earned. It is the exact opposite of what happens in the US. If I wanted to buy an expensive camera in the US, I would do a Google search, quickly find the best price and place the order. There is a 99% guarantee that the product I ordered will get delivered in the time committed and that I will have full protection for the 1% of the time there is a problem. I’d walk in to brand new restaurant on day-1 without any fear of falling sick.

In India, overcoming the very high level of mistrust at scale is at best extremely expensive and at worst takes a lot of time and staying power. It requires delivery at scale (which is a chicken and egg problem), advertising, marketing & sales at scale, and good corporate behavior at all company touch points.


The concept of “reliable” delivery means many different things in different businesses but let me take Canvera’s example of delivering photographic printed products to customers. We work closely with 15+ courier companies (not 2 or 3, but 15) that deliver mail-order product to 300+ cities. Over a 4 year period we have developed intricate processes to monitor each packet over and above the monitoring and tracking the courier company provides because in the end it is our responsibility to deliver the packet in a respectable timeframe and through experience we’ve learnt that we can’t rely on the courier companies. In a services business (think restaurants) in India scaling up is a huge challenge because of the quality of the available workforce and the training required (think hygiene in a restaurant). In the US, courier services are a given, no entrepreneur worries about them.  And while finding high quality workforce in a services business is always a challenge, the size of the US economy itself provides a reasonably large trained workforce. It may cost a pretty penny, but it is available. Think of how quickly Groupon scaled its team of Inside Sales to 1000s of people and how many different businesses were able to reliably provide the services to consumers through Groupon.


I have written at length about COD in an earlier post – that was more in the context of a B2C business. But there are other forms of collections including B2B, retail etc. FMCG businesses in India manage consumer collections by offloading them to the kirana store and managing their own collections from the store owner by creating enough demand for the product and holding back supplies if payments are not received. B2B collections are a nightmare in India that create massive cash flow issues with providers being forced to get in to the credit business since there really isn’t a legally enforceable collections framework and companies rely on personal relationships and holding back future services to enforce payments. No entrepreneur I know of in the US worries about collections as a bottleneck to business growth. It is a plug and play solution and largely a variable cost with a legal framework to fall back on.


Let’s examine the two aspects of profitability: Revenue and Cost. Take mail order businesses (or e-commerce);  consumers in the US are willing to pay a little extra for shipping solely for the convenience – as the economy has grown and time becomes a premium, there is an appreciation for opportunity cost and value based services that save time and headache. On the cost side – assuming healthy gross margins, given a reasonable cost of RELIABLE shipping and reasonable cost of collections, each transaction can be profitable in the US. In India however it is a very big assumption that consumers will pay for convenience (this is a separate blog post altogether) – real and perceived opportunity costs are low and even if they increase, it takes a lot of time for consumer behavior to change at scale. So far, especially in e-commerce, all we have seen is what we know sub-consciously: price is all that matters. I am willing to bet that volumes will drop dramatically if Indian e-commerce companies start charging nominal amounts for shipping or if prices are raised to even a few % points higher than offline retail. And then on the cost front, as I have mentioned before, cost of collections is high and logistics are very inefficient. The combination of not being able to charge for convenience/quality and a high friction operating environment results in highly unprofitable transactions.


Finally, any company can overcome all of the above at a small size, the real challenge is being able to do this on a large scale: in multiple geographies, in multiple languages all over India.

ignore underlying infrastructure at your own risk

Telecom companies are the best example of the power of distribution networks in India. This blog post by Abhimanyu Radhakrishnan summarizes it best:

The speaker line-up at the opening keynote … is the best evidence possible of the peculiarity of the mobile phone industry. The operator – the guy who provides you an easily commoditized, utility like service – has the biggest clout in the industry. Imagine the premier global conference for …umm… ‘Home Appliances’ having the CEOs of the electricity companies as the big draws.

How true! And yet the power is with the phone company because they have the distribution network. They have the trust of the consumer; they provide mVas players with a reliable delivery and collections mechanism which can in turn be profitable. And they have significant scale. No telecom player wants to be seen just as the dumb pipe. After having put in the capital and time to build out this network, they want a share of every innovative product that is distributed through it. It is a separate matter that their behavior and expectations are preventing innovation but the bargaining power is clearly in their hands (Side note: telecom companies are behaving like an investor in a capital constrained market who wants to invest at $0M pre-money. I digress).

This discussion is by no means suggesting that it is easy to build a business in a developed market or impossible to do it in emerging markets like India. In fact, the barriers of distribution that have been lowered in developed markets are lowered for all the entrepreneurs and therefore competition can be much higher as well. Product innovation and speed then become key differentiators.  India is unique; India is not the US, nor is it China or Brazil. In India FMCG companies have shown how a leveraged offline model of using kirana stores to offload a lot of these distribution costs and last mile headaches can lead to big defensible, scalable and profitable businesses. There is much to learn from these success stories.

The one cautionary fact for all entrepreneurs and investors to note is that TIME is the biggest price to pay for the lack of infrastructure. It is extremely difficult to overcome the barriers through just capital. There are no shortcuts and setting up the foundation correctly is absolutely critical to building defensible scale. It has been done, it IS being done and my money is behind the people who are keeping their eyes open to these challenges.

Why I don’t believe in COD


This post first appeared on the VCCircle blog here.


Payment systems (or lack of them) are a very big problem for India e-commerce and as much as COD is being touted as the panacea to get over the hurdles of reach and consumer trust, I am just not able to buy the story. To me it seems like optimizations to COD are like trying to make the horse run faster while what we need is an automobile (sorry Henry Ford!). Stretching the analogy, we not only need the automobile, we also need the support infrastructure of petrol bunks, repair shops, credit agencies etc. The equivalent of the automobiles in payments are well penetrated and reliable electronic payment systems, which in turn require good credit rating agencies, trust worthy law enforcement etc. A very tall order to achieve in a few years when there is generally low trust, low penetration and insipid growth if any (credit card penetration is stuck at 20M for last few years) of electronic payment systems.

I am firmly in the camp of people who believe that the e-commerce bubble is going to burst soon and while I think there are many fundamental reasons why this will happen, if you forced me to pick only one I would pick lack of electronic payment systems in India.

Before I explain any further, let me state two key assumptions:

  1. I am primarily talking about COD in the context of high volume, low gross margin (<Rs1000) products
  2. I am assuming that mid-term profitability is important

If both these assumptions don’t apply to a business then you can ignore my analysis.

Sources of Fixed Cost in COD

I explained the LTV impact of COD in a previous post. Let me expand on this further and list down the sources of fixed cost per collection:

  1. Cost of physically traveling to collect money: Number of trips that need to be made to collect the money and the cost of these trips especially given high petrol prices
  2. Cost of consumer changing their mind: If the money is not charged immediately when the consumer is in the buying window, large % will change their mind when the order is delivered (anecdotally I have heard 30% “return” rates being quoted in the industry today)
  3. Cost of managing cash and reconciliation: This requires people based processes to count, store and deposit money in the bank in addition to manually reconciling the money with the right order
  4. Fraud: This has many sources, including the collections person running away with the money, fake currency notes, collusion between  the collections team, law enforcement issues with someone carrying significant cash etc.
  5. Hiring, re-hiring, training and management of the collections team. This will be a huge fixed cost borne by the company

All of these translate to a FIXED COST PER TRANACTION, regardless of the money being transacted. Just this fact can make the business unprofitable unless, and this is important, that the cost of collection can be completely “variabalized”. There is no other way to do this except to switch to a completely electronic payment system.

Scalability challenges

And finally, there is the scale challenge. If the GMV value (not revenues – and can we please stop calling GMV revenues, it discredits everyone in the startup eco-system!) of e-commerce transactions in India is to reach $1B annually (so ~$100M in revenues) and the average GMV value is $20 then we need 50M transactions – even if 60% of these are based on COD, it means 30M transactions of $20 at a time being handled and reconciled by people. Let me say it again, for $100M in revenues there will be 30M POPLE-BASED transactions to collect and reconcile the money. There will be management layers to manage the team, hiring and retention (and replacement will be an issue) as the company will require low-skilled workers to try and make the math work. A rudimentary calculation assuming 6 collections per person per day means that the front-end collections team will have to be of about 15,000 people and probably half the size of this team to do the backend operations. Most of the front end team will probably earn Rs5-7K/month. To keep a reliable force out in the field (not in a factory where it is easier to manage a team) there will be significant management overhead and constantly having a pipeline of 10-20X  the team strength (i.e. 150K-300K people) for replacements. Then there are the costs of career planning and growth for this work force.

It just is really hard to see how this will work when transactions gross margins are low and nothing that I have heard or seen in the last 9 months of this bubble is able to convince me otherwise. The COD payments business is a services business and in order for the math to work, it needs to quickly transform to a products business, i.e. be replaced by an electronic payments system. That will take a lot of time. A.Lot.Of.Time.