Rise of competition in the food delivery business

He’s right, of course.

But, by the same benchmark, there aren’t too many examples of the Zomato model working either.

One of the reasons why it’s hard to make money from the ad sales model in food-tech is because you need a large sales team to make it work. The economics aren’t as bad as that of having a delivery fleet, but it’s quite formidable. This is what companies like Zomato and Yelp need to reckon with.

Of course, Deepinder does point out success stories in Japan, and while the jury is out, it’s quite possible the Zomato model may eventually work. Maybe. But it will likely take a while.

Unfortunately, you don’t have time. It’s important you get a working model immediately. So you decide to try something new. It’s daring, it’s risky, but it’s your best chance of success.

You give a rousing speech in the employee town hall about this new strategy. Fifteen employees resign the next day.

But you soldier on.

The Freshmenu model

In food, the biggest margin is inside the kitchen. It’s the difference between the cost of black dal in the market and the price of lentil soup on the menu. This is what funds everything else in the restaurant.

But here’s the problem, if you make the food yourself, you still need to find customers and you still need to make deliveries. So you end up fighting on all fronts, and the only way you can win is if your food is memorable, not too expensive, and reliably available.

It’s hard and it takes a lot of time, but if you manage to make it work, it can solve all your profitability problems. But it requires a massive one-time investment to solve a systemic problem, with the promise of reaping enormous, long-term rewards.

Your investors hear your plan, and point out that you are not the BCCI. They urge you to try something milder. Like a proof of concept. Just to see if it will work.

The Ola Cafe model

One reason why food-tech so expensive is because the customer has too many options which prevent you from doing any preparation. Your typical customer orders Chinese one day, a sandwich the next day and on weekends, he tries to convince you to get him tea and cigarettes on the way.

So there’s a simple way out. If you can somehow narrow down the options that a customer has to order, you can optimise delivery and perhaps start to break-even. If you have four or five options, your delivery guys can stock it with them instead of running to the restaurant for every order.

Remember – preparedness is profit.

Worth trying, you think.

But the bad news is that the Ola model didn’t work for Ola. It was shut down within a year of launch.

Ola Store, the groceries delivery service, and Ola Cafe, the food delivery service have failed to meet Ola’s expectations, said two people in know of the matter.

“It (grocery) is not a profitable business and has huge delivery costs,” said one of the two people cited above, requesting anonymity.

Your try to raise more money, but funding is impossible to come by. So you start shutting down offices in some cities. It’s temporary, you tell reporters, who don’t understand much about food-tech, but understand what drives pageviews perfectly well.

You ask around to find a buyer.

It’s over.

So you finally go for the only model left.

You set up a stall in front of a popular college. People can come and buy. Or not. You don’t care. You refuse to do delivery, even to a car. You claim to have 99 varieties of dosas, but you actually only have 8. The rest are just revolting variants involving pineapple, paan or Red Bull.

But at least getting paid isn’t a problem.

You realize that success in the food-tech business is a hard, long grind – like most other businesses. Sometimes the economics work out, and when they don’t, it’s best to be prudent and try other opportunities. Great products or services work best when customers value them as highly as your investors do. And are willing to pay for them. And often take time. Lots of time.

 

Kicking The Habit

Growth for the company was steady for the first two and half years, growing at 1.8X a year. And then came the boom — in 2014, the time when companies went giddy raising money. Urban Ladder, too, was one such that went from having a modest $6 million in the bank in November 2013 to a cash pile of $77 million in two rounds by the summer of 2015.

So what does so much cash do to a company?

Simple, you grow like you never did before.

Until 2014, the company was a team of 40 people operating out of a 6,000 sq ft bungalow making revenues of Rs 2.7 crore (as per Registrar of Companies filings) from selling an assortment of about 500 products on their platform in three cities.

But when you see competition raise money and see that the macro environment was upbeat (for funding), you start making bigger bets, says Srivatsa. Dressed in a pair of blue jeans and a black shirt with the company logo emblazoned on it, it’s evident he wears Urban Ladder on his sleeve.

So with two consecutive rounds of 20 million in July 2014 and 50 million in April 2015, Urban Ladder’s ambitions suddenly expanded into a 42,000-sq ft swanky office. Everything assumed larger proportions in those 16 months–it had more than double the people, eight times more products were made available, expanded to 10 more cities and surely to keep them going four times more coffee was consumed.

With this ramp up, the company was making close to Rs 19.2 crore in revenues by FY15, about six times more than what they did in the previous years.

And Urban Ladder was hooked to this.

“When you have more money, you start to think, we grew close to 1.8X earlier, so we thought let’s grow a bit more,” says Srivatsa confessing to the addictive nature of growth.

And that bit became 4X.

By now, the company also wanted to be seen as a big brand. So from 2015, it spent big bucks on running three different TV campaigns for nearly 15 months in prime time slots.

Besides chasing growth, what came with the money was the restlessness to solve problems. Goel and Srivatsa were emboldened with the money in the bank and felt this was the chance to put an end to all the bottleneck woes the business had, could have and would have.

Of slight build but big ambitions, Goel stands up to make a point: “Typically, you choose one or two areas that are bottlenecks today or are likely to be in the future. And you debottleneck them. This you keep doing as a principle. But when the growth you gun for is significantly high, you say let there be no bottlenecks.”

As a result, the company invested on a host of things much ahead of the curve.

Could there be a problem with stock?

“When an order comes I should have no bottleneck. I should have excess stock. Stock cannot be the reason that I cannot service the order,” says Goel. So pop went that block.

Next, was the availability of products.

The number of products grew from 500 in early 2014 to 4,000 by 2015 and they experimented by adding different designs in upholstered beds, dining chairs and bedside tables, for the first time.

One by one, the management targeted the entire supply chain and began busting bottlenecks that they could encounter anywhere from three months to even 10 years in the future.

Now, is that too much for foresight?

“An 18-month horizon is the maximum a young startup should take considering the changing consumer needs and ecosystem changes impacting new businesses,” says Sreedhar Prasad, partner – e-commerce and startups at KPMG.

Slowing the speeding train

By November of 2015, the founders began to see a mix of internal and external signs to have a more pragmatic approach to business.

Goel, a self-confessed nerd, saw it coming first. The signals from overseas began coming in that the ecosystem was headed for a funding crunch. Internally, they saw that their ballooned product catalogue was not pulling in more customers as it should have. It was time for Goel to have a conversation with his team.

 

The lessons Urban Ladder learnt from having more money than necessary

The first conversation I had with Goel about slowing down was easy, says Srivatsa. “But then I would have these bathroom moments and going-to-bed moments when I would rethink…” It took Srivatsa about 15-20 days to internalise the conversation and think about what it would mean for Urban Ladder.

Meanwhile, in November 2015, the government amended the FDI rules to allow foreign-funded retailers to sell products which were under a single label, both online and offline through stores.

As part of revisiting their strategy, Goel and Srivatsa looked at their cost structures and found that even though they were a brand offering curated products through their sellers, they were behaving like a marketplace in terms of offering a vast range. They had expanded their range by more than three times during the ramp up phase.

“We were in this hazy middle zone and we were confused,” he says.

The problem partly again was because of the large sum of capital they had.

“When we had a lot of money, we thought we were a marketplace, and functioning as one. In a marketplace, you offer the lowest price, an infinite range and convenience to customers. But as a brand, what we offer is a curated range, value for money and an experience,” explains Srivatsa.

While Urban Ladder had the cost structures of a marketplace, they were experiencing the growth rate of a brand. So we had to cut the cost to get to behave as a brand,” says Srivatsa.

Mistakes cost

As a response to these triggers, they set out on an expense cutting drive.

Naturally, the first to go was TV. It was not yielding anywhere near the returns they hoped for. “If we were expecting X as revenue (after the TV campaign), what we landed with was X/5 for that investment,” said Srivatsa. Marketing expenses were cut by nearly a third, and it went from accounting for about 60% of sales to 20%.

Then they looked at their catalogue. Data showed that many of their new designs and products they had made bets on didn’t have any takers despite discounting.

“Some things were not even core to our design philosophy. But we went with it. Those were some mistakes we made,” says Srivatsa. So they pared the number of products from 4,000 to 2,400.

Since November last year when it began the reality check exercise its total expenses have been cut by a third, and the burn has been reduced by 75%, says he.

But undoing some of those excesses have not been easy.

For instance, the company had invested in additional warehousing space back when warehousing was not even a problem and they anticipated it to crop up in a few months. But when they brought in more efficiencies into the system, suddenly there was more space.

“We have 40% additional warehousing space. And we have already spent on it and it is very painful to unwind from that. We have a three-year lock in on that and now my high-quality team members are spending time unwinding that position,” says Goel taking sips of warm water out of his flask.

In fact, Goel admits that had they raised $15 to $20 million less, there would have been a lot more discipline.

Life in the slower lane

With the accelerated growth, the company was able to get scale, but it was not a fun ride.

“It is not that much fun chasing crazy growth. You can see things breaking around you. Sometimes you see customer experience is suffering. You can’t be running a train at 100 kmph and checking the fuel tank at the same time. Just cannot do it,” says Goel.

The organisation took a conscious effort to maintain a more sustainable rate of growth. That meant growth fell– from a rate of 400% last calendar year to 60% now. Goel believes this steady growth is their path to profitability and expects to be operationally profitable in six months.

“In a market like furniture retailing, when the possibility of repeat sales to a customer is less, around 50% is a realistic growth target in the initial years unless there is a significant geographic expansion plan,” says Prasad of KPMG.

Investors seem to like this approach too.

“It’s easy to buy rapid growth through marketing and discounts. But usually customers acquired in this manner are not loyal – they are transactional.

The mounting NPAs of India’s stem cell banks

Now, six years later, Stemade claims to have 5,000 units stored but almost everything else is elusive. The company grossed Rs 3.4 crore in revenue for the year ending March 31, 2015, and booked a loss of Rs 3 crore. Most of the other stem cell banks, even though an established business in other parts of the world, aren’t in good shape either. LifeCell International, India’s largest cord blood bank, has seen a steady decline in profitability in the last five years, it recorded a loss of Rs 55 crore on revenues of Rs 141 crore in FY15.

Even if in fits and starts, these banks have been adding to their collection of cord blood units and dental pulp but they are at best non-performing assets. While Ghaisas would like to assign a good part of the mounting losses to the government’s withdrawal of service tax exemption to such banks since February 2013; the reality is a little more complex.

Is being ‘private’ the roadblock?

Storing umbilical cord blood in private banks is like ‘biological insurance’ against illnesses. That’s been a popular catchline for these banks. Parents who bank their children’s cord blood can use it for the family’s personal use. But the usage of these cells is so low that the need for this service is under question.

To set this in context, LifeCell, which its vice chairman Mayur Abhaya claims, has 200,000 units stored, is on its way to becoming one of the largest banks in the world. Abhaya says that it has supplied cord blood units to 45 transplants so far. That translates into 0.0225% use rate. Globally, for private banks, the rate ranges from 0.04% and 0.0005%.

Most of the uses are for blood-related disorders and are often expensive. Just to cite one case of thalassemia, which affects 35 lakh people in India. A child diagnosed with thalassemia can be cured if the bone marrow transplant happens within the first five years of age, beyond which the chances and complications of rejection are very high. Such a transplant would cost anywhere from Rs 12 lakh to 25 lakh. (Christian Medical College in Vellore charges Rs 8 lakh and provides the cheapest transplant.) Few children are diagnosed on time and fewer still can afford this treatment.

And this is why there are camps of medical professionals, which espouse public banks so that a wider pool of patients benefit from the science.

The use rate is distinctly higher, at 2.2%. In India, there is no overt support for public banks but in the US, the American Academy of Paediatrics and the American Society of Blood and Bone Marrow Transplantation have published guidelines encouraging parents to store these cells in public banks if they want to or have to. They do not recommend private banking as an insurance against an unforeseen future use of cord blood.

“In Singapore, Taiwan, South Korea and other Asian countries, what they have done is combine public and private banking and mandated certain expenses where it is reimbursed and where it is not. So it is extremely competitive as the rules are laid how the banks will work,” says Mahendra Rao, eminent scientist and former director of the NIH Centre for Regenerative Medicine in the US.

The changing environment

The moment you call Abhaya on his phone, Aishwarya Rai starts talking. Yes, she does. That’s Abhaya’s caller tune; of Aishwarya Rai Bachchan endorsing cord blood banking with LifeCell. The ads were all over on television when the company raised $4.6 million from Helion Ventures in 2013. An aggressive marketer, Abhaya first got Bollywood star Hrithik Roshan to be his brand ambassador when LifeCell introduced EMI for banking service seven years ago.

“We’ve created awareness in the country. We are in an investment phase,” he says, justifying his accumulating losses. “Moreover, it is an accounting thing, we are deliberately booking losses. We are cash flow positive but net worth negative,” he says. Most banks offer two options: One-time charge for 21 years of storage, in which case they count the processing fee as the revenue and the rest as deferred revenue over the remaining number of years; and annual subscription.

Ghaisas’ Stemade is planning to move towards similar pricing. “Our current pricing model is upfront for processing and future preservation till the child gets to 21 years. We also have an EMI scheme. Globally, people are moving on an annual subscription model for a fixed tenure. Umbilical cord guys have already moved on it. We may give that as an option. The idea is to get into the next economic level of parents and expand market reach,” he says.

 

YepMe Wants To Fly

Today, after five years of churn and burn, Gaur, as YepMe’s CEO, is turning back the clock and conventional wisdom to take YepMe offline. In a way, he’s once again giving up on the e-commerce dream, albeit only partially. He’s got company though, as many niche e-commerce players like Lenskart, Urban Ladder, Pepperfry and Firstcry are doing the same.

From opening its first flagship store in January this year, YepMe is now on a roll to create a network of flagship and franchise stores across the country. But not everybody is convinced offline is the way forward, especially investors. Ironically, neither is everybody convinced pure online retail business can become profitable in the near future, especially those running the businesses and taking the brunt.

So what gives?

It all comes down to numbers. India has about 40-65 million online buyers, depending on who you ask. This is the target market for all e-commerce companies, be it niche ones like YepMe or mass ones like Flipkart and Amazon. This customer base is growing, but at a pace far from what is required to sustain these companies. None of the e-commerce players have been able to become really profitable so far. Revenues are growing, sure. But are they making money?

Gaur’s first coup was getting Bollywood superstar Shahrukh Khan to endorse his relatively unknown brand. Using an informal set of connections tracing back to him and Khan having studied in neighboring schools in Delhi at about the same time, he somehow swung the endorsement. “He was interested in e-commerce business and asked us about what we were trying to do, and what would we do if it didn’t work out,” says Gaur. “I think he could connect with us, because we were not born with a silver spoon, but were trying to make it big.”

In November 2014, YepMe shot its first commercial with Khan. And the results were as expected. Over the next three months, they did twice the business than usual, reaching the run-rate of Rs 35 crore a month. But after signing “King Khan”, YepMe was also out of money. So even as its sales shot up, the question was how long could they sustain it?

The offline hypothesis

“When we thought about it, we were reaching out to 40 to 50 crore people through the television commercial. But we knew, not even one-tenth of that would come online to buy. It was like using a bazooka and killing a fly,” recalls Gaur. “It was that time, Sandeep suggested to venture offline, to maximise the opportunity.”

“But you can’t just shift gears and the investors have to be aligned to the move. Because those were the gaga days of e-commerce, nobody wanted to move to offline,” he adds. “That time we decided offline was the way to go. We wanted to launch in the upcoming summer season, but we had run out of capital by then.”

Going offline isn’t as easy. It comes with its own set of challenges. It is capital intensive, because rentals are high, and you need almost a separate team to run the business. There is enough and more competition already out there. And even if you can take care of all these things, it will be a hard decision to make, because, for online companies, the valuations are at stake.

“Firms that are straddling both online and offline will reach a fork in the road – are they online businesses or are they offline businesses with an online presence. If their economics reflect the latter, then they will be valued as primarily offline businesses with lower earnings and revenue multiples – may be higher than the traditional offline retailers but they would be a far cry from the online valuations we have seen,” says Haresh Chawla, Partner at India Value Fund Advisors. “Also, the metrics they will be evaluated and valued on will shift from hot-air oriented ones to real profits and margins – which itself will dampen valuations.”

Gaur was aware of the challenges. After raising $75 million from Khazanah Nasional in September 2015, he started the offline play as an experiment. He claims that the first flagship store in Gurgaon saw six times more sales as compared to online sales happening in that area.

 

Snapdeal – The Pivot Machine

In a move that mirrors that of its arch-rival and market leader Flipkart, CEO Kunal Bahl is stepping back from operational responsibilities to focus on fundraising and fund conserving, while his co-founder Rohit Bansal takes control of day-to-day activities. These changes are significant, especially with the drop in valuation as background. Snapdeal’s last confirmed valuation was at $4.8 billion last August, as reported by Mint, down from $6.5 billion in January. We also reported last week that VC firm Sequoia was apparently willing to sell its stake in Snapdeal at an even lower valuation of $3 billion.

Reorganising the workforce isn’t new for Snapdeal. It has tried everything in the book. Snapdeal has hired people, fired people; decided Gross Merchandise Value (GMV) was boss and then decided Daily Active Users (DAU) was the key metric. It focused on Exclusively.in to drive revenue and then wanted to acquire Jabong and is now hedging its bets on Freecharge. Snapdeal is doing everything that will return it to its glory days of nipping at Flipkart’s heels.

All of this would suggest, this is a farewell Snapdeal story. It is not. The company is evolving and is trying hard to stay relevant. Kunal Bahl and Rohit Bansal are not pushovers and they know how to put up a good fight. In June ‘16, Snapdeal had, according to sources, about $450 million sitting in the bank, giving it a runway of about 20 months. And buyers are circling. Snapdeal is the third biggest player in the market and you can swipe them away as an also-ran but do it at your own risk.

The evolution of Snapdeal will define how the market plays out. Not only for someone with skin in the game but also for a customer. Snapdeal has been wooing not just investors but buyers from across the world. Whispers of Bahl courting Alibaba to take a gamble and buy it out have been getting louder. This evolution, and our documentation of it, shows Bahl and Bansal’s well-intentioned but failed efforts to find a way out of the e-commerce maze. To piece this story together, The Ken spoke to several current and former Snapdeal executives. Most of them requested not to be named but had a lot to say in private.

Bridge over troubled water

Last year, in August, the top bosses at SoftBank and Snapdeal sat down to talk. The fight with Amazon was lost. They were asked to change track.

Snapdeal was reduced to being the third biggest e-commerce player. Orders peaked at 150,000 and then fell off a cliff, finally settling at 75,000-100,000. In context, currently, Amazon does 150,000 and Flipkart, just above 225,000 orders a day.

Last time Snapdeal knew it was losing a fight, it pivoted and that’s exactly what Bahl and Bansal tried to do. The first brainwave that came to the duo: build a version of WeChat in India. The five-year-old Tencent-backed company is the most popular chat platform in China. The company helps users chat, make e-commerce transactions and has a payment platform. It has 700 million active users, 200 million of them have linked their bank accounts to the instant messaging platform. A chat platform in August was commissioned to be tested on Freecharge and then expanded to Snapdeal.

Next, Bahl and Bansal started hiring. A big technology operation was commissioned. After a frantic round of hiring, Snapdeal stood at a bloated 8,000 people. Happy with the idea, in September, the duo went to the boardroom once again. And they were waiting for the question: What’s the plan?

Rise of the platform

“They said WeChat. And as soon as they said it, it was shot down,” says a former employee, who is part of another e-commerce major right now. He did not want to be named in the story as the current company does not allow him to talk to the media.

“There was no way Snapdeal could overhaul an e-commerce company around a chat platform this late in the game,” the person says.

It was September, and the two founders were back to the drawing board.

Then, another brainwave. Bansal in a brainstorming meeting suggested that users would take to Snapdeal if they were offered the e-commerce platform in their native language. So, the defibrillator was brought out for multilingual Snapdeal. In 2014, the company had launched Hindi and Tamil versions of the website but the idea was shelved.

 

An intelligent conversation around how big is ‘the nation’

What about Rajdeep Sardesai and Barkha Dutt, the two other news colossus of our times whom Arnab has left behind in a trail of dust and noise? They get merely 0.03% of viewers tuning into their show. Clearly, disproportionate to the level of abuse and trolling they get on Twitter, but then when did trolls ever need to watch the news to comment on its content?

Those amongst you, the naysayers, will perhaps say, how on Earth can we ever determine the viewership of a country the size of India? The truth is, the sample size of 22,000 TV meters is robust enough.

Let’s assume first that Arnab is primarily an urban, English-speaking phenomenon and hence he’s likely to be watched in 50+ cities with 1 million plus population.

To ensure that the 77.5 million homes are adequately represented, a little over 2500 TV meters will accurately give you 95% of the population mean viewership numbers. We have well over 2500 meters now. The potential margin of error is quite low (<1%). But let’s assume that is actually over 10%. That gives Arnab a potential audience reach of approximately 200,000.

200,000. Let this number stay with you

That’s how big is the nation – a tiny drop in the TV viewing audience. Truth is, there’s an obvious limit to Arnab’s viewership, but seemingly none to his influence.

Which brings me to another fascinating part of this story. Now, even as I was in the middle of this research, I was told, Arnab is not a TV phenomena. You need to work harder. The amplification which he gets on social media is the real reason why he is so popular. For instance, this person said, Arnab has been trending on Facebook for three days since he quit. “If you notice 50, 000+ people are talking about him,” he said. “He is not even on Facebook but he is trending. This is the audience which is not watching Newshour but knows about him and is talking about him. What you need is social media amplification data.”

Fair point. So, I reached out to Meltwater India, a media intelligence company which provides media monitoring and social media monitoring data. Meltwater looked into it, and remember, Arnab is not on Twitter and Facebook. The insights below are from social media. Primarily Twitter: 94%. Facebook: 2.17%. Youtube: 1.58%. Comments: 1.08%. Forums: 0.79%. Blogs: 0.45%.

Question 1: How popular are Arnab Goswami and Times Now on social media? How often is Arnab trending?

The graphs below are a keyword analysis of the conversations in social media around Arnab Goswami with Newshour (graph 1), mentions of Arnab with Times Now (graph 2) and Arnab Goswami’s overall social chatter (graph3). These are actual mentions.

There’s not much chatter around Arnab and Newshour. Or Arnab and Times Now. So, it doesn’t look like people are actually watching the show. Arnab’s individual mentions are far higher, maybe because people talk about him to fight the good fight as a journalist or also, simple caricature value.

Question 2: How popular is Arnab Goswami on social media, relative to Rajdeep Sardesai or Rahul Kanwal? Or someone like Barkha Dutt?

Picotrial Representation

The graph below depicts the share of journalists’ voices in percentages (graph1) and media exposure depicting the numbers (graph2).

In actual mentions, Barkha Dutt is far ahead of Arnab. Is that a factor of trolling? Quite likely.

Question 3: What are the conversations around Arnab like? I mean, what are people saying? Can we get an idea on the number of people tweeting about Arnab or Newshour between let’s say 9 PM to 10:30 PM?

The data below has the content stream depicting the nature of the conversation, the sentiment, and the tweets by time of the day.

Chatter does seem to go through the roof at 8 PM. But there’s nothing special about the Newshour slot.

Looking at the data and trends, where do we stand with respect to our original question? A very small, rather minuscule, percentage of people constitute ‘the nation’ in the ‘nation wants to know’. English news has very low viewership. And the nation that Arnab speaks about doesn’t go beyond urban India, beyond the posh living rooms, where the posh working professionals live. Social media is a valuable addition to his persona but seems like it has little to do with the Newshour. It amplifies his persona by polarising people — those who agree with his fire-breathing style and use it as part of their “country-first” narrative, and those who find the entire thing meme-worthy.

 

The Interconnect Wars

The launch was belated because Reliance paid over $3 billion for Jio’s core – the 4G spectrum over which it runs – in 2010. Since then newspapers, telecom experts and customers have been prophesying its imminent disruption of Indian telecom.

But the launch was also unexpected because the date Ambani set for Jio’s commercial launch was December 31, a full four months away. In the interim, Jio would offer its services free of cost till then, as a free trial.

Now Reliance is not a company that telescopes its big bang launches in advance. It likes to keep its cards close to its chest till the last minute. For instance, in 2010 no one expected a little known company, Infotel Broadband, to act as a Trojan Horse on behalf of Reliance to become the sole winner of 4G spectrum across all of India for over $1 billion.

It’s possible Ambani’s hand was forced by the powerful incumbent group he was going up against – Airtel, Vodafone and Idea – the three largest telcos in India who collectively account for nearly three quarter of Indian mobile subscribers. Their weapon of choice? Points of Interconnect, or POIs.

POW to POI

At its simplest, a point of interconnect is a place where two different networks are physically connected to each other. For instance, a cable representing traffic to or from Airtel’s network with another for, say, Jio. Without these physical connections, it’s impossible for calls made by one operator’s customer to end up on another operator’s.

POIs are neither very fancy nor expensive, at least at an individual level. Sitting deep within data centers spread over the country (including multiple cities within the same state), they’re boring black boxes with numerous ports behind them, each of which takes in an “E1” cable connection.

What’s an E1? It’s a widely accepted telecom standard that translates to 2 Mbps of capacity, which in turn means 32 simultaneous voice “channels” of 64 Kbps each. At the simplest level, it means one E1 connection is capable of carrying 32 simultaneous voice calls. Because no one ever makes continuous calls through the day, each E1 can handle anywhere from 5,000 to 10,000 subscribers on average.

This is where POIs get interesting.

Telecom regulators around the world, including TRAI in India, pay special attention to POIs because of their criticality to telecom networks and competition.

“According to ITU’ surveys, Interconnection-related issues are ranked by many countries as the single most important problem in the development of a competitive marketplace for telecommunications services” [Source]

Ask Bhavin Turakhia, co-founder of Directi, a group of Internet and tech businesses that claim revenues of over $250 million a year, and a valuation of $1.4 billion. In 2014, Turakhia introduced Ringo, a mobile app that allowed users to make international calls at prices that were significantly cheaper than leading VOIP calling providers.

In January 2015, Ringo was launched in India. Fast forward nearly two years and Ringo is still a non-starter in India and has had to suspend operations. Turakhia says he has spent “several million dollars” setting it up in India, including the money spent acquiring a telecom license to legally offer his services in India.

“Since 2010, we are the only company to acquire a new telecom license in India. We did it even though the capital requirements are high. But the incumbents are clearly flouting laws to delay competition,” he says.

And how are they doing that? You guessed it – POIs

According to a letter dated October 10, 2016, to TRAI, India’s telecom regulator, VMobi, Ringo’s parent company says leading incumbent telcos were asking it to put up a “bank guarantee” of Rs 11 lakh for every E1 POI it wanted. The graph below compares the cost of an E1 port that newcomers are meant to pay incumbents for each of the first two years versus the amount demanded from Ringo as a bank guarantee.

No, we didn’t forget to plot the cost of an E1. It is Rs 4000, or 0.364% of the bank guarantee amount. So small that a normal graph won’t suffice, but only a logarithmic one.

There’s neither a precedent, regulation or financial rationale behind incumbents asking Ringo to pony up 275 times the yearly cost of an E1 (not counting the Rs 10,000 one-time set up fee that must be paid separately) as a bank guarantee. Except, of course, to stymie it.

 

Subterfuge, opacity and regulatory warfare

Bank guarantees weren’t all. Ringo’s letter to TRAI mentions various other legal roadblocks incumbent telcos put in front of it to restrict the services it could offer, none of which they imposed on each other.

Who blinks first, Telecom Edition

In spite of claiming to have spent Rs 1,50,000 crore to build India’s first “all IP” network (essentially a computer network transmitting packets of data instead of the voice networks that require dedicated channels to carry voice), Jio’s Achilles Heel was the fact that its customers still need to make voice calls. And those calls need to land on the incumbent’s voice-based networks.

Which require POIs.

As per law, 0.5% are the number of call drops that are tolerable on POIs. But as Jio started adding more and more users to its network in what it called a “test” phase, its call failures started climbing to anywhere between 50-80% for calls made to the three largest incumbent operators.

And thus began an elaborate, high-stakes game of who-blinks-first between Jio and the incumbents, with regulations being the weapon of choice.

When Jio started complaining earlier this year that it didn’t have enough POIs for its users to make calls to incumbent networks, the incumbents said Jio was only conducting tests. So why did it need so many POIs? They said they weren’t obliged to provide POIs en-masse to any non-commercial network.

So Reliance announced a “commercial launch” on September 1, even though it wouldn’t charge any fees till December 31. Naturally, its subscriber base started to swell dramatically. Today it claims to have over 35 million subscribers, with 600,000 to 1 million new ones being added each day, though there’s no reliable way to verify this.

POIs again became the bottleneck. This time the incumbents said there was a “tsunami” of asymmetric traffic from Jio that was threatening to break down their own networks. Instead of ratios like 49:51 or 60:40 they were used to all these years, the ratio of incoming versus outgoing calls from Jio was 90:10 they said.

“Jio’s entry and scale is unprecedented,” said a representative for one of the three large incumbents. He did not want to be named to avoid prejudicing various regulatory and legal cases they are part of. “If you give anything for free, there will be a mad rush. Operators earlier undertook phased rollouts, so POIs could be added in a phased manner.”

To which Jio pointed out that it was willing to pay for as many POIs as required (Ringo says the same), but the operators then picked on the 14 paisa that Jio is meant to pay them for each call that “terminates” on their network. The figure was based on flawed assumptions, they said. The true figure was closer to 32 paisa, which meant they were losing money on every call that Jio terminated on their networks.

There is of course no evidence of this being true, except for the operator’s opaque reasoning. Globally, the trend has been to either drastically reduce termination charges to very low levels, or even do away with them altogether. Some countries like the US have even imposed it on the receiving operator, which can then collect it from its own subscriber via billing.

Then there’s voice traffic itself

It’s important to understand two things about voice traffic – one that it is broadly inelastic, meaning that if you drop tariffs then usage doesn’t keep shooting up. Thus, even zero tariff for voice calls on Jio doesn’t imply that most users will speak 2 or 3 times more than they usually do. And two, the quantum of voice traffic has plateaued in the Indian mobile ecosystem. Jio claims that its overall voice minutes are only increasing by 2-3% per annum per user.

This means at some point, extra voice traffic on the Jio-Airtel connection could merely be shifts from, say, the Airtel-Vodafone or Airtel-Idea connections. Which means existing POIs and E1s are be reconfigured, instead of being bought anew.

Meanwhile, call failures continued to climb for Jio. The graphs below represent outgoing call failures from Reliance Jio to the networks of Airtel, Vodafone and Idea during an average 24-hour period and busy hour respectively (the data is Jio’s) during the last month.

Frustrated at the stonewalling from incumbents, on October 21 TRAI recommended a fine of Rs 3,050 crore collectively on Airtel, Vodafone and Idea.