Like the Surf user looking for stains, you will be hard pressed to get decent internet speeds
Remember the Surf Lalithaji ad? For those who don’t, it had a middle-class housewife, Lalithaji, daring her neighbours to find any stains after she had washed her household’s dirty clothes with Surf. Much like that, telephone majors like Airtel and Vodafone may soon have to be placing similar ads for their data services. Except, they won’t be as flattering since they will be saying “dhoondte reh jaoge” about the internet services they offer.
Much of the problem, as an FE series last week called Spectrum First (goo.gl/AGTrCI, goo.gl/uhZeK6, goo.gl/CyfnBG) pointed out, relates to the fact that Indian telcos have far less spectrum than other global majors have. While the average is around a fourth if you look at all bands, things are a lot worse for ‘data’ spectrum bands, which is what this article is about. As compared to around 60 MHz of 3G spectrum per operator in Japan (see graphic), the Indian average is around 5 MHz. Combine this with the dramatically higher density of population, and Indian telcos simply cannot provide anywhere near the same level of service as their global counterparts do.
There are two aspects to this. One is the absolute quantity of spectrum and the data the spectrum can service. So, if a Bharti Airtel has 10 MHz of spectrum versus 100 MHz for an AT&T, AT&T can push through 10 times the amount of data that Bharti Airtel can. The second is the economies of scale that kick in.
A few years ago, the Plum Consulting report pointed out that the amount of data that could be carried per MHz of spectrum also depended upon the size of the spectrum chunk. So, if there was a firm with 100MHz, it got 50% more spectrum per MHz as compared to a firm that had just 10 MHz—in other words, AT&T would have the capacity to push through 15 times as much data as Airtel, not just the 10 times that the difference in nominal spectrum holdings would suggest.
Poor spectrum quality also means that, to provide very basic coverage, telcos in India have to put in more cellular towers than their overseas counterparts—combined with the high cost of spectrum in India, that’s also why EBITDA margins in India are so low (25%) in comparison with those in the US (50%). Indeed, as the last of the Spectrum First series brought out, telcos in India may not have the EBITDA levels to be able to buy all the spectrum the government puts out in FY16—this is also the reason why, in previous few auctions, some telcos have not participated. This, of course, is not unique to India; in Europe, after bidding insanely high values in the 2000 auctions, telcos never bid similar amounts again. In other words, while the government’s spectrum auctions in February will fetch good amounts, they will bankrupt the industry so much that later rounds of auctions won’t fetch high amounts.
To get back to the spectrum crunch, unlike their global counterparts, data taken for the top few operators in Delhi, for instance, show that while the number of cellular towers has increased by 75% over the past two years, the amount of voice traffic per day has nearly doubled and the data traffic by 3.5 times. In comparison with China, for instance, Delhi’s top 3 telcos have around 19 times the amount of data loading per MHz of 3G spectrum and around 11 times the amount of voice traffic.
As a result, two things have happened. For one, there is a sharp fall in the speed of the data networks. Two, the battery of a phone drains out much faster than it did in the past. So, if your battery lasts just 8 hours in Delhi, it will last 16-18 hours in the US; and that is without the calls dropping half the time—much of what we think is poor battery life is actually the phone expending its battery while searching for a signal from the nearest telephone tower.
From the point of user experience, Ericsson has some useful comparisons across the world (see graphic). For India, as the graphic shows, users of 3G have just a 55% chance of getting more than 1Mbps of data speeds (which is the minimum required for decent video quality) as compared to a 90% probability of getting such speeds in most developed countries. In Melbourne, for instance, users get more than 1Mbps of speed on their phones/tablets/computers 100% of the time; in Singapore, the comparable figure is 90% while it is around 95% in Beijing. (If users in Australia can get better quality signals versus those in Japan despite the latter’s telcos having more spectrum, this is related to the relatively lower density of population in Australia.)
So, at a time when data usage is the new revenue stream for telcos, and users are not able to get the kind of quality they expect, chances are telcos are not going to find it easy to charge significantly higher tariffs than they do at the moment. That, in turn, puts a limit to the amount of fresh investments they can make in setting up the infrastructure required to service Digital India’s growing needs.