5 Apr 2015

LTE as Disruptive Technology

Clayton Christensen’s theory of disruption is a widely known and respected framework for how lower-end products displace higher-end ones from the market. The term “disruption” is sometimes used loosely nowadays to mean one product or company succeeding at the cost of another (“Apple disrupted Nokia’s business”), but in Christensen’s work, which is how I use the term in this post, it means something specific: if you make a better phone or camera or something else, that’s not disruptive. A disruptive product is one that, when introduced, is criticised as being unambiguously worse than the incumbent, like smartphones were in comparison with standalone cameras. But they have other advantages (always with you, always connected, apps, easier to use, etc), so over time, they attract a market, sometimes bigger than the original one, as with smartphones vs standalone cameras. It’s important to note that smartphones still take worse quality photos than SLRs or mirrorless cameras. So, disruptive doesn’t mean “better”. A disruptive product is one that competes on entirely different criteria [1].

With that background, how might LTE compare with wired Internet access? Your first reaction may be that LTE is no substitute for a wired Internet connection, but that’s precisely the point: LTE is a disruptive technology. It has lower data caps than wired Internet, but might it have other advantages that would appeal to a different market (not people who use 100GB of data every month)? If so, phone companies can start from there and slowly broaden their appeal to adjacent markets.

To begin with, LTE has the great advantage that everyone has a mobile phone, and needs or wants an LTE connection [2] [3]. There’s a readymade market for LTE, with nothing on the horizon to supplant it. This provides a solid base to expand from.

The first thing the networks can do to encourage greater adoption is to eliminate bill shock. There are multiple ways of doing that. The obvious option is to make all plans unlimited, with a data cap after which you’re throttled, say to 100kbps. With an option to pay more if you need more data. And when you pay more, that resets your billing cycle, so you won’t be charged again for a month from that date.

A variant of an unlimited plan is a “bill shield”, such as 2x your normal bill — no matter how much data you use, you won’t be charged more than that. If we go this route, the bill shield should be no more than 2x your normal bill, to avoid bill shock.

A third option is to charge at the same rate when you exceed your data cap. If you use twice as much as data as your cap, you should be charged twice as much, not thrice or ten times as much. And in no event should you be charged more than what you’d pay if you were in the next tier.

For example, Airtel charges me ₹250 for 1GB of data. If I use 2GB, I should be charged no more than ₹500. But if Airtel offers a 2GB plan for ₹400, then even if I’m in the 1GB plan but use 2GB of data, I should be charged only ₹400.

Conversely, if I’m on the 2GB plan but end up using only 1GB of data in a given month, I should be charged only the lower fee.

Unused data should also be rolled over for 12 months. So, if I’m on the 2GB plan, but use only 0.9GB of data in a particular month, and there’s a 1GB plan on offer, then I should be charged as per the 1GB plan, and the remaining 0.1GB of data should be rolled over for a year.

Don’t look at data caps and complex pricing plans as a way to rip users off. Simple, flexible, fair pricing will encourage users to use more cellular data more without worrying about being ripped off by exorbitant charges. Users will also sign up for a higher tier plan knowing that they won’t be charged for what they don’t use. In fact, they won’t have to worry about what plan they’re on, and can consume as much data as they want knowing that they’ll be charged fairly.

In addition to simplifying pricing, phone companies should look at reducing what it costs them to deliver a bit to a user.

All towers that have been upgraded to LTE should be given a fiber backhaul, so that the capacity of LTE can actually be used. LTE is efficient enough in its use of spectrum that it moves the bottleneck to the backhaul, so we need fiber backhaul at each tower that’s been upgraded to LTE.

All towers should also be equipped with Wifi modems, to serve nearby customers. Given that cellular towers are often spaced close to each other, Wifi will cover a significant percentage of their customers. Don’t charge any more money for LTE customers to use Wifi, since it reduces the carriers’ costs by reducing contention for scarce spectrum. In fact, customers should be given an incentive to use Wifi offload, such as by counting only half the data consumed against their data cap.

The more aggressive version of this idea is to go after wired ISPs. If an LTE customer is using the phone company’s Wifi access point, it should cost them no more than getting wired Internet service at their home. This makes sense because these users are not using precious LTE spectrum. In fact, if you live near a tower, this is more convenient for you than wired Internet service, because there’s no reason to run a wire to your house, have an installer visit, rent or buy modems and routers and set them up and troubleshoot them, and so on. And there’s one less bill to pay every month.

Wifi service should also be opened up to people who live near a tower but aren’t customers. In other words, if Airtel sets up a Wifi access point, it should be available to everyone in the area, not just people who use Airtel for their phone service.

And, just like wired Internet, make it work on all their devices, via a username and password. Don’t play cheap tricks like charging for each device.

The cost of setting up Wifi access points [4] is trivial compared to the tens of lacs of rupees it takes to set up a tower. So, it makes sense to derive as much value as you can out of each tower.

Phone companies should also lobby governments to free up more spectrum, tons of it, either for a fee or for free.

Phone companies should also strike roaming agreements with all other providers to have their customers’ phones use other companies’ towers at no more cost than using their own tower. That way, phone companies can co-operate to reduce costs for everyone, by avoiding redundant towers in an area. Note that when I say “tower”, I’m not talking just about the towers, but also about everything else that comes with it, like the transceivers, whatever equipment is on or below the tower, fiber backhaul, diesel generators, and so on. Sharing all these reduces costs as compared to each company building redundant towers to cover a city. Sharing also leads to better coverage and speeds for users, when my phone can connect to whichever tower has the best signal, or has LTE rather than 3G. Driving users to LTE further reduces costs for the phone companies.

Data can also be made free and un-throttled during off-hours, like 11PM to 6AM, rather than letting it go waste. The cost of setting up a network depends on its peak capacity. You pay for peak capacity, which goes unused and therefore wasted a significant portion of the time. So you might as well make it available for your users to use, free and un-throttled, even for users who have exceeded their data caps.

In fact, the industry should work together to evolve a standard protocol to let the network tell devices when data is unmetered, at which time things like OS and app updates, Bittorrent downloads, bulk Google Drive syncs, and so on can happen [5]. Traffic at times of extra capacity reduces the cost per bit of the network.

With these ideas, LTE can grow to provide a more important channel for internet access, rather than just something to use when you’re outdoors. Carriers and users all benefit.


[1] If you’d like to read more about Clayton’s theory, in 10 minutes instead of the hours the book takes, I suggest you read my summary.

[2] There are exceptions, like people who use mobile phones only for calls and SMSs, but these people are not likely to have a wired Internet connection, either, so we can ignore them for this analysis

[3] Most people use 3G or EDGE to access data, not LTE, but it’s just a matter of time before they upgrade to LTE, with LTE phones becoming as cheap as ₹7000. In any case, as long as you’re paying for cellular data, it provides an opportunity for the phone company to upgrade you to LTE, which is much easier than convincing someone to sign up for a service they currently don’t use, like wired Internet.

[4] These should be dual-band, with directional antennas, to serve more users without competing for bandwidth and slowing everyone down.

[5] This applies to wired ISPs as well, since they have data caps, too.

No comments:

Post a Comment