Connecting the unconnected: The device challenge

Authors note

This is part of a series which analyzes the challenges in bringing reliable and affordable broadband to the undeserved parts of the world. The first provided the overview, while the second looked specifically at the infrastructure aspect. We now round up the series by analyzing the “ultimate” mile — that of the device.

For webscale players such as Google and Facebook, the end-user, in more ways than one is the product. Greater the number of users, higher their level of engagement (read — amount of time spent on their assets) all contributes into a complex set of algorithms which allow advertisers (the real customer) to hyper-target their audiences. On hindsight, it wasn’t a surprise that Facebook’s stock suffered on July 26 with investors spooked on slower user growth during an earnings call. What did definitely shock me was the drop; at USD 119 Bn the biggest one-day loss in US stock market history. Did investors pay that much attention to user growth over revenues in determining the long term future of a company? If so, what — if anything can folks such as Facebook or Google do? One of the biggest hurdles they face is that the developed markets are saturated and the “great wall” of China has prevented access to one of the most attractive (but fiercely competitive) markets. Growth in these markets is primarily an upsell game e.g. Facebook engaging users across Instagram, Messenger and the likes, and Google doing the same across Android, Chrome, Maps etc. The only way out is the developing world which still has significant growth potential — but comes with its own set of challenges — from lower spending capacity, lack of infrastructure and unconnected and underserved population.

The efforts to connect the underserved

It is not that they have been standing still. On the contrary both firms have been active in this space. There are ongoing efforts ranging from submarine cables, to access and backhaul networks including balloons, drones and WiFi. Another set of initiatives have been focused on developing OS, apps and services which can adapt to the often bandwidth (and cost) constrained environment the end user is confronted with. These efforts are bearing fruit as indicated with the latest Inclusive Internet Index; the percentage underserved is steadily shrinking — and will continue to do so in the next few years as incomes grow and new technologies and players emerge across the landscape.

To some extent these efforts are also an endless catch-up game. While infrastructure improves, the underlying apps and services also advance, often at a far more rapid pace than the infrastructure replacement/ upgrade cycle. And in countries which are mobile first (in some cases mobile only), when competition or other factors drives prices down usage surges — often straining already overloaded networks. An example of this is in India where the current usage is estimated at 1 GB/day — up from 4GB a month just 20 months ago!

The device conundrum

Within this revolution of sorts, there are a couple of challenges which have still to be fully cracked. I call this the device conundrum and it consists of two parts

  • How can we make the end-user device affordable to the consumer
  • And how can this affordable device actually reach this consumer

The cost challenge

While terminal devices have been dropping, they seem to have settled around the $35–45 mark. This remarkable drop has been aided by the availability of reference designs allowing low cost OEM’s to churn these capable (albeit undifferentiated) devices out at scale. As app/ service demands continues to increase — so does the component costs. A 2015 tear-down indicated a component cost hovering around the $42 mark for a capable device, cheap in USA, but terribly expensive in the DRC where the annual GDP hovers around $468. And in a place where creditworthiness is unknown, and financial inclusion is low — loans or installments are an alien subject.

What ends up happening are that devices circulate in the market for many years, some estimates between 6–7 years (in comparison replacement cycles in USA is around 2.47 years). These old devices may be smartphones — but are ill-equipped to handle newer technologies (such as 4G) or support new data hungry apps.

Another interesting observation (one that is most obvious in India) is that mobile phones are not only essential devices, they are also aspirational devices. Hence amidst an undifferentiated market, there is a tendency to save up for “better” devices. This is reflected in the upward trend in device pricing — with an uptick in spend in marketing budgets. These incentives drive the price in the opposite direction of what the webscale players would like; at the same time the devices present in the market are better suited to handle the requirements of the modern app.

The distribution challenge

This issue is endemic especially in areas where population clusters are spread out rather than concentrated — such as in Uganda and DRC. While the devices may reach the main population centers, they are still a long ways off in reaching the tier 2 and tier 3 townships. In areas where infrastructure is poor, and distribution networks limited and unreliable, it results in increasing the cost and complexity to reach the end distribution point. In a nutshell, it doesn’t make sense if the person can afford the phone, if he cannot find a place to buy it.

Where do we go from here?

Cracking the affordability puzzle

With the challenges as illustrated above, this is definitely one which needs to be addressed. Similar to initiatives such as OCP and TIP — having a reference design created and offered in a license free manner would reduce costs associated with often expensive Intellectual Property. Incremental benefits can be obtained by simplifying devices — offering only a 2G-4G chipset rather than including 3G capability as well. There are other indirect options such as lobbying governments to reduce/ eliminate levies for goods or forcing local production (especially when limited volumes, lack of skills etc. do not lend themselves to economies of scale).

Webscale players can take inspiration from startups such as Juvo, who are actively engaged in building what can only be called — the “credit score for the unbanked”. Perhaps it isn’t an accident that Samsung is among their investors. If one could accurately ascertain the probability of regular payments — then OEMs would be willing to work with operators and other distributors to get their devices in the hand of the end-consumers.

An alternative is working with operators and their payment platforms (e.g. Safaricom’s mPesa) since in many markets, telecom operators have the best available data on location, spend (as a proxy from mobile top-ups) and behavior. Other upcoming alternatives leverage off data rich platforms such as Uber. With this information, there may be a case for simply discounting the cost of the firm via a cash infusion, with the subsidy recovered over time as users spend more in broadband usage. Here the risk could be modelled and an insurance hedge built in

Solving the distribution issue

The distribution issue is more tricky — since it requires the physical movement of goods in regions with poor or limited infrastructure, and do so in a safe and secure manner. Here too, a solution can be found looking at what startups are working on in developing countries — take Copia in Kenya or Ruma in Indonesia as examples.

Customers liaise with a local agent to order from a catalogue of products — including smartphones. Customers may be able to take advantage of an individual or group-based savings plan, depending on the model. Products are delivered to the agent in a short time frame through an established distribution system. Models may be connected to savings groups or offer savings plans. They can also use mobile money as a payment mechanism. While there is certainly a time-lag between need and actual acquisition of the device — it eliminates simply stocking inventory in low density regions. As more data of the user base in a location is obtained, one could consider applying AI techniques to determine the number and type of devices that may be desired — and simply pre-ordering them to the location.

Eventually as infrastructure improves along with GDP growth — this problem will slowly ease, until it ceases to be a challenge.

Rounding up

While there have been failures such as the Android One, the device and associated distribution challenge are just too big to ignore. It is imperative that the web-scale players tackle this head on in parallel with their efforts in the infrastructure space. A success here would ease device adoption, enhance user engagement and serve to further accelerate the user growth story that investors love to here.

About the authors:

Joseph Noronha is a Director at Detecon leading their practice in Emerging Technologies and Infrastructure. He has extensive “on the ground” experience with infrastructure players around the world spanning the Americas, Europe, Middle East, Africa and Asia. His interests lie in Next Generation Connectivity, IoT, XaaS and more recently in Edge computing.

Friederike Göbel is a Business Analyst at Detecon She is working on strategic and digital transformation projects in the Telco and travel industry. In addition, she co-authored in a publication at Springer Verlag on the successful positioning of Telcos in the digital age. With a strong marketing background, she is a digital technology enthusiast looking for customer experience enhancing solutions.