Monday, January 6, 2014

Access is Solved, so where are the users? Part 2


As I discussed in my previous blog post, Access, in other words, letting people that are interested in connecting to the Internet, have the telecommunication infrastructure that allows them to connect, is solved.

Kabul, Afghanistan, late 2011

In the past 5 or so years, the telecommunication industry has expanded so dramatically in developing countries, that today over 90% of the world’s population lives in a place that has Internet connectivity, at least via Edge (2.5G) and in most urban areas in developing countries, via 3G or better.

This is a dramatic change that happened in the recent few years. There is still a lot of room for improvement in the speed and cost of this connectivity, but the most important thing - the connectivity itself, is there.


In many developing countries, there is an astonishing gap between the number of actual Internet users today, and the number of people that could be Internet users - people who have the financial means and the mobile Internet coverage at their home - and yet are not. Why not?

To understand why, let’s do a deep dive using Cambodia as an example, as in many ways it is a typical developing country.

Please bare with me as I walk you through some background and numbers - I promise this will only last 3 paragraphs and then I will get back to the narrative:

First a quick intro to Cambodia, in case you are not familiar with it: Cambodia is poor country of 15M people, with a GDP/capita of $950, 80% of the population is rural, the population is very young with half of the population under 24, and the majority of the people live on less than $2/day. It also has a small but rapidly growing university-educated middle class, that uses the Internet, can survive in English and either owns or aspires to own a smartphone. Sounds familiar? This is almost a generic description of a developing country. Cambodia, similar to some but not all developing countries, also has its own national language, Khmer, and the vast majority of population can only speak, read and write in Khmer.

Access is solved. Personally, I am paying $5/month for 1GB on my mobile phone in Cambodia. This is more than enough even for a heavy user like me, including using it for tethering with my laptop I am on the road. There are much cheaper packages too, of course. In addition, the city is covered with a seemingly infinite number of free Wi Fi access points, and many people who are cost-sensitive can be seen hanging out on the pavements near coffee shops or restaurants to use the free Wi Fi with their phones.

Looking at this nice infographic on Geeks in Cambodia, the number of Internet users is somewhere between 1.1M (the number of active Facebook users) and 2.7M (the number of mobile data users - which would also include people who have once downloaded one ringtone - not my definition of an Internet User) - let’s pick the middle of that range - 1.9M Internet users or about 13% of the population, including 840K smartphone owners (~6% of the population). However, there are 5M unique individuals that own phones in Cambodia (out of a population of 15M - the rest are either children, too poor, or share a phone with family members). In other words there are 3.1M people (20% of the population) that are phone owners (probably not smartphones though) and yet are NOT using the Internet at all - these are people that have already spent money to buy a phone, and spend money every month to use it, and yet are not using the Internet.

It is inspiring and exciting to look at the poorest people and dream about getting them to use the Internet. However, it is more realistic to look at the next 10% of the population that are most likely the next Internet users - beyond the current 13% of population that are already Internet users. Who are these people? And why aren’t they already using the Internet today? How dare they be so unappreciative of all the effort we, the ICT industry, put into solving the access problem for them?

Well, I don’t need to look too far, one them is my mother-in-law, Kimrong:

Mrs. Kimrong Hor, my mother-in-law, at her shop

Kimrong is an intelligent, capable, middle-aged, middle class business woman in Phnom Penh. She owns a nice house, a business, a car and of course a phone. In fact, she owns a smartphone, a Samsung Galaxy Mini, which she uses only for making and receiving phone calls. She is the classic person to be in the next 10% - she can afford it, she is swimming in 3G radiowaves, and she is a very social person who loves to communicate. Yet, the Internet is not part of her life in any way shape or form.


Two big reasons:

Big problem #1: Language. 

Kimrong, like most Cambodians, is monolingual. She can only speak, read and write Khmer. Unfortunately, the Internet does not speak, read or write Khmer. At all. Very few products or services from any company are localized to Khmer. How come?

The problem originates far away from the familiar tropical heat of Cambodia, in a remote land along an ancient trail called Highway 101, an exotic territory known as Silicon Valley. There lays the beating heart of the Technology Industry.

The technology industry is organizationally structured around products. It consists of small companies that have just one product, and large companies that have many products - and these large companies in turn are divided internally into departments, each which is responsible for one product - and can be seen as a small company within a large company. At the end of the day, for each product out there, you have a product team lead by a product manager that makes the call regarding which countries and languages should be supported.

Imagine this hypothetical product creation cycle: after many months of long days, and all too often also late nights and weekends, the product team is finally ready to launch v1 of its new awesome product. Originally, they hoped to launch in many languages, but like many features that got cut out during the development process in order to try to finish the product on time and launch, “internationalization” is postponed to version 2, and the product launches in English only.

Then comes the work on version 2, and the product manager starts thinking about which languages should be included. These would normally be those of the most lucrative markets - markets that have lots of users and money. So version 2 launches 9 months after v1, and it now supports a bunch of European languages.

In version 3, the team adds Japanese, Chinese and Korean (postponed from version 2 to version 3, when the team discovered that they are technically difficult languages), and maybe some languages of smaller markets like Greek and Finnish. Now the product is launched in 20 or so countries, and the team announces that it is “fully global”. In other words, of the 192 member countries of the UN, 20 are supported, and 172 are not.

Just before version 4, the product manager is promoted and moves to another role, the engineering lead leaves to pursue his Ph.D, the designer quits the industry and goes to meditate in an Ashram in India and the finance manager announces that he needs to see solid justification before approving budget for adding more languages. Khmer will not be supported.

When this scenario is repeated (with variations on the theme) product after product after product across the entire industry, the experience of a user in Cambodia is that nothing works, everything is broken. And my mother-in-law Kimrong is excluded from the information revolution, from the Internet, by a series of decisions made by product managers in Silicon Valley.

I repeat: she is excluded not because of poverty, not because of lack telecommunication access, but because of explicit decisions to exclude her made in Silicon Valley.

This is a fundamental problem that needs to be thought about: each individual decision by these product managers makes perfect sense in its own context, but in aggregate, the result is that Silicon Valley is creating and growing the digital divide between developed and developing countries. No amount of good intentions and good CSR can offset the simple fact that the technology products on which Internet users depend to be able to use the Internet just don’t work at all for users in these countries.

This is not only a problem of the platforms owned by large companies - like the popular mobile operating systems or the major social networking services. Even if those were fully in Khmer, my mother-in-law could still hardly use the Internet, because what makes the Internet great are all the specific apps and web services that comes from small companies - the thousands of apps on Google Play and App Store and Facebook, not to mention the millions of websites out there. Clearly, these small companies or individual developers don’t have the financial means or the linguistic knowledge to localize their product to all relevant languages, including Khmer, with today’s technologies and processes.

We, as an industry, need to get way more sophisticated in the way we approach and execute localization, so that platforms as well as small specific apps all get rapidly localized to the relevant world’s languages - and at a cost that a small startup can accept without thinking twice. If anyone is interested in innovating to get more people online in developing countries - please think less about telecommunications, a problem that has already been solved to a reasonable extent, and more about how to solve THIS problem.

Big problem #2: Local content.

If the localization problem was not hard enough, this problem is even harder, mostly because all the easily conceivable solutions are on an individual country basis, and it is hard to find a solution that scales globally across the 150 or so developing countries:

Internet users in all countries interact primarily with local content. All politics is local, all retail banking is local, all neighborhood gossip is local and all eating-out is local. Some music, TV and movies are international, but most are local. Sometimes this local content is hosted on global platforms like Youtube or Facebook or Wikipedia, and sometimes on local websites or mobile apps for news, entertainment, business listings etc.

For local content to emerge in a country, you essentially need three ingredients:

(1) Skilled and talented people to create it: These are the software engineers, designers and artists.

Every country has its few super smart and talented people. There is a lot that can be done to improve their skills, given that the formal education system is usually quite rudimentary compared to the world’s top universities - however from my own discussions with thousands of these creators in dozens of countries, I have come to the conclusion that the main impediment blocking them from achieving world-class abilities, is the lack of local competition - they are recognized as the local superstars early in their carrier, and have little competition at their level, nothing to challenge them and push them forward. This is the classic “small-town syndrome”.

Startup Weekend in Phnom Penh. Hoping for more local content.
Organized by +Channe Suy Lan and +Chantra Be, I was one of the judges

(2) Investors to fund it: Someone needs to put down enough capital for the creators to work for 6 to 12 months on a product, without income, and without knowing if it is going to succeed.

The almost non-existing tech-friendly financial ecosystem of angel investors and venture capital in developing countries, is very significant. Most startups I have seen are forced to do contract work in parallel, such as developing applications or websites for clients, and work on their startup idea on the side. There is just no possibility of surviving for months without income. And we all know that it is very difficult to succeed with a startup if you don’t give it your whole. The need for immediate income also pushes nascent startups to pursue quick-revenue business models for their startup so that they can get money flowing in as quickly as possible - sometimes resulting in rather trivial products that are sub-optimal even from a revenue perspective in the long run.

(3) A business model for ongoing funding: The idea of an equity-based business model (i.e. build the user base and hope for an acquisition) is (almost) not available in developing countries, at least not for companies that focus on products for the local market. Advertisement-based business models are challenging, because the online advertisement market in these countries is still in its infancy, and commerce based models are also challenging, because of lack of remote forms of payment, such as credit cards (some countries, like Kenya, have popular mobile payment systems, but this is quite rare), as well as issues of trust. The startups that are financially successful, usually have some element of real-world operations in their business model, feet on the street, to collect cash for some service. These are viable models, but of course the downside is that they are very difficult to scale nationally, and extremely challenging to scale beyond national borders.

Similar to the language problem, if Cambodians had the creators, funding and business models to develop cool websites and apps that my mother-in-law would find fun, addictive, and cannot-live-without, she might just want to use them. However, these do not exist today, at least not on a sufficient scale.

These twin problems of language and local content are the real problems that need to be solved if my mother-in-law Kimrong is to ever be interested and value using the Internet.

Can you help?


You can also check out my other posts:

Developing Technologies for Developing Countries - Introduction to my blog

Technology products for rural people: Why do they almost always fail

Languages: How to decide which ones and how many?

Access is solved, so where are the users? Part 1

Access is solved, so where are the users? Part 2

Copyright (C) 2013. Divon Lan. All rights reserved.