Lessons From Loon & Connecting The Unconnected

News that Loon was winding down operations because “we haven’t found a way to get the costs low enough to build a long-term, sustainable business” came as a disappointment to many, and rightly so. It is hugely difficult to provide connectivity in very remote areas, and any new technological innovations that could do so would be welcomed particularly in Kenya and Mozambique where operations were quite advanced. What may we be able to learn from this announcement? There are lessons around costs, and more importantly, around revenues.

The first lesson is around costs. A typical base station can cost anywhere from $20,000 to $80,000 depending on factors like how it gets backhauled to the Internet (e.g., via satellite, microwave, fibre or LTE relay), the size of the actual tower itself (the taller a tower the easier it is to provide wider coverage), power costs (e.g., solar-powered with batteries if grid electricity is not available and can lower long-term operating costs), and the equipment on the tower.

The balloons that Loon used also incurred costs, including the balloon itself, 4G equipment, and the use of lasers for backhaul. These lasers transmitted the Internet from the ground to a single balloon and then between different balloons – innovative technology which is still being used in other projects outside of Loon. For Loon, one challenge was that each balloon only lasts a few months and thus needs to be recovered and then relaunched, adding a great deal to the cost even if the technology on the balloon can be re-used. Consider, for example, balloons that are currently launched only from Puerto Rico and Nevada and then spend weeks travelling across to Africa.

In the industry, we talk a lot about Capital Expenditure (CAPEX) and Operating Expenditure (OPEX). For a base station, CAPEX is buying the equipment and building the base station, which is a one-time cost and may last a decade or more. The tower itself can last multiple decades, even if equipment on it may need to be replaced. OPEX for a rural site may be quite low if the tower has no power costs and uses solar panels with batteries and the site rental is low.

In this case, the main operating cost is backhaul costs, which partly varies by demand. A telecom company thus faces significant up-front costs and then hopes to reclaim the revenue over years, ideally around 2 to 3, sometimes around seven. But some sites may never recover the costs, and thus those sites rarely get built without subsidies.

Loon seems like it has quite significant CAPEX and OPEX if the balloons have to continually be replaced to refill the helium, and that may be one of the reasons the costs are not as low as desired. Maybe with time, the balloons could be improved to last longer, or maybe with greater scale, more manufacturing, launch (and helium refill) sites could be built closer to where they are needed, which would reduce costs. As the balloons move around, there is a large minimum number of balloons necessary to ensure good coverage of any particular location which is a large cost, even with Loon’s impressive algorithms managing the balloon’s movements.

It seems that Loon has decided it might not be able to get the costs low enough, or is not willing to keep investing any longer to reach the point at which it can get them low enough.

The second lesson is around revenue. As well as costs, perhaps the bigger challenge is revenue. Enabling Internet use is the key, not just providing Internet coverage. But in rural areas revenues are typically very low due to few users, the low income of those users, and thus the low average revenue per user (ARPU) for service providers. Loon was supposed to help overcome this by covering a very wide area, which could have many more users per balloon than a typical base station (200 times more, according to Loon). However, providing wider network coverage does not mean there will be more users, or that they will spend money on telecommunications.

In a remote area that previously did not have coverage, very few people will have a phone or smartphone. Why would they, if there was previously no coverage? They need to be able to afford one, as well as the electricity to recharge the phone on a daily basis and the Internet service fee for voice, the Internet, mobile money, or other value-added services such as agricultural information services or educational content that might not be free.

Few actual services may be useful for that person—in other words, there may not be a website or app that meets that person’s day-to-day needs in a language they understand, because they are likely to represent a minority market. As a result, they may not spend much money once they are online. On top of that, they may not know how to use a smartphone or the Internet in general.

To overcome all of these issues would require a tremendous ground operation: making the people on the ground aware there is the Internet, explaining why and how it would be useful to them, helping them get a device and power, which might require donations or loans, showing them how to use that device, and so on. Then there is still the challenge of those users being on a low income. Thus even if there was an incredible use case, they might struggle to pay for it at least in the short-term.

Maybe if such an operation took place, it would have generated more revenue, and thus made Loon worth it. On the other hand, such an operation may just increase costs even more (especially an operation in very remote and sparsely populated areas) that the business model is even harder to be viable. Providing network coverage is not enough without these complementary activities; many mobile network operators are increasingly investing in such initiatives to enhance their business case, but these may not have happened with Loon’s partners—or they may not have shown results quick enough.

So, the lesson is not just about costs, it is about understanding the complexities of the revenue side of the equation as well. These cost challenges are similar for traditional base stations, though they may be easier: when someone builds a tower you see it and know it is there; the telco staff who build the tower could also market and sell the benefits of the Internet or devices, or provide skills training for a few days after building a tower, or they could license an agent in the village who would sell internet services. This requires people on the ground who would not be there if there is just a balloon floating overhead.

Finally, for those who invest in base stations, and design them with low operating costs, they can afford to wait several years until they recoup their costs, and then begin to make profits, since the infrastructure will stay there for a long time. It also makes it easier for governments to offer subsidies to build that infrastructure, as a one-time subsidy, than a recurring subsidy.

Loon has announced a package of $10 million to help support various related efforts in Kenya to support those in Kenya who might be affected by the loss of Loon’s service. This could make a real difference. As could the many programs that Google, a sister company to Loon, has been running for many years to provide digital skills, one of the barriers identified above.

So the lessons from Loon are not just about the needs for innovative technology that might reduce costs of providing Internet coverage, though that is still true, but about the needs for more investments in increasing revenues—providing devices, power, relevant content, awareness of that content, and skills to use the devices amongst many others.

Loon’s efforts were welcome and some of their technologies no doubt will be re-used in other ways, whilst the lessons learned may well help others working on similar initiatives. Their willingness to take risks, invest, and innovate is commendable and hugely necessary if we are trying to solve the problems of rural Internet use (not just coverage).

There are many efforts underway around the world in this area, but it is not enough. Let’s work together to do more!

First published on the Huawei blog.