Why is Internet Expensive and Slow?
At ZOL we often get questions like, “I just got back from the UK where my son/daughter has a 10meg Broadband connection for only US$20 a month, why can’t I get that here?”, or “How come in South Africa ADSL starts from US$10 a month?” Some less pleasant questions are, “Why is broadband such a rip off in Zimbabwe?”, or “Why are you like every other Zimbabwean business and just giving poor service at ridiculous prices?” Given that we do indeed charge so much for small amounts of bandwidth, we do owe you an explanation. The ISP market in Zimbabwe is very competitive with low barriers to entry – literally anyone can be an ISP for an investment of just a few thousand dollars – unfortunately, we really are not making huge profits! There are about 12 licenced Internet Access Providers (the companies that provide and operate the international and local links) and a score of Internet Service Providers who resell these services to the public.
There are five major contributory factors to the high prices:
1) Internet via Satellite
By far the biggest issue is that nearly all Internet in Zimbabwe is via satellite. Despite a recent fibre link installed by PowerTel (see article in this magazine), the majority of our Internet capacity still comes via satellite. The new fibre link is, in fact, more expensive and less reliable than satellite anyway! Unfortunately, satellite space segments are extremely expensive. In the US, larger customers can buy transit across the USA for as little as US$5 – 10 per megabit and you can buy gigabits at a time. Access between large ISPs is in fact free, as they interconnect to each other on a “settlement free basis”. Running a link from Harare to New York costs about 300 times more than a link from New York to London.
Fortunately, the effects of fibre are already being felt in other parts of Africa, such as Ghana, where an ISP recently said it had seen the monthly cost of bandwidth drop more than a quarter to US$1,625 per megabit per second, from US$2,250. The international carriers, fearing they could lose customers, have started cutting prices ahead of the new undersea cables landing. Countries lucky enough to have a coastline, are likely to see price drops over the rest of the year to about $350 per megabit per second. By the end of 2011, when two other cables have gone live wholesale prices could drop to as low as $225. However, compared to the US and Europe this is still very high.
Another cost factor is that, worldwide, providers oversubscribe or “share” their links on the basis that not everyone is online at the same time. In the US and Europe, almost everyone has broadband, hence a much smaller percentage of those customers are active at a time. Business broadband is typically shared 50:1 and residential at 200:1. We cannot do even half that – because when someone in Zimbabwe pays so much for broadband you can bet they are going to use it! In fact, most of our broadband connections are sold with no contention at all.
Satellite also explains why our broadband feels slower compared to the same speed in Europe or the USA. Data needs to go from Zimbabwe up to the satellite down to the USA – and then reverse back along the same path. Satellites operate at about 36,000km above the earth and the signal, which, even at the speed of light, takes about 500ms to make the round trip. This may not sound a lot – but a web page is made up of many images, often as many as a dozen. Each image has to be requested individually – so these 500ms all add up to several seconds.
2) “Last-mile” Costs
In Zimbabwe the “last-mile” costs almost as much at the entire satellite Internet link. The “last-mile” is the link from ZOL to your house or business – typically WiMax or UHF radio, or copper ADSL or fibre. Since the formation of POTRAZ in 2000, ISPs are not allowed to run their own infrastructure. Only Internet Access Providers (IAPs) are allowed to own and operate these “last mile” links. Prior to this ruling, ISPs like ZOL did self-provisioning of links at a fraction of the cost of current systems. Although POTRAZ has licenced a good number of IAPs, very few have managed to install a comprehensive network. Unfortunately, the licence costs US$4 million. With 12 licenced providers, the total cost to the industry to Government is US$48 million – before infrastructure is even deployed. Having said that, some providers were lucky enough to pay US$2 million in the Zimbabwe dollar days – meaning their effective price was very low. However, all IAPs have had to “top up” with another two million to upgrade to the Class A licence. Although POTRAZ does offer generous payment terms, one way or another these companies need to recover their licence fee money from consumers. This means not only is your choice very limited, but the service quality is low and prices are high. On a per megabit basis, it is more expensive to send a message from the centre of Harare to Newlands or Borrowdale, than it is to send to New York. So not only do you get hit once by high fees for your International bandwidth – but you are forced to double up and pay a high fee for local connectivity.
3) Business Costs
As everyone knows, costs of doing business in Zimbabwe is high. After years of under investment and erosion of capital, businesses are struggling to recover in a high interest rate environment relative to the US or South Africa. Whether you look at salary costs relative to productivity or communication costs (SMS, cell phone, landline charges), or just electricity or city rates, they are all much higher than those of our neighbours. Although government has done a tremendous job of lowering duties on communication and computer equipment, and has reduced many taxes, there are other efficiency gains that could be found. Due to power issues, we have to spend a disproportionately high amount on generators and UPS systems. ZOL has 8 generators at our key network points around the country, and there is never a time when none of them are running!
Finally, regardless of your political views, any business will tell you that sanctions are hurting all business in Zimbabwe and therefore all citizens. It raises the cost of supplies since some companies simple refuse to deal with us, whether or not we are “on the list”. PayPal, an online payment provider that allows you to send money cheaply to any one with an email address, is a perfect example. They have banned anyone located in Zimbabwe from using their system on the basis that Zimbabwe is a sanctioned country. ZOL has also had personal experience where companies have refused to ship to us once they know we are in Zimbabwe due to “sanctions”. Just last week we had an argument with another online services company that refused to deal with us on the basis we were under sanctions. Luckily we have some entrepreneurs like Mukuru.com that are willing to help us.
4) No Local Content
Zimbabwe has very little Internet content stored locally that is in high demand in the rest of the world. This is a subtle but significant point. The Internet works on content – he who has content has power. If your friend calls you from the UK, the UK carrier has to pay TelOne the portion of the call that is in Zimbabwe – i.e. the termination fee. Part of your friend’s phone bill gets paid to Zimbabwe. But if your friend sends you an email, the Zimbabwean provider (and therefore consumer) must pay 100% of the cost and no revenue is passed back to ISP in Zimbabwe. The same goes for websites – when you visit a website in the US and download a video – Zimbabwe (i.e. you) must pay 100% of the cost. Put another way – as an ISP in Zimbabwe we have to pay to link to “the Internet” in the rest of the world. No ISP in Europe or the US has ever called us asking for the price for a link to Zimbabwe because they want to communicate with us! Those living in Europe, and even more so in the USA, are essentially always using “local traffic” to surf the Internet – which is close to free!
5) Size and Density of Market
The scale of the current addressable market is just too small. Addressable market is the number of people who live in an area that can be easily and cheaply connected to the Internet (radio, copper line etc) – and can afford to do so. Consumer broadband lines in the UK and USA are shared up to 100 times over. This means that the bandwidth cost can be split over 100 users – since most of the time, most users are not using it. A similar principle applies to your landline or cellphone – for example most of Econet’s millions of subscribers are not on a call all at once – this means the “links” between your cell phone and tower can be “oversold”. In Zimbabwe, because Internet is expensive, and there are so few subscribers, we are not able to share more than about 10:1 in the best case – and for “business quality” broadband we share at 3:1 or not at all – depending on your package. Only when Zimbabwe connects hundreds of thousands of broadband lines, will we be able to share out those lines at more economical numbers. Another factor is the density of the population who can afford Internet. Laying out fibre, copper and even radio systems becomes much more cost effective where there are a large number of customers in a small area – for example areas with high-rise apartment blocks.
How can we improve all this?
Fortunately, things are changing fast! Fibre is coming, new networks are starting up, and competition is heating up. In the meantime, we must encourage all licensed carriers to provide us with alternatives to satellite. Zimbabwe is geographically well positioned to be a central switching point for Southern Africa. Government could give incentives for operators to create a large Internet Exchange – which would lower costs of Internet in Zimbabwe, whilst creating revenue and jobs for generations to come.
Finally Zimbabwe needs to develop content, which is not only a long-term project, but also is a chicken and egg problem. We cannot host content cheaply until we get cheap and reliable Internet access. However, there are many sectors such as call centre hosting, Voice over IP and outsourced services that can be hosted in Zimbabwe to increase local usage and content.
If all these plans can be put in place, we will see an increase in competition, more choice for the consumer and a dramatic drop in broadband prices with higher quality.