Zambian internet is priced right; Here’s why
A recent Facebook post regarding the high cost of the internet in Zambia has me unusually jumping to the defence of service providers. The post was by a leading Executive in Zambia; someone who has financial information at their behest. If he believes internet service providers (and mobile telcos) are taking advantage of the consumer what more the average user of the service? Consumers hear stories of unlimited 100Mbs internet for $35 a month including TV and think the service providers are making insane profits! I am here to clear up that misconception!
Let us start with the basics of where does the internet come from and what is this thing we surf? In general the internet is everywhere but specifically the internet we like is hosted in data centres of Europe. The internet then is Facebook, Google, Youtube, YourCompanyNameZambia.com (see our piece on that here), the list is genuinely inexhaustible – we like the internet in Europe. Now to get the internet from Europe to Lusaka, Ndola or Kitwe there has to be interlinking infrastructure from a data centre in London, England to your house in Zambia. Rhodes found it an almost impossible dream to build a railway line from the Cape to Cairo, the cable running the breadth of the earth is no small feat.
I will start with an analogy on why the cost is high before going into the specifics of the costs of the internet in Zambia. A farmer in Dublin grows potatoes to be sold at his local market and eaten by patrons in a restaurant in Lusaka. The distance from Dublin to Lusaka by road according to Google is 12,039.8 km. With that distance we know it does not make sense for the farmer to grow and transport the potatoes to the end user on his own. He has to sacrifice a higher final selling price for lower input costs meaning he must rely on a chain of middlemen who aggregate transport loads and have many market locations to sell his potatoes. These transporters in turn sell to other transporters who specialise in dealing with local border and import controls. This process continues until the fried chip is presented in your plate. As you can well imagine if the farmer sold a potato for $0.10c your plate of chips will cost many times that! The analogy has a few flaws but it serves the introductory purpose well.
Here are the specifics of why you pay what you pay to surf in Zambia. When a group of coders and content creators make a single episode of House of Cards for Netflix they have just added 450MB (megabytes) to the internet. That episode of House of Cards will live in a data centre in Europe (to the geeks out there yes Netflix operates CDNs and peers freely but that is beside the point). The data centre and its tenants need to make money, they make money in 2 ways, by selling rack space and by selling bandwidth. Because the data centre is large bandwidth tends to be sold in large chunks of 10Gb/s or greater (Zambia as a whole does not consume 10Gb/s of international capacity). Capacity at this magnitude is bought by Tier 1 providers like Level3 or Hurricane Electric not MTN or Airtel (they are not Tier 1). The Tier one providers pay approximately $1.00 per Mbps for 3Gbps in London (source: Telegeography).
The Tier 1 providers sell the capacity to wholesale providers who tend to specialise in regional (Africa) capacity provision. This leads to the first major bump in price, crossing the ocean. The map above shows the undersea cables that service Africa. The cable owners have multi-million dollar investments to recoup; for example SEACOM on the East Coast cost $375million to build initially. SEACOM now has a stated capacity of 4.2Tbps (terabits) and is valued at $600million after several upgrades. The Tier 1 providers are in the business of aggregating capacity and selling it to providers who specialise in “African capacity” like PCCW and WIOCC. The specialised providers (Tier 2) buy their undersea capacity in the form of Indefeasible Rights of Use (IRUs) from the owners of the transoceanic cables. IRUs are contracts that last for many years usually a decade or more. When you have to recover an investment of $600million you want a tenant who will be with you for a long time! The regional providers are responsible for finding a market on the continent and purchasing a large enough capacity to get a good price. The Tier 2 providers initially have a loss making asset on their books. We know they bought it for approximately $1.00 a Mbps in London but they now have to make a profit whilst paying for the following: equipment in London to connect them to the Tier 1 provider, staff in London, cable lease, equipment on the coast in Africa and staff in Africa. This is where it gets a little complicated. The Tier 2 providers leases capacity from the cable operator and also pays the Tier one provider for the internet (House of Cards episode). Capacity is often quoted in multiples of STM1 (155Mbps) and for the rest of our calculations we will use a single STM1 as the foundation cost:
STM 1 lease cost from Cable Owner $8000 (source: ISOC)
STM 1 internet capacity from London $155
The cost to Tier 2 provider for 155Mbps landed in Dar es Salam is $8155.00 or $52 per Mbps. The internet now has the terrestrial segment to Lusaka. In Dar es Salam a local provider may buy capacity and onward sell it but in our example the Tier 2 provider will transport the capacity all the way to Lusaka. As we have already established Tier 2 providers are not in the business of building cable systems – they are in the aggregating capacity. To get from Dar es Salam to Lusaka they have to engage companies like Tanzania Telecommunications Company Limited (TTCL) to transport the 155Mbs from Dar es Salam to Tunduma. To get from Dar es Salam the cost of an STM 1 is $5000.00 this is because terrestrial capacity has fewer efficiencies than undersea capacity resulting in higher per km installation costs, higher per km maintenance costs and the companies running them tend to be state controlled resulting in less financial efficiencies.
At Tunduma the Tier 2 provider has to figure out a way the internet to get to Lusaka. Border crossings entail inter-government agreements and changing terrestrial carrier.To get from Nakonde a terrestrial provider like ZESCO or Liquid must be used and we presume the same operating model as Tanzania with the STM 1 from Nakonde to Lusaka costing $5000.00. Because the internet is defined as an imported service in Zambia, the final buyer of the capacity must add on 20% Withholding Tax to the $13,155.00 making it $15,786.00 plus the cost of Zambian terrestrial provider to Lusaka giving you a total of $20,786.00. Add on 30% markup for the tier 2 provider and you get $27,021.80 or $174.33 per Mbps or 17,300% more than the cost in London!
To neatly summarise the costs so far we have tabeled the figures below. The table gives you the price (volume purchased dependant) that MTN, Airtel, iConnect, etc would typically pay if they bought capacity on the lower end of the spectrum. Remember this is a single link to the internet and all providers currently have multiple links to the internet (maybe not to your house).
Internet Costs
| Segment | Price | Price per Mbps |
|---|---|---|
| London | $155.00 | $1.00 |
| Undersea | $8,000.00 | $51.61 |
| Tanzania STM1 | $5,000.00 | $32.26 |
| Witholding Tax | $2,631.00 | $16.97 |
| National STM1 | $5,000.00 | $32.26 |
| Tier 2 Provider Margin | $6,235.80 | $40.23 |
| Price to Service Provider | $27,021.80 | $174.33 |
If the provider has 2 routes to the internet all the above costs will double giving you a cost of $54,043.6 per STM or $348.67 per Mbps.
To deduce any costs beyond this would be conjecture. However, we can list typical costs for a service provider:
- Staff (technical, support, finance, marketing)
- Equipment purchases (routers, switches, servers, radios, fibre)
- Space leases (towers, rooftops)
- Transport costs (installations, maintenances)
- Marketing expenses
After all of the above the service provider then adds their mark-up and the relevant taxes. The taxes will include: ZICTA 3%, VAT, any associated WHT and corporate taxes. $1000.00 per Mbps is very justified – we must enjoy the shared links because without them we would not be able to afford to surf.
Next time someone complains why the internet is expensive explain to them it is because of the two infallible truths – geography and taxes!
I did not capitalise the i because the internet has become part of Maslow’s hierarchy of needs 🙂
we understand all that, looking at the volumes of subsea internet capacity these people recouping there investments would need less than 3years on average and then would you say the cost of transporting the internet will still be as expensive? Most of these systems have been around for atleast 5years and am sure the cost and maintenance is so minimal its hardly noticeable but still the prices are still higher if we compare to nearer countries like RSA where they now have a 1gbps line for slightly above $200, can you still justify us getting a 1gbps line coming at almost $200, there must be something wrong even if we need to build long terestrial networks to get the internet here
Hi I think to recoup $600million would lead to extremely high prices. I believe the next major pricing reduction will come in the next 2 years as some of the initial IRUs come to term and additional cable capacity is activated. We hope our terrestrial companies are ready and have also upgraded their networks.
I don’t think Geography plays much of a role anymore, nor does taxation. Yes each hop does add to the cost, but until the data actually hits our shores, it’s quite cheap. Using your logic, Namibia should have cheaper monthly costs per mb as opposed to say South Africa or Zimbabwe, by virtual of its closer proximity to the European IX, but, no. South Africans pay half of what Namibians pay, while Zimbabweans pay around a third less than what Namibians do. Going by one of the reports you referenced on internet bandwidth trends in Africa, costs have dropped from nearly 50,000 dollars to under 5,000 dollars, which is around 90% in reduction. And even if the entire market, wouldn’t compensate in a similar way, the expectation is, there should be some semblance of gradual drops. But that never happened. So I am quite certain the high costs of internet services in Zambia has more to do with an unregulated market, and an overly controlling regulator than it has to do with tax and geography. This can be seen in the very large discrepancy between the cost of typical Zamtel link and an iConnect home link.
Hi thanks for your comment. The main reason geography plays a part is the closer to the coast you are the cheaper your capacity costs are. South Africa and Namibia have very low costs. South African and Namibian companies are part of the consortiums that own the cables – Zamtel owns a portion of an East Coast cable but their price into the market makes it more expensive. Zimbabweans pay less than Namibians because of market size and terrestrial distribution costs – Namibia has a sparse population making it costly to distribute internally. The national cost of Internet tends to be blended so as not to disadvantage rural or remote users by having pricing tiered by geography. The non-tax or geographical reasons are the availability of in-country CDNs (content distribution networks), owning the infrastructure in another country (like Liquid in SA) and demand volume. An ISP in Zimbabwe uses more capacity than Airtel and MTN combined leading to better volume discounts. As Zambia we need to host stuff locally, get past Facebook as the internet and have a government that views low cost internet as a cog in growing the economy.
Well explained. I hope the public now understands why it is so expensive to bring Internet to Zambia.
The Zambian government is pursuing an economic diversification program to reduce the economy’s reliance on the copper industry. This initiative seeks to exploit other components of Zambia’s rich resource base by promoting agriculture, tourism, gemstone mining, and hydro-power.