The future of South Africa’s telecom towers

Keith Boyd is a 13-year veteran of the African tower industry having served first as a Director at Plessey then as CEO of Venture Communications, later acquired by DPI and merged with Eaton Towers. Keith is now Managing Director of Eaton’s business in South Africa, and he’s kindly agreed to host the South African market round table at the TowerXchange Meetup, taking place on October 1 and 2 in Johannesburg. By way of a preview of that round table, TowerXchange caught up with Keith to ask his views on the South African tower market.(select style).

TowerXchange: Thanks for speaking to us today Keith and for hosting the South African market round table at the TowerXchange Meetup. Please introduce our readers to the South African market.

In South Africa, latest figures seem to show that Vodacom has around 45% market share, MTN 37%, Cell C is up to 16%, and Telkom Mobile (formerly 8ta) has 1-2%. Vodacom and MTN are very successful, cash rich operators. Telkom Mobile would appear to have a difficult road ahead, but they have options as the mobile arm of the fixed line operator Telkom. Cell C are within reach of becoming a sustainable, self-funding operator.

Cell C, which recently benefitted from a US$350m injection of equity from parent company Oger Telecom, are pushing for asymmetrical call termination rates, or to have call termination rates dropped from their current level of R0.56 (about 6 US cents) to nearer to 2 US cents. It will be interesting to see how the Independent Communications Authority of South Africa (ICASA) views their request given the recent pricing competition from newer entrants Cell C and Telkom Mobile.

About half of each of MTN and Vodacom’s subscribers are ‘data active’, and data usage is nearly trebling annually, increasing 193% last year according to MTN, driven by data bundle price reductions, and the uptake of smart phones.

TowerXchange: Would you characterise South Africa as a mature telecoms market?

Whereas SIM penetration in most of Sub-Saharan Africa (SSA) is around 50-85%, there are 70m subscribers in South Africa, representing 115% SIM penetration. Dual-simming is not as common as elsewhere in SSA, but anecdotally I do see an increasing number of people with more than one mobile device, whether it be a dongle for their laptop, a tablet or a business and personal phone.

When Eaton Towers first started looking at South Africa, we wondered if the market had matured beyond the point at which a towerco would have a significant opportunity for growth. So we commissioned two independent consulting firms to prepare a report on the future expected demand for towers, and PoS (Points of Service). The levels of growth in demand for PoS they forecasted were surprisingly high. Both reports - although utilizing slightly different approaches - came to a similar conclusion, while our own conversations with key market stakeholders painted a similar picture.

When we conducted these studies in the second half of 2011, there were just under 15,000 greenfield towers in South Africa, with 25,000 PoS in total including the towers plus DAS, rooftops, lamp posts et cetera. That was forecast to treble to just under 75,000 PoS in 9-10 years, with the bulk of the growth in the next 5-6 years. We knew if we could capture just 10% of the market growth, that would be a good business to have. With an increasing preference for infrastructure sharing to avoid spending capital on ‘poles and holes’, a 3:1 growth ratio suggested an opportunity to secure 1.6-1.7 tenants per tower relatively quickly.

TowerXchange: So how has Eaton Towers’ South African business evolved since your initial market studies?

We realised we couldn’t control the timing of when South Africa’s existing operators sold their towers, but we could control an organic market entry by building our own towers, which is what we decided to do at the end of 2011.

We spent 2012 acquiring, permitting and developing sites, signing up operator contracts and tenancy agreements, and have an expansion plan through 2013-17. Demand has been higher than we expected, and the average number of tenants on Eaton’s towers in South Africa are higher than in our original assumptions and business plan. Now it’s simply an issue of getting to scale in the next year and more.

To date, we’ve acquired approximately 400 sites, and fully permitted more than half of those. Permitting can take 9-18 months in South Africa, so you can’t build 1,000+ towers overnight. We have already completed construction on over 80 of these sites (August 2013), and are constantly adding more sites as per increasing customer demand.

The simple reality of our business is that, at the pricing levels we charge, any tower with one tenant is a liability - or at best, a very poor investment. Towers with two tenants are an asset. Fortunately, Eaton has signed up more than two tenants per tower on average already in South Africa.

When we market a site to an operator, it’s already fully permitted, and we have a framework agreement in place, so both parties know the price. All they have to do is sign an ISA (Individual Site Agreement), and we can have them installed on a new site in 30 days. We break ground as soon as we have a single anchor tenant signed up, and we’re able to react quickly to operator demand - in fact the passive infrastructure can often be ready faster than the time taken to get the transmission links in place and to get the active equipment installed.

TowerXchange: How does running an organic growth towerco differ from a business model built on a sale and leaseback transaction?

We have to make informed guesses where coverage gaps in the networks will appear, as voice and data capacity demands grow. It can be as simple as choosing sites more than 500-600m from an existing tower, where people live in the shadow of the site. The truth is that it is not that simple for network planners at operators to be able to predict accurately more than a year ahead in terms of their network densification requirements, and the acquisition and permitting work can take several months, so towercos offer significantly reduced time to market, by speculatively acquiring and permitting sites.

Our formula isn’t complex. As soon as one tenant needs a site, we build it. If we can add a second tenant in under 2-3 years, we’ll be successful, so it’s a case of assessing each site to determine how strongly we should market it.

TowerXchange: What would be the impact on your business of the rumoured transaction between MTN and American Tower in South Africa?

I expect to see American Tower make more investments in South Africa. But I can’t comment on speculation linking them with another deal with MTN.

Quite frankly, I’m glad American Tower is operating in South Africa as well as Eaton Towers - it makes operators more familiar with the towerco business model and price points. We’re not really in direct competition with American Tower - because uniquely located sites are exactly that - uniquely located. In our entire portfolio there are probably two sites where our tower is within 500-600m of American Tower’s. As an organic build towerco, if we saw a new site go up where we had acquired land and completed permitting, we probably wouldn’t market the site, as the supply:demand ratio would be out.

TowerXchange: How would you anticipate the South African tower market evolving if MTN and American Tower did get together?

It would be interesting to see how Vodacom and Telkom would react. Both have larger portfolios of towers in South Africa, but there is overlap with the MTN portfolio, so I doubt it would make sense for them to deal their towers into the same vehicle as all the others. That would be very poor for competition reasons - and no towerco would want to buy assets where they already have significant numbers of “near neighbor” sites.

At Eaton, we are keen to expand and accelerate our investments in South Africa. And I expect that our local commitment and presence, coupled with our well funded position, in-market experience and customer base, would make Eaton Towers an excellent counter-party for any future towers transactions in the country.

Vodacom specifically has an excellent opportunity to use their towers to drive their cost base as low as they can over the next 10 - 15 years and beyond. They could structure a deal with a towerco to secure an immediate opex subsidy in the region of 25-40%, creating industry-leading low opex, thereby putting them in a great place to prevail in any future price wars. Dropping opex from around ZAR9,000 per month to ZAR6,000 simply cannot be done through efficiency programmes, so the towerco counterparty would have to agree an EBITDA-negative deal.

The only place that the guaranteed “opex subsidy” comes from is the tower company’s shareholders, through their income statement. Obviously, at some point in the future, we would need to believe that the tower portfolio becomes EBITDA positive - but that is not something the operator would need to worry about. Their greatly subsidised site opex would be fixed by the towerco. Eaton Towers is the only towerco that has done a deal on this basis in Africa. Whilst our shareholders have a positive view on these types of transactions, I don’t see how publicly listed towercos, that are currently quoted as trading at a multiple of their EBITDA, could bring their management and shareholders along to do such a strongly EBITDA negative deal.

Meanwhile state-owned fixed line operator Telkom is estimated to have over 6,000 shareable structures towers. It’s the largest ‘poles and holes’ network in SSA. Telkom has a great opportunity to do a tower deal. It’s not about the Telkom Mobile rollout, they can do that using other infrastructure. It’s an opportunity to become a carrier of carriers, to release cash and drive down opex, securing a stake in a towerco that would be a boost to the valuation of their business. In order to do this, Telkom would have to clarify the implications of the Facilities Leasing Act, which forces “dominant” operators to sell slots on their infrastructure on a “cost plus” basis, but whether Telkom can really be considered a dominant operator in the mobile-era, and the exact definition of “cost plus” should both be up for debate.

TowerXchange: Have there been a lot of bi-lateral tower swaps in South Africa? I recall the suggestion that one operator in South Africa had third party tenants sharing 50-60% of their towers.

Historically there have been quite a few swaps of slots, but on a piecemeal basis. Bi-lateral swaps are not customer service oriented, and the turnaround time between requesting and securing a slot can be as long as 12 months.

I’d describe bi-lateral swaps in South Africa as significant but not strategic. Mobile networks have been functioning since 1994, so over the nineteen and a half years they’ve been operating towers, if the lease up rate is still only 1.5, then you can see they haven’t given the same attention to adding multiple tenants that an independent towerco would have.

TowerXchange: What impact will LTE have on the South African tower market?

LTE could be a shot in the arm for the telecoms business, and ICASA is expected to issue licenses in the coming months, but we don’t know yet what their strategy will be.

The impact on the market will depend whether ICASA gives licenses to incumbent operators who then supplement their existing 3G networks with LTE, or if they license new operators who are then motivated to build as much LTE coverage as they can in high demand areas, catalysing the data market.

Although operators are already running some LTE sites, there’s still a lot of ground to cover in South Africa’s LTE story.

TowerXchange: Beyond the four leading mobile network operators, are there any niche players that are significant potential tenants on South African towers?

Internet Solutions and Neotel are two of the more interesting operators. And I expect that LTE licensing in future will see new entrants launch.

Neotel is a non-mobile player targeting the SOHO (Small Office Home Office) segment and using WiMAX, WiFi and CDMA. They have a lot of rooftops in their infrastructure. While they may not be completely self-funding yet, they have an interesting business model in not competing in the retail market but focusing on broadband for business and high end users. Their opportunity seems to be partly due to current data capacity issues in many areas of all the GSM operators. If the last year is anything to go by, demand seems to be increasing faster than the mobile operators can expand and densify their networks - and I think Neotel are capitalising well on this.

TowerXchange: For the benefit of our supplier readers, what capabilities does Eaton Towers South Africa have in-house, what do you subcontract?

We retain project management oversight, quality, health and safety responsibilities, but we subcontract site builds to full turnkey contractors.

Our field maintenance is outsourced, but we manage service levels, call escalations and keep the NOC and SLA management in-house.

With the exception of a few permitting activities, our property team is managed in-house and permits 15-20 new sites per month.

TowerXchange: What’s the energy logistics state of play - how extensive and reliable is South Africa’s grid?

We have over 99.3% availability of grid power across our portfolio. Uptime is so high that we haven’t got fixed generators on sites - we simply use mobile gensets that we can dispatch to sites within three hours of a grid failure,  before the operators’ batteries are depleted. So the focus on energy issues in South Africa is lower than anywhere else in SSA.

TowerXchange: What are the implications of Eaton Towers’ success in South Africa for the rest of the African tower market.

I am pleasantly surprised that the South African market has already offered significantly better than expected growth for the tower sharing business. If you’d asked me in 2010, I would have thought South Africa was a mature market, but now we see a great opportunity for the tower industry here.

Data-wise, South Africa is the most developed market in Africa. I recently met a consultant who forecast data consumption would rise to an average of 1GB of data being used per day per data subscriber. I don’t know whether we’ll reach those kind of numbers in my time, but that’s only an hour of HD entertainment. We’ve all been guilty at times of underestimating where technology can take us, and if 1GB per subscriber per day were the future, that would represent a thousand times increase on the data demand we have today. If active equipment becomes ten times as efficient, you still have a hundred times capacity to make up, and it still presents a tremendous opportunity.

That’s why there’s upside in the South African tower market, and if you consider the relative data consumption in Eaton’s other tower markets, Ghana, Uganda and Kenya, then you can see there is still a lot of potential runway for towerco growth elsewhere in SSA as increasing data consumption drives cell site densification.

The insatiable demand for voice coverage that drove the initial land-grab network rollouts in Africa is going to be replaced with an insatiable demand for data. Operators are going to need to accelerate project management again to densify the last mile, or the last 400m as it is today, and to keep up with transmission requirements.

If operator Points of Service treble between 2011 and 2021 in South Africa, as our forecasts suggested, operators will continue to escape the huge capital requirement this generates by using independent towerco sites. They’ll have to keep all their capex focused on active equipment. Eaton Towers has made a great start to our organically grown South African business, and we see plenty of room for growth in the future. And, if South Africa is not a mature market for the towerco business model, then there is even more room for growth elsewhere in SSA.

Click here to register today - join Keith Boyd’s South African market round table at the TowerXchange Meetup on October 1 and 2.

Cell C are rolling out aggressively and currently have around 4,000 sites in South Africa. Speculation continues about potential M&A activity, either the consolidation of Cell C and Telkom Mobile, or the potential market entry of Airtel through acquisition.

In the tower market, American Tower acquired 1,400 sites from Cell C in 2010, in a deal structured to release cash. American Tower is currently marketing 1,629 sites in South Africa, of which just over 1,000 are greenfield sites.

Recent press speculation seems to indicate that MTN South Africa are in the later stages of agreeing a tower deal with American Tower. MTN has already established partnerships with American Tower in Uganda and Ghana, in both cases MTN retained 49% equity in the local towerco. MTN are estimated to have over 3,500 towers in South Africa, with a presence on another 2,000 sites.