Joint venture infrastructure companies CTIL (Vodafone + O2) and MBNL (Three + EE / BT) operate parallel yet subtly different siteshare businesses, and operate more than half the UK’s ~40,000 active towers. Each has different models of sharing and asset ownership. CTIL and MBNL each has around 20,000 sites, around two thirds of which are their own, the rest co-located on a variety of independent towers. The UK’s leading independent towerco Arqiva has 10,550 towers, Wireless Infrastructure Group has 2,000, Shere Group 500, and a variety of smaller towercos and other stakeholders make up the other ~1,450 sites.
Consolidation among UK MNOs; implications for the tower industry
The spectre of a third major restructuring of the UK tower landscape within the past few years seems to be receding, with increasing indications that the O2-Three (Telefónica -Hutch) merger will be blocked by Brussels. At time of writing the formal announcement was expected on May 19, 2016. However, with significantly reduced dialogue about remedies, many designed to foster the introduction of an alternate fourth operator, the likelihood of the merger being approved appears increasingly remote. Speculation has already commenced that Telefónica may seek an alternate counterpart through whom to exit the UK, with Liberty Global almost inevitably mentioned.
In the increasingly unlikely event that O2-Three was permitted, there was a growing view that regulators might force the divestiture of O2’s towers, giving a prospective new entrant third operator an opportunity to ‘hit the ground running’ by acquiring a network.
“We’ve looked at being a fourth entrant MNO in the UK, and there is a model that we think works,” said one participant in the UK roundtable at the recent TowerXchange Meetup Europe. “We know the number of towers we need, it is doable. But any new entrant would have to make it economically advantageous – they would have to reduce capex and opex relative to incumbents – which implies a need for different business models like some of those we’ve seen at the TowerXchange Meetup over the past two days.”
Another participant concurred: “New market entrant options to compete and differentiate on price may be limited, given the finite room to maneuver offered by handset prices. This would suggest any four to three consolidation include remedies so severe that the cost base for a fourth entrant is discounted, otherwise the experiences of Three, which struggled to achieve profitability, suggest the cost base is unsustainable.”
Parallels were draw with the Irish tower market, where Three was also the fourth operator, where they struggled to compete with Vodafone, Meteor (Eir) and O2, and where eventually Three had to acquire O2 to achieve scale. The parallels were even more apt given that O2’s partnership in NetShare with Vodafone had to be broken up as a condition of the merger. While in Ireland the merger was permitted, in the UK it seems less likely.
Whether the UK MNO landscape consolidates from four to three or remains at four, some roundtable participants suggested it was interesting that the CMA opposed O2-Three when BT-EE passed relatively seamlessly. Perhaps as a vertical merger, in which spectrum holdings didn’t materially change, there was less immediate cause for concern, even though the transaction will leave BT in a very strong position. BT acquiring EE introduced a maverick new operator, whereas O2-Three was a contraction affecting both the primary and wholesale markets.
Whilst MNO consolidation remains uncertain, the principle for business operation in the UK tower market has been and shall remain “business as usual.”
What does ‘business as usual’ look like in the UK?
The structure of the UK tower market is unique because the two joint venture infracos, CTIL and MBNL, are both the primary clients of the UK’s independent towercos, and are sitesharing businesses in their own right.
Back in 2009, Vodafone and O2 established a joint team called Cornerstone to share passive infrastructure. Cornerstone evolved into CTIL in 2012 and now programme manages Beacon, the parent companies’ joint 4G rollout. “CTIL are currently focused on acquiring sites for Beacon to meet the license obligations of our parent MNOs by 2017.” Ofcom requires that O2 cover at least 98% of the UK population with 4G by 2017.
“CTIL is a passive share,” said UK roundtable moderator and CTIL CEO Malcolm Collins. “We manage telecom property and real estate; we own assets of both our parent operators, whereas MBNL’s parent operators own their assets, but MBNL’s model extends to active equipment and transmission sharing.”
While CTIL is an AssetCo, the UK’s other joint venture siteshare business, MBNL, is a ServiceCo – MBNL don’t own the assets. Both joint ventures have yielded significant network consolidation, decommissioning and cost reductions, while accelerating 3G rollout.
MBNL was envisaged in 2007. It’s a RANshare using the MORAN model between Three and EE (becoming BT). MBNL claims to have enabled ~£1bn savings each over their initial seven years and that has been realised.
CTIL and MBNL have a good relationship each other and share around 500 of each others’ towers.
CTIL has over 11,000 of their own structures, including streetworks and rooftops, with a total network approaching 20,000 sites, including third party sites, which they use for reasons of “speed, cost and quality”. MBNL has similar tower numbers. “We still have to build a bunch of towers for 4G,” said one of the UK siteshare businesses. “We take the independent tower companies’ maps and work out where they aren’t – we have to build such sites ourselves.” While there is some new build in the UK, both JV siteshare businesses have undertaken a deep process of network rationalisation; TowerXchange estimate there a little over 40,000 active towers in the UK, and a little over 50,000 towers in total.
Tower ownership and co-location in the UK
A call for greater transparency into MNOs’ network capex
“While five or ten years ago the UK’s MNOs were keen to deploy their own network capex, their ability to invest in their networks is now reduced,” said one UK independent towerco. “Yet despite this, there isn’t much transparency into their plans. Towercos have capital to spend – tell us where to spend it!”
Another towerco concurred: “Especially as UK networks extend into rural areas, MNOs are going to need taller towers, and to get transmission out to them. While towercos like us have such sites, we’re not seeing footprints disclosed – we’re desperate to find a way to better engage.”
What would the tower industry ask of MNOs in terms of understanding projected demand over a longer period?
“I can get coverage demand, but the site specificity doesn’t come until late in the day,” said one towerco. “If we get a batch of nominal (search ring), we respond almost overnight, but then we’ll often not hear about that for a long period of time. When communication is patchy, we don’t know if the requirements is real enough to resource it,” added another.
Towercos and siteshare businesses in the UK appealed for progression toward more of a partnership model, wherein MNOs would share their forward requirements. It seemed no one was seeing even one year forecasts of site demand from MNOs, let alone five years – a time horizon some felt unrealistic as too much would change.
The fact that UK networks are shared along an East-West divide was felt to represent a challenge to this kind of partnership model. Similar split country models have been used in Sweden and Spain, one participant suggested the approach had worked well in the latter because the project managers from the partnering MNOs were cousins!
“It’s all about trust,” said one participant. “You sit down and define the value chain starting from the MNOs’ own patchy information, to the siteshare business seeking to serve an ill-defined demand. They in turn ask towercos for help… the towerco makes money, the siteshare business saves time, the MNO saves money… There’s enough pie for everyone to have a slice!”
“Even in Myanmar, where towercos build everything, there is more disclosure of where the MNOs need the sites,” said another roundtable participant. “Towercos transparently facilitate rollouts in other markets, but we haven’t achieved that degree of integration in the UK.”
UK mobile coverage, by % of premises
Uncommercial coverage
“We’re talking about uncommercial coverage, driven by regulation,” countered a towerco. “So the environment is different.”
“The current 98% coverage requirement would need 1,000 extra sites in our network,” said another participant, “so we’re digging through spreadsheets finding out what’s there, and what we need.”
“The MIP (Mobile Infrastructure Project) had noble intentions to provide coverage where there wasn’t even 2G, but it hasn’t been particularly successful,” added another participant. Under the MIP, the UK government provided £150mn to fund the connection of “not-spots and partial not-spots”, yet when it concluded at the end of March 2016, the project was expected to have delivered just 60 of the 600 masts identified in the original plan.
“MIP failed because of lack of transmission infrastructure – microwave is expensive – so the per site cost went through the roof,” said one participant. “The MIP programme came in the middle of the 4G rollout, and the reality is that the MNOs never wanted it. It should be revisited in three to four years when there’s more fibre. But if the government wants to make the infrastructure free in uncommercial rural areas, they need to fund transmission as well as towers.”
One of the outcomes of MIP is an increasing government appetite to permit taller masts. “the regulators’ relaxation of planning laws to allow higher masts will accelerate site acquisition, possibly as soon as Summer 2016,” said one participant.
New build volumes in the UK
One towerco estimated there were around 2,000 new points of presence (PoPs) per year being added by the UK’s four MNOs via CTIL and MBNL.
CTIL’s remarks suggest an upward revision of that estimate: “We’ll add 700-800 rural sites this year, 500 in London, a couple of hundred infills in other cities; maybe 1,500 total. We prefer existing tower structures to accelerate time to market in rural areas, but sometimes we see the ‘site’ on a tower company’s database yet when we get there it’s a green field!”
A towerco added: “We expect more demand to upgrade structures than build new sites. Many MNOs built smaller structures, some of which are upgradeable, others will have to have foundations and/or towers replaced.”
if the government wants to make the infrastructure free in uncommercial rural areas, they need to fund transmission as well as towers
Declining NIMBY concerns
Whilst the battle to provide rural coverage and urban capacity continues to be a challenge, one side effect is a reduced NIMBY (not in my back yard) mentality on the part of ‘Joe Public.’
The concerns of the British public are apparently now more focused the implications of coverage for the value of their real estate (and the importance of getting good coverage on the toilet!), than they are concerned about the health implications of cell tower proximity.
Rail infrastructure
The UK is one of several European markets where existing and potential new trackside masts could be very valuable both for coverage in adjacent communities, and potentially for coverage on trains.
“We use network rail infrastructure in a big way, but we’re very restricted in frequencies and power if we put sites on their land,” said one participant.
“There’s often a degree of engineering inertia when dealing with rail and other State or former State infrastructure,” said a towerco. “No-one knows the assets are worth anything.”
“Delivering free Wi-Fi on trains requires near complete transmission,” said one participant.
“I’d anticipate increasing use of LTE relay technology for localised coverage,” suggested another. “It won’t work at that speed. We use a wide band repeater on each car for this in other European markets,” countered a third.
