stds-802-16: MRD to FRD input

[Peter Ecclesine <petere@cisco.com>]

[Notice: It is the policy of 802.16 to treat messages posted here as non-confidential.]

A cautionary tale from Ionica(as published by TotalTele):

A matter of life and death

 01 April 1999 

 Ionica had a cunning plan: to become a UK national operator via the use of WLL technology. But it failed - and Ionica went the way of all flesh. Caroline Wallace and Dr Stephen Unger conduct a post mortem on the operator's brief life and find a lot can be learned from a corpse 

 Ionica floated on the stockmarket in July 1997. On the back of its innovative wireless local loop (WLL) technology, the company's shares attracted a price of £3.90, (US$6.50, £5.80), which put the company's value somewhere in the region of £640m ($1 billion, £960m).

 Fifteen months later, on October 29 1998, after suffering a catastrophic collapse in its share value and having failed to secure any finance to continue its growth, Ionica's management team put the company into administration and the would-be mould-breaker became an ex-operator, overnight.

 So what caused such a dramatic disintegration? Did Ionica bite off more than it could chew, or were more subtle and complex factors at work, which could not have been foreseen?

 Ionica had an ambitious vision: to use innovative technology to become a major player in the UK - one of the most competitive telecomms markets in the world. In choosing such a battleground, it took on a far more demanding task than if it had decided to operate in the developing world, or to become part of hybrid wireless-wireline networks.

 At another time, and probably in another place, Ionica might have been able to realise its vision. That it didn't is partly due to the prevailing financial climate, along with a degree of corporate hubris.

 Ionica's WLL technology was based on Proximity I, a proprietary TDMA-based radio system operating at 3.5GHz, manufactured under licence by Nortel, to deliver voice and data traffic across the 'last mile' between radio base stations and customers' premises.

 The optimism that surrounded Ionica's flotation stemmed from the claim that the costs of such a system are scalable with customer numbers. Wireline operators incur a large proportion of their infrastructure costs per home passed - whether or not customers take the service. The cost per home passed for a wireless system is far lower; perhaps £30 ($50, £45), compared with £500 ($850, £748) for a wireline operator. As most of Ionica's infrastructure costs would be incurred at the point of customer installation, it was argued that a revenue stream would quickly have been established.

 One of the most widely reported causes of Ionica's downfall was its failure to meet network rollout targets. This is illustrated in Figure 1, which shows the coverage Ionica achieved, compared with the targets set by the PTO licence, the prospectus dealing with the Initial Public Offering (IPO), and the banks which provided Ionica's senior credit facility.

 The PTO licence requirement was extremely demanding, requiring Ionica to cover 75% of the population of the UK (excluding Scotland) by February this year. In the IPO prospectus, Ionica acknowledged that it was unlikely to meet this obligation, and Oftel did subsequently propose its removal, but only two weeks before Ionica went into administration.

 The coverage targets agreed with Ionica's financial backers were less ambitious. They specified coverage of 3.8 million homes by the end of September 1998 and this target may have been achieved had financial pressures not resulted in network rollout being put on hold at the beginning of that year - when a substantial number of base sites had been acquired in London and just before service was due to be launched in the UK's capital city. Had these metropolitan sites become operational, coverage figures would have been boosted significantly, but, as things turned out, Ionica had covered only 2.8 million homes when it went into administration.

 The target for customer numbers was much more challenging. Banking covenants required Ionica to achieve 195,000 customers by the end of 1998, but, by the time it went into administration it had just 62,000.

 Even had the coverage target been met, Ionica was still required to achieve a penetration of about 5%, (within a few short months of launch of service in some areas), to meet its customer number targets.

 Thus the company not only had to improve market penetration in the areas where it did provide coverage, but it also had to exceed the coverage targets originally agreed with its bankers, to provide a bigger market in which to obtain customers.

 Ionica failed to achieve its penetration targets for two main reasons:

      The capacity of the network as built did not match the pattern of demand for the Ionica service, leaving some base sites loaded up to (and beyond) theoretical capacity limits, while other sites were embarrassingly empty. 
      A software upgrade that would have solved some of the capacity problems was badly delayed. (When it was, eventually, rolled out into the network in April 1998, the capacity of a typical base site was effectively doubled.)



 Figure 3 shows the average base site utilisation (or fill-factor). We have estimated it using publicly available information on customer and site numbers, and average sectorisation, and assuming industry accepted figures for peak traffic per customer (70mErlangs) and grade of service (99.9%). The fact is that the fill-factor remained below 40%, even during the period when the capacity limitations hit hardest in summer 1997.

 So why did Ionica's network capacity not match the pattern of demand for service? When customer service was launched, in May 1996, Ionica positioned itself as a service-oriented provider of telephony to high-end residential and small business markets. Alongside much-vaunted promises to undercut BT on line rental and call charges, Ionica's advertising emphasised its hi-tech features, aiming to attract high-spending, technically literate users. Ionica's network was primed on the basis of this marketing strategy; capacity being provided on each site according to how many households in Ionica's target market segments were covered.

 In practice, Ionica found its service appealed largely to the lower end of the residential market, where price sensitivity is much more acute. Up-market households remained indifferent to the price message, and were also disinclined to change service providers.

 To make matters worse, Ionica faced significant competition in the local loop from the CATV franchise operators. Their bundling of television and telephony services was attractive and subscribers failed to respond to Ionica's technical and service blandishments. Ionica's initial network rollout strategy did not adequately address the need for accurate cable intelligence and network dimensioning.

 Because the network had been dimensioned according to Ionica's initial marketing strategy, some sites with insufficient capacity began to fill up unexpectedly, while other sites that were over-provisioned remained empty. As a result, Ionica had to implement point-of-sale capacity restrictions. A nightmare scenario wherein sales to new customers were prevented, because an uptake of new subscribers would have compromised the grade of service offered to existing customers.

 Had Ionica's network, and network planning, been flexible enough to respond quickly to the evolving patterns of customer demand, neither the delayed software upgrade nor the ineffective, and misguided, marketing strategy would have been so problematic.

 The upgrading of a mobile network is fairly straightforward - since customers can receive traffic from any transmitter in range, temporary degradation of the signal from an individual base site is not critical.

 However, where WLL networks and conventional systems differ is that with WLL each customer has a directional antenna pointing at a single transmitter site, which greatly reduces flexibility. Thus, every time a new site or a new sector is added to the network, the risk of interference to existing customers must be modelled, and all subscribers at risk of interference from the new sites must have their antennas repointed towards a new transmitter site.

 The continuous upgrading of a network in this way becomes a tightrope walk, as the operator balances the number of extra subscribers gained against the costs incurred in re-aligning the antenna.

 Thus, efficient processes for the installation and re-installation of customer premises equipment are integral to the proper economic management of a WLL network. However, as there is no close analogue in either mobile or wireline networks, Ionica had to develop such a regime from scratch.

 Furthermore, to compete with wireline networks, a WLL network must deliver a better grade of service than a mobile network, while supporting more traffic per subscriber. This results in several practical problems.

 First, a detailed traffic model is necessary to ensure that radio spectrum is available to carry the expected traffic. Then, in order to prevent airside call congestion and too many customers being allocated to busy sites, restrictions at the point of sale may have to be imposed. Which means that highly accurate coverage predictions must be available.

 Site-acquisition becomes more difficult, since the grade of service target implies 24-hour maintenance access. Furthermore, the need for long-term leases, driven by the high costs associated with subscriber antenna re-alignment, plus the fact that most WLL sites will need to be in built-up areas, reduces the number of available sites.

 A high-capacity diversely-routed backhaul network was needed to take traffic from the radio transmitter sites back to Ionica's switch sites. The company's initial focus on the local access network meant it had to fill gaps in its backhaul network with lines leased, expensively, from BT.

 In isolation none of these difficulties were insurmountable. Indeed, by the time Ionica went into administration most had been solved, at least in principle. For example:

      Analysis of Ionica's customer base had improved the knowledge of the target market, allowing new network capacity to be targeted in a highly effective manner. 

      Network-planning software had been developed, providing full support for a WLL network. 

      The process of acquiring and building base sites was about to be improved. 

      Potential third-party suppliers of backhaul transmission had been identified, resulting in substantial reductions in cost, and greatly increased rollout flexibility.


 Had these solutions been identified earlier, both the coverage and capacity of Ionica's network might have been much improved.

 So the tactics for network rollout were faulty, but the corporate vision was even more seriously flawed. Ionica's strategy was to use WLL technology to become a new national operator, aiming, with 80% coverage of the UK, to provide close to universal service.

 Accordingly, Ionica embarked on an expensive programme of regional TV advertising. The rollout strategy was driven by the need to provide substantial coverage in each TV franchise area where service was offered. The network infrastructure was designed with a national network in mind, resulting in the purchase, for example, of a large number of switches.

 This capital-intensive approach is not wrong, but neither does it allow maximum exploitation of the key competitive advantage of WLL technology - scalability of costs. The priority of any WLL operator with national coverage ambitions should be to ensure that all costs scale with customer take-up, using better targeted marketing and operational outsourcing. This would, in turn, allow a more gradual network rollout, initially focusing on densely populated urban areas, where take-up is expected to be high, and using this revenue to fund future growth.

 The photo on the previous page is a still from the 1947 film, A Matter of Life and Death. In it, a World War II pilot hovers between the two states while his fate is determined by greater powers. There is a certain parallel with Ionica's brief life - after a brave launch it existed in limbo until market forces put it out of its misery. It wasn't the first company this has happened to, and it certainly won't be the last.

 Caroline Wallace was Ionica's network planning manager. Dr Stephen Unger was Ionica's switch planning manager. Contact: cwallace@ scigen.co.uk and sunger@scigen.co.uk.


Peter Ecclesine, Technology Analyst, Biz Dev  
MS SJ-J4 255 West Tasman Dr, San Jose, CA 95134-1706 
Ph 408/527-0815, FAX 408/526-7864 
"Without Prejudice" U.C.C. 1-207. "Time doesn't fool around"