Hotspot aggregators, such as iPass, Boingo and GRIC (which has renamed itself GoRemote) make a lot of noise about how many locations they cover - and rightly so. Aggregators allow users to connect to different providers' hotspots, through a single contract, so users get a better choice of places to connect (see our feature on the development of roaming services).
Corporate customers will choose a service at least partially on how many places they can get access. Unfortunately, in at least one case, the way those figures are calculated appear to be deliberately misleading.

There are three leading Wi-Fi aggregators:

  • iPass, which started as a dial-up aggregator and added Wi-Fi to its arsenal (see our recent interview with iPass)
  • GoRemote which recently changed its name from GRIC and emerged from a similar background
  • Boingo, a recently-arrived Wi-Fi only aggregator.

iPass claims to have almost 10,000 hotspots, while GoRemote claims 7,800. As the new kid on the block, Boingo has 3,300. Each one also talks about how many are in the pipeline - for instance Boingo reckons it has another 6,000 locations coming up.

But the tricky question is: what exactly do they mean by a "hotspot"?

Every hotel room is a hotspot?
Under close examination, the hotspot aggregators' listings turn out to include duplicate entries, in some cases counting every single access point at a certain location. The problem has been highlighted by writer Glenn Fleishman of Wi-Fi Networking News, who used the providers' client software to download a list of locations and compare the real figures.

The worst offender in inflated numbers appears to be GoRemote. In New York, for example, although GoRemote claims to have 61 locations, Fleishman found that 37 of those referred to individual access points at one location - the Warwick Hotel, which provides Wi-Fi coverage to its rooms. Take out the duplications, and GoRemote has 21 unique locations.

This would take it below Boingo in that city, whose 62 claimed New York locations all turn out to be unique, according to Fleishman. Meanwhile, iPass comes out as a leader there, with 258 New york database entries, which refer to 247 locations.

Elsewhere, the story is very similar. In Seattle, Boingo's 26 locations are unique, iPass's 160 entries fall back to 149, while GoRemote's 35 entries become a meagre seven locations on closer examination. "Consistently, Boingo had the most conservative, unique count; iPass had an overlap of the same address of about two to three percent for non-airport locations; and GRIC [GoRemote] listed every access point in a given venue, floor by floor and concourse by concourse," says Fleishman.
"GRIC doesn’t appear to be differentiating in their marketing between the necessary additional entries required for roaming across a venue, and the unique number of locations that a purchasing decision might be made on."

Explain yourselves!
GoRemote is keeping very quiet about the criticism so far. It has not responded to our calls yet. More crucially, the company did not respond to Fleishman, over a period of weeks.

iPass, however, counts itself pleased with the results, and has explanations for the duplications in its figures.

"GoRemote say their network is just as big as iPass and, in some countries, bigger," said John Sidline, iPass's director of corporate communications. "We are gratified it has come to light. We take a great deal of pride that we are building the world's largest WiFi network and that we are accurately communicating our strengths to prospects. It is frustrating when we come up against a competitor that goes out of its way to mislead."

The databases are clearly visible in the client software, he pointed out, so any user can make the comparison, without taking out a contract (or even owning a Wi-Fi laptop).

The anomolies in the iPass data come from two sources he says. Firstly, the directory is pulled together from data provided by service providers, whose databases themselves may contain errors or different spellings.

The residual discrepancies are small he says: "They give a two percent overlap." While third-party access point directories, such as that provided by Jiwire can take out duplications in their combined listing, it would be futile to do so in the iPass client because the underlying database is regularly refreshed with the service providers' listings, which would restore the error.

An airport IS a multiple location?
Secondly, airports are handled differently. Many of the iPass duplicates are at airports and Sidline reckons they are a special case. Fleishman found that the Seattle-Tacoma International Airport (Seatac) was listed as a single location by Boingo; GoRemote pumped this up to 26 locations, while iPass listed it as four locations.

"Airports are a unique environment," said Sidline. Some have coverage across all concourses, while others have partial coverage, for instance with Wi-Fi only available at the international terminal, or in individual airline lounges. To simply say that there is Wi-Fi at the airport does not give the user enough information to know whether he can actually use that hotspot, he said: "It would be no use saying there is Wi-Fi at an airport, if it is only available at the Swiss Air lounge. Airports are the only place where we count multiple locations.

Boingo's record is impressive, but Sidline reckons it might have something to do with the clean slate the company has started off with, which allowed it to put a sophisticated database in its client. "Some of our architecture was designed when we did dial-up aggregation," said Sidline. "It was designed to make it easy to support customers."

Whatever the outcome, let's hope this clears the air a bit. "When we try to interpret the scale of this industry, having an accurate count of unique numbers offers a better gauge of the relative scale of aggregated network and a better tool for customers making decisions," says Fleishman.