The wireless industry is reaping another bountiful harvest. But digging into the subscriber basket can yield a surprising amount of chaff.
Service providers like
have roared to better-than-expected results in 2005, as new customers have stampeded into prepaid packages, cheaper rollover minutes and multiline family plans.
But lumped in with all the pretty numbers, skeptics say, are some practices -- ranging from tallying dead accounts to double-counting users -- that can muddy the telcos' true subscriber growth and user loyalty performance. And while Wall Street loves a technology growth story, those who have predicted a slowdown wonder how long the industry can keep its foot on the gas.
"The growth has been surprising," says Roger Entner, an analyst with Ovum, a tech strategy shop in Boston. "I've been trying to come up with good reasons to explain it before I think of nefarious reasons."
It could be that the U.S. market has a much greater appetite for mobile phone service than conventional wisdom would dictate. In other countries, as cell-phone use approached 70% of the population, the pace of new subscriber additions started to taper off.
But in the first quarter all the big players, ranging from Verizon and Cingular to Sprint partner
, posted stronger-than-expected numbers. That has some people wondering whether the U.S. market is really different -- or whether those numbers are really what they seem.
"This quarter is when we expected to see some slowing," says Yankee Group analyst Marina Amoroso. "It was right around now when we saw an end to strong net adds, but these numbers are challenging that."
One obvious explanation for the recent growth is the surge in new lines on family plans and the success of youth-aimed prepaid services. Both approaches help telcos reach people outside their main customer pools.
But raising some doubt about the wireless companies' unflagging success is the fact that all the key numbers are tallied by the telcos themselves. There is no outside verification or auditing process to bless the findings.
Also, industry consolidation has helped make the numbers even more suspect in some eyes. The merger of Cingular and AT&T Wireless, for example, has added a layer of confusion to an already opaque accounting situation.
But it's not always clear whether the creep of dirty numbers is a function of abuse or neglect.
One person familiar with Cingular says the company adopted a policy after the merger, in at least one region, of counting AT&T Wireless customers that switched to Cingular as new subscribers. Cingular strongly denies that claim and says AT&T Wireless customers who switch to Cingular plans are called migrations, not new customers, and they are charged an $18 upgrade fee.
And though Cingular claims it is innocent of subscriber-counting abuse, the company did recently have to repair some of the problems it may have inherited from AT&T Wireless. Cingular recently restated its fourth-quarter gross subscriber additions, cutting about 200,000 customers from its original 5.7 million figure. That may seem insignificant, but the move helped the company trim its monthly defection rate to a more respectable 2.4% from 2.6%.
Cingular blamed a lot of the difference on "conformity issues" between how it defines a customer and how AT&T Wireless did.
For example, Cingular found that 81,000 AT&T Wireless accounts were nonexistent. About 65,000 of those so-called customers were actually expired prepaid accounts that were never taken off the subscriber rolls. Cingular also cleaned out 12,000 AT&T Wireless customers that the company said had two lines.
Ovum's Entner isn't surprised by the disparity. In one particularly illuminating industry report he put together in 2003 for Yankee Group, Entner found that 10 wireless telcos calculated a seemingly standard and clearly important number -- average revenue per user, or ARPU -- 10 different ways.
On subscriber growth and customer retention, Entner says there are "three ways to make the numbers pretty."
The first is the most common, promotions. Telcos typically dangle free phones to lure customers and pump up growth. Second, carriers can lower credit qualifications that would normally screen out potential deadbeats. And third, companies can delay account closings to protect churn rates.
There are no hard and fast rules about when a company has to close the account of a nonpaying customer. Analysts say some companies will leave the account open until the end of the billing period and others may leave it open for three months, on the outside chance that the customer may come back.
In these instances, the subscriber is counted twice, once as a new customer with a new carrier and again as an existing customer at the previous carrier.
Cingular and Verizon Wireless dispute the notion that there is any wiggle room and say they immediately close the customer's account when service is canceled.
The policy is less clear among the resellers for Cingular and Verizon Wireless. Typically resellers -- the name given any wireless outlet not controlled directly by the parent telco -- are given an allocation of phone numbers and told to report the net gain or loss of customers.
Analysts say a lot of the subscriber accounting is left to the discretion of the resellers. Normally these include a lot of prepaid accounts that are on a month-by-month status, making them particularly difficult to track.
Industry insiders and analysts also say there is not a lot of vigilance when it comes to cleaning out dead accounts. One reason is that coming clean on customer defections could damage churn numbers, say industry watchers.
Churn is calculated by dividing the number of disconnects by the total subscriber count. If the deadwood were culled from the total customer count, the smaller roster would make defections proportionally larger.
The industry averaged about 2% monthly churn in the first quarter, and any company deviating significantly from that number could end up being punished by investors. As we have seen, Wall Street might love a growth story, but it has little patience for deception and laggards.