Speaking Engagements & Private Workshops - Get Dean Bubley to present or chair your event

Need an experienced, provocative & influential telecoms keynote speaker, moderator/chair or workshop facilitator?
To discuss Dean Bubley's appearance at a specific event, contact information AT disruptive-analysis DOT com

Monday, February 19, 2007

Thought for the day – is reported mobile data use real?

I regularly see statistics around mobile data that seem to indicate decent levels of market growth. I read recently in a magazine, for example, that Orange UK is doing around 100,000 music downloads per month, or about 3k per day. I don’t want to pick on Orange specifically, but this is a good case study.

I’m wondering if all of these are actually “real punters” paying for music on their handsets. So, for example, can Orange’s statistical reporting discriminate between “proper” use, versus demos given in shops? Maybe there’s a separate APN for retailers’ SIMs, or maybe not. Or maybe there’s just internal billing reconciliation between the various divisions of the operator and its channel partners. Orange has 280 of its own shops in the UK, plus countless other retailers stock its wares, so it doesn't take an awful lot of demos-per-salesperson-per-day to rack up an appreciable fraction of 3000.

And what about testing? I’ve seen a number of presentations around the concept of “user experience monitoring”. Basically the idea is to stick probes in the network, or drive a bunch of phones around in a van, doing user-type stuff like sending MMS or downloading music. Now imagine a network with 10,000 cells, and you want to to do ongoing quality testing with various types of phones, in various network conditions. How many tests might you make per cell, per month? 1? 10? 100?

I’d guess a van with 10 phones in it could visit 100 cells a day, doing a variety of quality-of-experience tests. A fleet of 10 such vans seems reasonable for a decent-sized operator in a medium-sized country. So, maybe the operator conducts 10,000 tests of any given service per day. Or 300,000 per month.

And, of course, both distribution & testing could be outsourced, and any costs effectively just billed back as part of the contract’s expenses terms, making it even more difficult to separate out from real-world commercial use.

And lastly... how much reported data usage isn't from real Joe-Bloggs-in-the-Street, but from people like us in the industry, trying stuff out to see if it works, using it during meetings with advertisers and content owners, checking up on competitors, showing it to the world at Barcelona or wherever....

I could well be overlooking something very obvious here… but equally, it could be that the split of customer vs internal-ops usage of services is misreported by some operators.

No comments: