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 see recent presentations, and discuss Dean Bubley's appearance at a specific event, click here

Thursday, May 27, 2010

What are the side-effects of speed/vol tiering for mobile broadband?

I'm at the Open Mobile Summit again today. I've still got a load of things to write up from yesterday as well (there are some too-brief comments up on my experimental and much-hated Twitter feed @disruptivedean).

I've just seen an interesting presentation from TeliaSonera, talking about its early deployment of LTE in the Nordic market and its pricing/tiering for mobile broadband.

It's worth noting a few things first - Sweden and Norway are quite "special" markets: affluent people, initial introduction of cheap and "very flatrate" USB modems for PCs in 2007 (€20 for all-you-can eat, I think was stated), lots of fast fixed broadband, and quite a lot of quite cheap spectrum, split between relatively few operators (mostly fixed/mobile integrated).

The speaker highlighted an interesting tiering scheme being introduced in the summer, as it extends its LTE rollout beyond its current (very) early-adopter phase of a "couple of thousand" users.

It is going to charge LTE access ("4G") for 30GB at €36 per month, at "the highest speeds feasible on the network", including dropping down to 3G where necessary.

It will charge lower prices of €4, €23 and €32 for 2GB, 10GB and 20GB at 3G-only speeds, dropping down to 2G when the cap is exceeded. It's not clear if this is a throttled "3G speed on a 4G" network, or whether it just involves an non-LTE modem.

So basically, they are using 3G vs 4G to provide price discrimination on a proxy for speed performance - and then combine it with different caps to enable hybrid speed/volume tiers.

But when asked, they are not doing speed discrimination within 3G (eg 3Mbit/s vs. 7Mbit/s vs. HSPA+ at 21Mbit/s) because it is too difficult to guarantee real-world perception of tiering. It's much easier to discriminate between HSPA at "1-10Mbit/s" vs. LTE at "10+".

This all makes sense - trying to have more than 2 or 3 tiers is difficult even in fixed broadband, where there isn't the variable of changing radio conditions involved.

In future, they're planning to offer "top ups" for your cap, if you run short at (say) week 3 of a month. They will also do zero-rated traffic for specific applications supplied by themselves (eg a deal with Spotify will mean that streamed music may not count towards the quota).

There are prepaid plans for 3G but not LTE at the moment.

I'm curious about a few things, though. The first is what the performance of the LTE network will look like when it starts loading up - it's all very well to claim customer-achievable speeds of 20-40Mbit's when you've got maybe 2000 users and 500 base stations (1500 sectors) - most of the time, any active user is going to be alone or perhaps sharing capacity 1-3 other people. What's going to happen when there's a conference room with 50 people with laptops and dongles, or even just a branch of Starbucks with 5 users?

The other question is around unexpected side-effects of price discrimination.

Ultimately, it sounds like we are getting to a position of differential pricing for what is a (relatively) commodity service. The big question is how "tradeable" that commodity is - and how arbitrage might work.

In prepaid voice markets, it is common for users to have multiple SIMs, and swap them over frequently to take account of pricing oddities. Many have multiple phones, to enable them to always call "on net" and benefit from the lowest prices rather than pay extra for cross-network fees.

It seems likely that while me may see more clever operator pricing, we'll also see more clever user behaviour to circumvent this. Given that data services don't need fixed phone numbers, I see much greater opportunity for dual-SIM, connection-sharing, tethering and MiFi-style products to facilitate arbitrage. There is also a lot of potential for new forms of least-cost routing and collaborative applications that pool different users' allowances.

So for example, how long before we see an application that can suggest:

"If I'm in week 4 of the month, and still have >30% of my quota left, then share or trade that capacity with my Facebook friends automatically, when we're within WiFi range of each other".

I think we're on for some fascinating device/web-based techniques for "gaming" the new data pricing structures.

No comments: