Fixing the NBN debacle

So those who follow my various missives on the NBN and its issues will know that while the technology of the rollout will be an issue in the future (probably soon unfortunately), it’s not the big issue right now that’s causing slow internet. The issue right now is the CVC pricing – it’s priced in Mb/s and is essentially the amount of bandwidth allocated to all of a telco’s customers in a geographical area (POI). Due to the (quality/price) race-to-the-bottom we tend to have in telco, the temptation to skimp on buying CVC is just too high and we get the result we have today.

So how do we fix this?

VHA came up with some interesting suggestions but I don’t think they go far enough. Let’s start by looking at why the CVC charge is there in first place.

NBN (then NBNCo) was setup as a Government Business Enterprise (GBE) so that it didn’t appear as an expense on the budget. The rules of a GBE are that it must generate a commercial rate of return for the Government. This puts a floor on the revenue that NBN must generate overall. If they were to just wipe various charges, then NBN would be forced back on to the government budget and taxpayers would pay for it instead (oh the politics). The irony here is that all the discussion about the costs of the NBN have been as if the NBN actually was an expense on the government budget (like say submarines), not as if this was an investment and whether it was a good or bad investment.

There are two main components to NBN’s charge model:

  • The AVC charge which is levied per customer and scales with the peak speed of the connection. A 12/1 is charged at $24, a 100/40 connection $38.
  • The CVC charge which levied per Mb of bandwidth. It starts at $17.50/Mb but since a new discount model was introduced now decreases the more bandwidth the RSP allocates per customer, all the way down to $8/Mb.

When the pricing model was originally conceived, the AVC charge was kept low because NBN didn’t want to penalise the infrequent internet users (some of this thinking is over 10 years old – remember all those arguments about the NBN going to cost too much for the elderly) – so the charge was shifted across to the CVC as a usage charge.

The CVC charge currently accounts for about 35% of NBN’s revenue. So every time they reduce the price it appears they would drop revenue. Or would they?

Imagine you run a high quality hotel charging $400 per night. However due to the price, most nights your hotel is only 30% occupied. The revenue you get from the guests has to cover all the staff, rates, mortgage, power etc. regardless of the number of guests. How can you increase revenue? Drop the price! If you charged $200 per night, you would sell way more rooms – more than the 60% occupancy that it would take to recover the same revenue. Net result: a lower price actually creates more profit.

It’s the same with bandwidth and the CVC. NBN is in the main a construction project with a fixed cost, paid for with money borrowed from the Government which they need to pay interest on regardless of the amount of CVC they sell. So the trick is to make sure the telcos buy more CVC so that NBN’s revenue doesn’t suffer as they reduce the place. We need to break the Mexican standoff with the telcos saying reduce the price, while the NBN is saying buy more CVC, then we’ll reduce the price.

So here’s my fix.

Currently the telcos are allocating less than 1Mb per customer, sometimes much less – 500kb, which means NBN is receiving only between $10 and $20 per customer from the CVC charge. For any speed over 25/5, this needs to be raised to 5Mb per customer – yes that’s excessive, but stay with me. For 25/5 perhaps half that (and let’s kill the 12/1 speed tier). The price of all the AVC’s over 25/5 should go up by $20 but this 5Mb of CVC would be included in the price. For 25/5 it would only be a $10 increase, with 2.5Mb of CVC included. You will note this effectively reduces the price of CVC to just $4 per Mb, but also forces the telcos to actually buy it – thus protecting NBNs revenue model.

So why so much bandwidth? I think it’s crazy that we have created the CVC as the choke point on the bandwidth. Frankly a carrier’s back end connection to the internet should be the choke point, not the CVC. This would encourage carriers to do more “on-net” peering with content providers – just as they do on ADSL. This would push the adoption of more carrier backhaul bandwidth which would again bring down the prices.

Yes we would all end up paying a little more for our internet, but not much more (probably a few dollars a month) but the difference in performance would be absolutely dramatic. It would end all the current issues with the 6pm slow down and then we could get back to focusing on the access technology issues.