One question in the fibre network debate relates to whether a new open access infrastructure will destroy ISPs’ ability to differentiate their services in the last mile. Since everybody will be using the same network, won’t all their access services look the same?

There are two responses to this. The first is that it might be true. The second is that it’s irrelevant. Just how much of the differentiation between today’s infrastructure-based ISPs rests on their last-mile deployments — apart from the difference in geographic reach that ISPs achieve, based on the extent of their DSLAM investment?

I’m not privy to every aspect of the ISP’s business operations (outside the ISPs, nobody is), but a high-level overview should suffice for this discussion.

Today’s infrastructure-based ISP buys access to the copper from Telstra, installs DSLAMs in the Telstra exchange, and buys backhaul from the DSLAM back to its own network. Network access is regulated, which leaves backhaul and the DSLAMs themselves as the points of differentiation.

In the cities, where most competitive DSLAM deployments are concentrated, the backhaul market is contested by a number of fibre owners. Once you’ve got the best dark fibre price available, you can’t differentiate yourself from other ISPs getting the same price.

The same goes for the DSLAMs themselves; ISPs might get some negotiating power over the DSLAM price, and engineers will spend a lot of time and effort assessing the characteristics of different systems, but in the end, there’s only a limited number of vendors and a limited difference between their products.

Such differentiation as exists happens not in the access network, but inside the ISP itself: its skill in managing its network; the way it handles customer support (an expensive part of any ISP’s operations); its success in creating the kinds of bundles (Internet, phone, IP-TV, a cheap Chumby, Internet steak knives and so on) that attract customers, and success in keeping those customers sticky (because churn is another costly piece of the ISP’s operation).

Take a good look at the plans on offer, and it quickly becomes clear that with all the knick-knacks on offer, product differentiation in broadband is mostly illusory. There are huge outliers of services that offer very poor value or brilliant value—and with Telstra's decision to get rid of its more notoriously poor-value caps, the field is narrowing.

Most plans cluster around a very small centre.

Frankly, broadband access is like razor blades—the products are more different in their branding, advertising and promotion than in the products themselves. Where there are real differences between ISPs, they are inside the network where the customers don't notice but the bottom line does.

Both the “real” and the “promotional” differentiators will still exist in an NBN world. The operational aspects that make customers stick to one ISP and churn away from another (some of which can be inferred from a close reading of the TIO’s regular complaint summaries) will still exist, even in a more distant future in which the ISPs have retired their DSLAMs and bought wholesale fibre access from NBN Co.

It may even be that new differentiators emerge.

For example, while the fibre will be 100 Mbps-capable (and upgradable, I should imagine, to higher capacities in the future, if this becomes a requirement), that 100 Mbps can be sliced in lots of different ways. There’s nothing in my reading of the Communications Alliance’s draft reference model to exclude a provider from applying capacity management to the customer connection.

It’s feasible that an ISP, freed from its own direct investment in DSLAMs but with access to better link management than you get from the copper, might find more ways to differentiate its last mile services from other ISPs — even if there’s no longer a difference in its last mile deployment. It’s a subtle difference, but one that might matter as the NBN rolls out.

Richard Chirgwin