NBN study: the analyst view

Under the hood of El Reg's crowd-funded study

Pledges have reached almost $7,000 for The Register's crowd-funded NBN study, so we thought it's a good idea to let analysts explain how they'd approach the project. Today, IBRS analyst Guy Cranswick provides insights into his view of how to conduct the study.

Of all the vexed issues surrounding the NBN debate, an assessment of the technology is the easiest, Cranswick told The Register. It boils down to determining which network will be the most durable.

There, he said, the analyst is dealing with “products and accounting practices … when you assess those, you only really need to explain your methodologies and the principles you applied.”

Durability, however, doesn't point to the future requirements. As noted by Telstra CEO David Thodey recently, a well-maintained copper network could last a hundred years. Trying to understand future requirements is where the analyst has to bring skills and experience to bear.

The NBN is a consumer play, Cranswick pointed out. “It's about wider distribution [of broadband] and better quality of service for all.”

However, that poses a challenge: even well-conducted consumer surveys can founder in trying to assess the uptake of future applications. “People give lukewarm responses to things they haven't seen yet,” he told The Register.

It's when the product, service or application is in front of the user that they begin forming their attitudes: “When they experiment with something, when they see how others use it, they jump on board fairly quickly.”

Trying to assess the future is “where you ground your credibility,” he added.

We can't predict all of tomorrow's services, but some things remain constant, he said: price factors, the value proposition, all affect the kinds of services that people will use and pay money for in large numbers.

That's what makes the international experience important, whether the examples are drawn from the US, the UK, Germany or South Korea, he said.

Don't fall into advocacy

There's always the risk that people taking an advocacy position on one side or the other of a debate will attack the validity of the examples an analyst uses, he said. Transparency is the key: differentiating between those parts of the analysis that are known, those parts that are models, and making sound judgements about which data is applicable to Australia's experience.

“It's not something that will be bolted down – we look at scenarios that look realistic from point of view of 2013. Even by 2015, it will more than likely change, because there will be different factors that will have greater or lesser strength,” he said. “An analyst is compelled to use the standard set of research and statistical techniques in answering these questions. That involves gathering the relevant data sets, looking to overseas instances for analogies or patterns if the data is not available here and using evidence from the past as the best set of guides.”

Data interpretation matters

The data sources that exist are vast and of high quality, but they need to be handled with care, Cranswick said. A good example is the Cisco Visual Networking Index (VNI).

“The Cisco VNI is really valuable to people like us, that anlayse these things. But it lacks a price tariff: basic finance tells you that you can't test the usage of a product unless you know the price it will be,” he said.

The changing adoption of broadband in Australia in the face of falling prices illustrates this. “The VNI is interesting as an illustration how usage patterns per device, or data type, is growing. That tells us that the area is growing enormously – but we need a context: what if the price is $100 per 100 GB per month, what would that do to customer behaviour?

“If you put in various price impacts, you change it – this has been brought up in the current debate of quotas and bundles and additional services It's known by the DBCDE and the telecoms industry – the way access to services is priced is important.

“We're pleased that Cisco produces the information – but it's only part of what we need to know. An economist then looks at models of usage depends on price and speed, and work out the optimum levels of pricing – that's how you use that information, in my view.”

Discussing prior reports by Deloitte and Allen Consulting, Cranswick explained: “The craft of analysis and use of data will suffice to a large part but the answers require interpretation and scenario analysis. This was seen in two of the better NBN reports in which the analysts framed the replies by saying 'some things cannot be known but it is probable that X may result and this report provides an illustration.' “That is responsible analysis. This highlights the importance of making logically valid inferences, that is to propose future conditions on the basis of propositions that are valid. What we have seen sometimes is that sound rules of logic and scenario analysis have been overlooked to demonstrate or prove a preset advocated position. Such an interested stance is not independent.”

He also emphasised that IBRS' approach will not try to attempt being a “cost-benefit analysis” of the NBN, because such a thing would have low value anyway.

“The problem of doing a CBA of something that isn't built is that you might assess that some inputs may have variable outcomes and different cost influences on the project.

“It's not like choosing one CRM versus another – those are done within one period of time, say two years: we can work on that easily. But that's not the same as something that won't be completed for nine or ten years,” he said.

“A virtue of this project is that it's coming from a technology publication – one of the problems of any analysis is that analysts have tended to deliver summaries and recommendations suitable to the client that's paid for the reports

“What we're doing here is producing reports that can be said to be genuinely independent: that we are serving no particular client or interest group. We just want to explain the position as best we can at this time.”

The study is at Pozible, here. Share the link and help us reach our targets! ®

Sponsored: Network DDoS protection