'

Here's how NBN Co rolled out FttP better in Melton: Part 1

NBN Co reacted hysterically to the publication of an internal trial review that found that FttP could be rolled out 61 percent faster and 50 percent cheaper than in previous rollouts. Here's what the company's Melton deployment trial taught it about managing contractors better.

Depending on how closely you follow National Broadband Network (NBN) news, you may or may not have noticed NBN Co's media meltdown over the past week. Either way, it was a doozy — and it came about mainly because the company's own internal work demonstrated that it is in fact capable of rolling out fibre-to-the-premises (FttP) NBN connections faster and more cheaply than the current government wants to admit.

Melton aerial 1: CC BY-SA 3.0 Graeme Bartlett
Battlefield Melton: Successful trials of FttP optimisations raised hackles at NBN Co. Image: CC BY-SA 3.0 Graeme Bartlett

Faster, cheaper FttP, after all, would completely undermine the claimed cost benefit of adopting the Malcolm Turnbull Melange (MTM) alternative strategy, which has focused on replacing the construction of a modern and future-proof network with a dog's breakfast of technologies that has at its core the renationalisation of the same Telstra copper access network (CAN) that the government sold to the private market a decade ago.

Charged as it is with delivering this MTM model at all costs, NBN Co was less than enthused for people to know the results of an internal trial of new deployment techniques, as it demonstrated, in that particular 2484-premises fibre-servicing area module (FSAM) at least, that fibre could be rolled right to the premises 61 percent faster and 50 percent cheaper than comparable earlier rollouts in Ballarat, just a bit further down the Western Freeway.

The changes introduced in Melton were largely the same as those floated to the former NBN Co board just before it, like the design recommendations that it was told would cut AU$4.5 billion from the cost of the FttP project. However, all were dismissed by on-message incoming Communications Minister Malcolm Turnbull. Turns out that Labor wasn't as bad at rolling out FttP as Turnbull needed everyone to believe.

The point here is that by and large, the changes implemented in Melton were far from magical; no, they were just the kinds of things that you figure out as you review your progress to date in any large project, then introduce changes to make things work better.

The over-the-top nature of NBN Co's response beggared belief, raising eyebrows across the telco community and putting NBN Co in the position of publicly flogging one of its well-meaning employees in the media.

One major element of the new design related to the communications between NBN Co's subcontractors for the trial — Celemetrix, Comms Connect, and Linktech — which were engaged with "a uniform schedule of rates" to wire up what the document called "a relatively large number of 5- and 1-acre lots" in the area, described as being "a rocky area" predominantly composed of single-dwelling units.

Recognising the significant problems that NBN Co has had in keeping contractors supplied with a steady stream of work, the company trialled a deployment management methodology called Render, which the document says was introduced "following a technical proof of concept".

Render — which I cannot find formally documented anywhere online, and which NBN Co naturally refuses to tell me any more about — is described as having enabled 10 steps to improve contractor communications. These include:

  1. Automatic creation of the work scope, including asset data and the construction sequence

  2. Allocation of work to contractors in advance, using this work scope information

  3. Automatic creation of the work schedule, based on committed allocated resources

  4. Releasing four days' worth of work to crews each day, topping up their pipelines and ensuring "continuity and visibility of work"

  5. Only releasing jobs that were ready to be done — in other words, those for which all preceding tasks had been completed

  6. Daily review of completed work, with remaining work rescheduled

  7. Capture of as-built information on the day it was completed

  8. Automatically updating progress status and scope, resource, asset, location, cost, and program data centrally, so it could be viewed on Google Maps

  9. Building the network "as it was designed" to eliminate "'air gaps' and 'red-line markups'"

  10. Sequencing and managing build drops concurrently with the local network/distribution network (LN/DN) rollout.

Render used a task-based approach that staff took a little while to get accustomed to. However, while there is "still some work to be done on the functionality of the system ... it performed well".

If this all sounds logical, you're right. The procedural changes that NBN Co introduced in Melton mainly consolidated previously separate work steps, took a more proactive approach to identifying and allocating outstanding work, and improved visibility of every aspect of the rollout so that problems could be remediated quickly and effectively.

The changes that NBN Co introduced in Melton consolidated previously separate work steps, took a more proactive approach to identifying and allocating outstanding work, and improved visibility of the rollout so problems could be remediated quickly and effectively. This is hardly controversial stuff.

This is hardly controversial stuff, but it seems to have worked. According to the document, these improvements reduced administration effort and simplified the definition of work scope; improved visibility of the entire rollout; allowed variations to be measured at the task, asset, and resource level; allowed NBN Co to commit work volumes to contractors at the task level; automated and improved progress reporting; expedited "jeopardy management" relating to unexpected issues at any level; and integrated health, safety, environmental, and quality concerns at the task, resource, and asset level.

The NBN Co report also flagged improved contractor productivity, automated optimisation of daily schedules, better forecasting of future work, better utilisation of rollout crews, and the availability of as-built data on a daily basis.

Riveting stuff, and it may be common sense for anybody involved in construction. But there's more: A few technological changes, which I will outline tomorrow, supported the previous NBN Co board's contention from almost a year ago that a modified technological approach could indeed speed up and reduce the cost of the FttP rollout.

What do you think? Was NBN Co's reaction appropriate? Should we actually ignore the suggestion that FttP can be rolled out faster and more cheaply? Or do you agree with NBN Co's warning about broadly extrapolating the benefits seen in Melton?

After publication of this story, ZDNet received the following response from Karina Keisler, NBN Co executive general manager of corporate affairs, which is reproduced in full.

The article is inaccurate, because the conclusions it draws are not supported by the facts.

The work under way in Melton delivered no such conclusions. Further, the scope of the work being done in Melton has been overstated. The PowerPoint used as the basis of the article has not been endorsed by management, due to a number of shortfalls in the methodology and metrics.

Importantly, NBN Co continues to identify efficiencies in our rollout processes across the country, and across all technologies. We have an obligation to the taxpayers who are footing the bill, and we know an efficient rollout is key to our company successfully reaching its target of 8 million customers connected by 2020.

Update: NBN Co comments at end of article.