Getting technology platforms over the line

With Paul Clarke

The New Zealand government rightly puts emphasis on proper business cases to justify major investments. But this doesn’t always work well when realising opportunities presented by technology. By being smart, we can stop due process being a roadblock.

A few years ago, our family decided to buy a sail boat. Or rather we decided to buy a bigger sail boat — the temptations of increasing size being a familiar hazard for yachties. Being compulsively impulse-free, I set about drawing up a list of criteria: small enough to be easily sailed by two; big enough to cross the Cook Strait or take friends out for a potter around Wellington Harbour; sporty enough to race with the local club; but not so sporty it would lean over beyond my wife’s ‘angle of divorce’.

This led to a set of specifications that narrowed things down a lot, and so after much trawling though TradeMe Boats, we set off around the country to view and test-drive suitable models. The time taken felt worthwhile because this was a major investment for the family.

Major investments in the public sector are subject to considerable scrutiny for at least a couple of good reasons. First, and most straightforwardly, it is taxpayers’ money, so you have to do it right. Second, governments care about the full range of outcomes and often there are indirect consequences of the purchase. A force acting against these imperatives is that the starting point for thinking about an investment is often an idea for how to do existing things better rather than other larger potential outcomes. To take the boat analogy, my starting point was simply buying a bigger boat, not how to have fun, meet new people, and get to some nice places.

Out racing with family and friends in the new boat (source: Paul Clarke) 

The New Zealand Treasury’s response to this is the ‘Better Business Case’ process, which includes, among other things: being clear about what you are trying to achieve, spelling out the problem with current arrangements, and what can be a tortuous long- and short-list process. (As an aside, although these are reasonable matters to consider for investments in the public or private sector, the Better Business Case process can sometimes feel constraining if applied inflexibly. At MartinJenkins, we usually find there is reasonable flexibility from Treasury if you make the case for it.) This process can be pretty effective at filtering out what is referred to in the trade as ‘solutions looking for a problem’.

The trouble is that sometimes solutions looking for a problem are a good idea. Technological innovation is a particular case in point. Many technology innovations are platforms or enablers (think electricity distribution, internet, and smartphones) whose uses have developed beyond the wildest speculation of their early proponents. Government is making these kinds of investment all the time — think archive digitisation programmes, making national geospatial information open source, mechanisms to share information between ministries, and mechanisms to allow research on administrative data without infringing privacy. These investments potentially support a wide range of objectives, some of which will only emerge with time.

Faith - based investment

Justifying such investments can be hard work because of the speculative nature of the benefits. Sometimes, the response is to give up on the analysis and invest as an article of faith. An extreme example of this was Australia’s 2009 decision to invest AUD43 billion in the National Broadband Network with the aim of reaching just about every home and business in the country. Given the famous isolation of the outback, this reach was very ambitious. At the time, there was no cost benefit analysis. The high cost was driven by the decision to take fibre to most doorsteps — rather than local nodes. (It was also a very interventionist approach compared to that taken in other jurisdictions.)

It hasn’t gone well. By 2016 it was only capable of servicing 260,000 premises (compared with 7.7 million households). There were only 78,000 subscribers, 77% of whom were choosing service plans with speeds a quarter or less of what was available. Escalating costs led to a decision to back off from the fibre-to-the-doorstep model.

The 1871 Alice Springs Telegraph Office (source: alfotokunst on FeaturePics.com) The Australian National Broadband roll - out was forecast in 2009 to cost AUD43 billion

There is of course plenty of commentary about what went wrong (see for example the Centre for Public Impact’s case study). A decent business case would certainly have helped — but would it have been enough? One of the problems with platform-based technology innovation is that there are lots of dependencies that can be hard to predict: how quickly will internet services requiring a high speed connection grow? how many researchers will have the funding and tools to use a new platform? if we are enabling data sharing, how many participants have the tools to analyse the resulting data and share the results where it is needed?

Business cases can struggle in these situations. But there are at least three approaches that can help, according to circumstances.

The first is to develop a range of scenarios and consider how the investment options play out against them. This will inform a judgment as to whether a precautionary approach is appropriate (as in the case of the National Broadband Network going for the less expensive option and then upgrading it). This works best when government does not have control over the dependencies.

The second is to start not with a business case, but with a wider strategy that covers most, if not all, the dependencies. This works best when government has control over those dependencies. Business cases can follow — but on firmer ground. An admittedly non-technological example of this was the New Zealand Refugee Resettlement Strategy. The core investment at issue was redevelopment of the Refugee Resettlement Centre in Māngere. In easing new refugees into life in New Zealand it acts as a platform: it is a place for a range of service providers, from dentists to housing providers, to make contact with new arrivals. Having a strategy in place meant that when MartinJenkins helped MBIE develop the business case, we were able to provide confidence that the redevelopment was not the answer to ‘what do we do about the rundown state of the centre?’ but ‘how do we best settle refugees given the outcomes we want for them?’.

Starting small

The third approach is to start small, and see what the response is. This can be appropriate when there are a lot of inter-dependencies, or when developing an overall strategy is too hard because there are too many moving parts, or when you’re able to carve out a small demonstration effect. There is of course a tension in that many platforms depend on network effects, so starting large can be of the essence — but not always. An example of starting small is Statistics New Zealand’s development of their Integrated Data Infrastructure. As explained on their website, it holds anonymised linked longitudinal data about people and households from an increasing range of sources and provides a controlled platform for research. The IDI started as a specific data integration project linking two or three data sets, moved through a data integration service, and then through to a cross-agency data sharing platform with frequent increases to the data sets included as opportunity and funding allows.

Starting small is exactly what we did with sail boats. When we bought our first we expected to use it mainly in pottering about Wellington Harbour and trailing around the country for multi-day cruises. It turned out we used it in those ways only a bit, while we used it in ways we hadn’t expected — racing — rather a lot. So when we finally got a bigger one, we used it almost exactly in line with our revised expectations. The surprise was my wife’s willingness to renegotiate the angle of divorce.