Stats NZ to tout $2m start-up unit
Agency offshoot’s first product taps into anonymised location data from mobile phone towers. David Williams reports.
In February last year, as access codes were being sent out for the looming census, Statistics Minister James Shaw was briefed on the emergence of Stats NZ’s commercial arm, Data Ventures.
The website for the nascent agency unit had gone live and it had “launched” through a story in The National Business Review, “with the aim to create greater awareness of this new venture”.
“We are working towards building a pipeline workflow, recruiting for a customer advisory group (which Data Ventures will use to validate ventures) and starting the first venture by 30 April.”
That turned out to be overly ambitious.
The business unit, which has eight staff, was created to be a data “broker”. It doesn’t sell Stats NZ data – its job is to find new sources of data, and revenue, through partnerships with the private and public organisations. Official information shows that it initially struggled to use Stats NZ staff and missed its revenue and product development targets.
Almost two years into the experiment, it has cost the taxpayer more than $2 million ($500,000 under its two-year budget), conducted one pilot project, attracted few customers, and earned an undisclosed amount of revenue.
Minister Shaw remains upbeat and excited about its prospects. “I’m really pleased that Stats has taken it upon itself to set that team up and to see what opportunities can be created through the use of data or the position that Stats has in the country,” he tells Newsroom in a phone interview.
“I frankly think this is a pretty controlled experiment. We’re literally talking about a handful of people, and stepping through it with all of the normal safeguards that you would expect from a public sector organisation but with a much more entrepreneurial mindset.”
Data Ventures is about to step out of the shadows.
An official launch is planned in the coming weeks to explain the Stats NZ unit’s first product. Population Density, as it’s known, is a pilot project that provides a suburb-by-suburb, hourly snapshot of the population, using anonymised location data from Spark and Vodafone NZ’s mobile phone towers – an innovative idea that sparked controversy in June.
There’s cautious optimism the business unit can deliver, and the concept might be rolled out more widely.
BusinessNZ chief executive Kirk Hope says in an emailed statement there is scope for a number of state agencies to generate revenue from their focus of work, and provide innovative products and services.
Richard Llewellyn, head of external affairs for Vodafone NZ*, says the pilot has only been going for a matter of months but “it’s looking pretty exciting from our perspective.”
Tourism Industry Aotearoa chief executive Chris Roberts has been in discussions with Data Ventures over the last year or so. “This is the sort of initiative that Government departments should be taking – devoting a small dedicated resource to look at innovation and new ways of doing things.”
The question is, after the census debacle, is Stats NZ capable of pulling it off?
“It is an experiment to see whether we can innovate in a commercial way within a Government department.” – Liz MacPherson
Data Ventures was the brainchild of outgoing Stats NZ boss Liz MacPherson.
In August last year, as guests picked at food and drank mocktails at Stats NZ’s Wellington office, the chief statistician said she hoped the business unit “feels like a start-up”.
According to the presentation that evening, its operating model might extend to sharing risks (and revenue) in joint-ventures. MacPherson said: “It is an experiment to see whether we can innovate in a commercial way within a Government department.”
Reports written to Stats NZ management by Data Ventures executive director Drew Broadley, and released to Newsroom under the Official Information Act, show that “feel” wished by MacPherson was paired with the real-life headaches you expect in start-ups – including not meeting targets, and product development delays.
(Shaw, a former management consultant, says: “My experience was that it takes a bit of time to work out what your business model is and what your value proposition is, and my sense is that they’ve been through a fairly typical start-up cycle.”)
Data Ventures missed its first target date – September last year – to generate revenue, and the Population Density pilot took much longer than expected to get started. It was a struggle to sign up customers. Broadley missed his goal of signing up five customers by last Christmas.
“Things are taking longer than expected with both data providers and customers,” his briefing in February of this year states. “We were obviously hoping to have a handful of customers – this again is taking longer, but we are making progress. We acknowledge the pressures to really start delivering and we’re confident of having signed customers soon.”
The reports are peppered with potential sales leads. Broadley confirms only one customer – the Business Ministry, MBIE. According to his reports, other nibbles have come from the likes of Treasury, the Department of Conservation, Sport NZ, Ministry of Education, NZ Transport Agency, the Electoral Commission, and Tourism Industry Aotearoa.
In April, a restructure within the population team and “priorities of the census”, delayed the building of its first model. There was talk of having to “get around the limits and restraints we’ve had accessing Stats NZ resource”.
A month later, a frustrated Broadley pushed to have Stats NZ people seconded to Data Ventures. (It now has two seconded staff from elsewhere in Stats NZ.) “Sales progress has slowed in terms of closing any deals,” he noted, adding it was about to advertise a sales position.
Broadley tells Newsroom he doesn’t think if his unit sat outside of Stats NZ it could have gone any faster as the concept had to be thoroughly tested and validated, and scrutinised by its board, which has independent members. It deliberately pulled back from an idea of tracking people’s travel patterns, in case privacy concerns were raised.
“It always takes a while to build a new product and build a new service, but I feel very comfortable with the amount of spend,” Broadley says. “You could say there have been things going on around the edges that can create some challenges, but all in all I feel very positive.”
From stats to data
The world of statistics is morphing into the world of data. Companies are scrambling to be the first to spot patterns in data – and make money from the results. But given the misuse of mobile device data by other Government statistics agencies, including the Australian Bureau of Statistics, and privacy concerns raised in the wake of the Cambridge Analytica scandal, Stats NZ is wise to move cautiously.
Population Density has received ticks from the Privacy Commissioner and the Government’s data ethics panel, led by the Prime Minister’s chief science advisor, Juliet Gerrard. That’s the gold standard, Shaw says.
The pilot’s package of data – the initial set is from February 2018 to February of this year – arrived from the telcos anonymised and aggregated. Data Ventures doesn’t get access to individual records.
The telco feed is then supplemented by other datasets and modelled by Stat NZ experts. (Gaps in the data include areas of low mobile phone tower coverage, devices that are indoors and behind a number of walls, and people carrying multiple devices.)
Between censuses, accurate population data is hard to come by. That can create problems for the likes of Government agencies, Crown entities, and local authorities. Potential uses for Population Density’s count indicator include planning for council infrastructure, estimating tourists peaks and troughs, and the patronage of national parks. Ministry of Health could use it for pandemic modelling, Broadley says, while Civil Defence could apply it to emergency planning.
(Data Ventures’ initial concept was for the mobile data to be used to track movement patterns, which, Broadley confirms, is something potential customers still say they’d like. But that might take the business unit beyond its “public good” mandate. “Right now there’s a lot of work before we get to that.”)
The prospects for Population Density seem good and customers are apparently lining up. Can Data Ventures commercialise its data? “I guess proof is in the pudding,” Broadley says. “I feel this is an opportunity for Stats to demonstrate that it does have the goods.”
Dr Jian Yang, the National Party’s statistics spokesman, says via email Data Ventures’ business model is a good one. “I haven’t had an in-depth briefing about the company but I hope the Minister has asked more questions about this and has received all of the assurances he needs.”
Shaw retorts with a dig at Yang’s boss, Simon Bridges, saying that in the wake of the census debacle he hasn’t seen any credible third-party commentator undermine the credibility of Stats NZ as a whole. “I have complete confidence in the capability of Stats NZ to do the job that it has always done for us as a country.”
Broadley says he views his budget as a loan, and his intent is for the business unit to be self-sustaining – although there’s no set time by which that should be achieved.
Most people would think start-up companies, including those involving selling data, will fail more often than not. So is taxpayer money at risk?
Minister Shaw says: “I think you have to say, if you’re going to get innovation in the public sector and if you’re going to figure out new ways that the public sector can add value to the lives of ordinary New Zealanders and businesses, then you have to be willing to experiment. And in being willing to experiment you have to have something at risk. You have to assume that not everything is going to work out.”
Data Ventures is only a small chunk of Stats NZ, he assures, and he’s happy with what he’s seen to date. “My view is let’s let that run and see how it pans out.”
Stats NZ has made a commercial decision to stop providing a commercial accommodation survey. That leaves a gap for tourism organisations, like those represented by Roberts, the CEO of Tourism Industry Aotearoa.
Roberts says Population Density contains “quite powerful information”.
“It looks really interesting. Just how it can be applied usefully and commercially packaged up for the industry is the next step, really.”
* Vodafone Business is one of Newsroom’s commercial partners