The cultural change at IP Australia making data an asset
- 17 September, 2018 12:52
‘Treat data as an asset’ sounds like a straightforward and sensible maxim for any government agency to follow. It was one published in bold type in IP Australia’s corporate plan from a few years back.
There is a lot of data that comes into the agency; applications for patents, trade marks and design rights are steadily increasing each year. In 2017 the agency received 28,905 patent applications; and a record-breaking 76,594 trade marks and 7708 applications to register designs.
Each application - of which the agency has more than a century's worth - can be broken down into a multitude of data sets, including its place of origin (the majority of patent applications and a quarter of trade mark filings come from overseas), information about the applicant, and the design files and imagery.
There is also an abundance of internal data, from the agency’s various business functions.
Putting the pithy ‘data as an asset’ phrase into practice, then, requires more than simply writing it down. Getting the agency’s 1200 staff to embrace data and appreciate its value has taken true cultural change.
Leading that change for the last two years has been IP’s chief data officer – the agency’s first – Kevin Jeffery.
“You can say you're treating it as an asset but you also have to actually do it,” he told CIO Australia at an Amazon Web Services event in Canberra.
Getting people and processes aligned behind the data asset mantra, means IP Australia can now embark on an ambitious project: becoming a "global leader in IP data provision and analysis".
Changing culture with custodians
One of Jeffery's first moves in changing the culture around data within IP Australia was to appoint one or more of what he calls a 'data custodian' in each and every function.
"The custodians are the ones who are responsible for the data within their business area which owns the data. And I use 'own' carefully because the term carries baggage," Jeffery explains.
"We specify the responsibilities very carefully. They have responsibility for data quality, for data protection, so they have to know their data and who's allowed to use it and when," he adds.
When the new governance model was rolled out, the CDO often heard the same phrase: "The common refrain was 'that's an IT thing'."
"What that means is you can abrogate your responsibilities to IT. But IT don't have responsibility for your data. It's the result of your business process, or you collect it or buy it. You have responsibility for it," Jeffery says.
Attitudes changed after the custodians, who take on the role on top of their other responsibilities, were also required to increase the value of their department's data.
"So that way they can't just lock up their data and say 'alright I'm going to protect the data' they actually have responsibility to create value with it," Jeffery says, "These are things they wouldn't have had to do before."
The custodians are provided a lot of support to meet their new responsibility, "so they're not left out on their own," Jeffery says.
There are regular get-togethers, and an education and training programme "to help those people and bring them up to speed".
The data stewards are also clear on "all the people who will help" Jeffery says. "So you've got IT who look after your data for you physically, you've got the privacy people who will give you advice on security … and the data governance office we'll give you advice."
Executive buy-in means the nominated custodians get allocated time and resources to do their data duties.
How did they take being given the new data responsibilities?
"It's a question that's asked of me a lot. How have areas handled taking on that responsibility? And it doesn't seem as hard looking back on it, compared to the amount of times I get asked," he says.
"I was very clear. I recognised this as a cultural change issue. If data is an asset it takes cultural change to see it as an asset.
"That means every decision you make that impacts on data, you have to recognise that. I explained the outcomes and benefits of that and you don't try and impose massive compliance burdens and so on but just show where there's some responsibility and where you have to be careful," Jeffery adds.
No more two year, $2M tools
IP Australia is not new to sharing data. As part of the act that established the agency, all applications go on public registers. The registers are searchable online and over the past few years the agency has been putting out datasets in bulk on data.gov.au. At last count its IPGOD (IP Government Open Data) dataset was the seventh most downloaded from that site.
"It's really well used," says Jeffery. But there are limitations to its access and use. There's also increasing amounts of internal data, and a number of disparate applications that have been built by analytics teams.
"As chief data officer I want to make sure I can find my data, I can value it, and I can get it shared and used in a better way. I want to get rid of this proliferation, spaghettification of data ... data spaghetti that's been built up over the years," Jeffery says.
"One of the challenges, and this is a reasonably common challenge, is we've built a range of useful applications, both in collecting data better, using specific datasets better, enabling our new data science teams, our reporting teams, or our analysis teams to do better work with data, but what they tend to do is do it in an ad-hoc way," Jeffery explained.
In April, IP Australia contracted Capgemini to develop a cloud-based data management, analysis and reporting platform using leading edge technologies and data management concepts.
The $11million project – dubbed business intelligence and analytics or BI&A – will see Capgemini provision and manage the infrastructure through Amazon Web Services (AWS), while providing operations and support after the capability is implemented.
"One of the benefits of doing it in the cloud is that trusted third parties whether that's other agencies or academics or however we maintain that trust, can come into our platform, use the data, potentially bring their own data in, and do things within the platform we provide," Jeffery says.
"Traditionally you would have to take a copy of data and take it to someone else's location and put it on their computers."
Over the past five months data pipelines have been built to feed into a data lake and lab environments stood up; analytics tools and the third party platform will soon follow.
"What we want to do is get us to a place where it's not two years and two million dollars for every analytics tool we want to build," Jeffery adds.
The promise of a unified analytics platform – which is being delivered using a modified agile methodology - is already proving itself in a proof-of-concept called TM-Link, a single, internationally linked trade mark database.
The tool, developed in collaboration with Swinburne University of Technology, can track where companies are registering trademarks and use the information as an indicator of expansion plans. It will be hosted on BI&A.
"One of the keys to [BI&A] is it's enterprise-wide, we wanted to get away from analytics teams building their own things ... it really is a cultural shift. I think probably what I underestimated there is the number of cultural shifts we're trying to do at once," Jeffery says.
"The cultural shift to make data an asset, the cultural shift to move to the cloud, and the cultural shift to move to that rapid processing and continuous integration that is enabling these sorts of practices that really the analytics people need."
George Nott travelled to Canberra as a guest of AWS.