Provide decision makers with the right data at the right time, says Julian Hebden.
The world is swimming in data, and each passing year only adds to the deluge.
All this data can facilitate decision-making — but can create its own problems if not managed carefully.
Sample and save 50% on a yearly subscription.
Offer ends 08/12/2020.
“We’re really not short of data,” Julian Hebden, Victoria’s chief data officer, told a recent YIPAA masterclass.
Decision-making with too much information can be an experience akin to drinking from a firehose — and dashboards, with their unfiltered numbers, often suffer from the problem of too much information.
“I hate dashboards”, he says.
“The reason I dislike dashboards is because for me they don’t guide the users enough, or bad dashboards don’t anyway. Actually they don’t help you make decisions, they just give you more and more data to look at.”
This is not a new problem — pilots have come up against similar issues.
“In the old days in aircraft they used to have loads and loads of dials and a big dashboard,” Hebden explains.
“Now it’s getting narrower and narrower. It’s one screen, and the systems on board the aircraft are feeding the information that the pilot needs at that point in time to make the right choices.”
It’s about getting the right information at the right moment.
“I think we need to adopt that kind of approach to the way we surface data to the decision makers in our organisations.”
A different perspective
One of the valuable things about data is its ability to provide a new way of looking at a problem.
“I think where you go wrong with data is when you try to torture the data to say what you want it to say,” he says.
“For me, the data brings a different perspective on things — challenging your own assumptions.”
Artificial intelligence will increasingly be a useful tool to strengthen decision-making by providing a counterpoint to professional, human judgement, Hebden believes.
“AI it isn’t necessarily right, but it’s two perspectives you can compare to each other. Here’s my world experience as a public servant, and this is what the AI is bringing. Are they the same? If they are, it’s probably okay. If they’re not, let’s investigate why not.”
It’s important to acknowledge data is prone to flaws, though. Understanding how the data may be skewed is “an incredibly valuable asset”, he argues — though the people who have spent years working with the dataset and know its limitations aren’t always highly valued.
“If you apply your algorithms without understanding the data may be skewed — in many dimensions — then you run the risk of getting some bizarre results.”
Blood donors and traffic
For all the potential pitfalls, smart use of data can be incredibly useful, says Helena Hamilton, analytics and cognitive consulting partner at Deloitte.
The Red Cross Blood Service provides a good example.
“The biggest business question is how to manage the supply and demand of blood products,” she explains.
“If they get the supply right, they’re saving human lives, if they get it wrong, they’re wasting blood products that are really quite hard to come by.”
The challenge isn’t so much around finding donors — there are plenty of those.
“The curveball is people like myself might wake up in the morning and the weather is nasty and there’s a crash on the Monash [freeway] — pre-COVID — and you decide to make a last-minute cancellation of that donor appointment, which then disrupts that fine balance of supply and demand for the blood product,” Hamilton says.
“One fantastic experiment the blood service ran was to tap into external data sources like Bureau of Meteorology forecast data for bad weather, City of Melbourne open data for events, which is a lead indicator for congestion, as well as Google traffic data, and internal data sources that point to the likelihood of an individual donor cancelling based on past behaviours.
“By combining those datasets, and it was just in a sandbox type environment, they were able to predict with a degree of accuracy the likelihood of someone cancelling their appointment well ahead of that person even waking up in the morning and thinking about cancelling their appointment.
“The really exciting part of that particular data experiment was they were able to take it a step further to prescriptive analytics, where they didn’t stop with that prediction of who was going to cancel, but then routed a text message to an alternative donor with a matching blood type asking that they bring their appointment forward to the date they were anticipating having a gap.”
Data can also be a useful tool for pushing back on poor decisions made on “gut feel”, she says. Public servants who want to ensure their work becomes more evidence-based should “be prepared to challenge the decisions and the data that’s being presented”.
Sharing data is often culturally difficult in the public service, too.
“There is a culture of risk aversion, a historical culture that really is quite embedded,” says Hamilton.
She gave an example from her work with a federal government client.
“There was discussion around some data that was going to be released onto an open data platform. One of the people put forward — and I think this really shows both the culture and the vulnerability people feel when it comes to exposing data they are custodians of — the comment that the person made was, ‘we can’t release that data, because what if it doesn’t support the policy position we just made?'”
“That’s disturbing in one sense, but in the other sense it is a really good representation of [the culture] that exists that makes it really challenging to share and release data.”
Subscribe today and save $220 on an annual subscription
Because we are reader funded, we’d love you to join Mandarin Premium. Without your support, we simply can’t do what we do. And we’re looking forward to doing a whole lot more in 2021.
If you subscribe now, you can save 50% ($220) on an annual subscription*. Just enter promo code PREMIUM50 when you subscribe.
*Offer ends 08/12/2020.