Volvo analytics chief: 'Teradata is giving value for money, but there's a journey ahead'
"It's about us utilising the data well," says director of business analytics
Big data analytics company Teradata's solutions offer good value for money, providing the data you feed into them is worthwhile, and you're prepared to wait for a data-heavy future.
That's the conclusion of Jan Wassén, director of business analytics at car manufacturer Volvo, as the firm starts to make some serious in-roads into the Internet of Things (IoT), and hopefully get a good ROI on all the money it's been spending on big data for nearly nine years.
Now boasting 80 to 90 per cent of its cars ("with customer permission") connected to share their data not just with Volvo, but also with each other, the company is on the analytics warpath.
"The car detects when the road is slippery, so together with the Swedish road authorities we are working to make it so when a car detects that slippery road it sends the information to the cloud, and an algorithm - providing there is a threshold of [relevant] info - tells the road authority it's slippery," says Wassén, speaking to Computing at the Teradata Universe 2016 conference in Hamburg.
"It then warns other drivers who have this feature that they're coming into this area."
Volvo has been using Teradata for big data since 2007. As well as using the firm's traditional relational databases, autumn 2015 saw Volvo take on Teradata's more advanced big data solution, Aster - although Wassén admits they're "still discussing how to utilize that", with some "new datasets to bring in".
"Web-click information is now being brought in from the [Volvo] website, and that's data we will use in more detail," he says.
In the main, Volvo is still using Teradata for its "original and most advanced" utilisation: handling warranty coverage for cars by using sensors to keep an eye on failure levels of components. This can be anything from whether the brakes are about to fail to whether there's enough washer fluid (which probably isn't a warranty issue, but a nice-to-have anyway).
"We use it to action all kinds of problems before they get too big," says Wassén.
The original reason for choosing Teradata back in 2007 was "performance", he says.
"We were stuck with the warranty use case, and performance was the key question to sort. It's continued to be an important question, and now also in other areas too. We've continued to expand on our warranty journey, as that's the most critical when it comes to response time.
"There's now more analytics coming into it, and what we're looking into now with Think Big (the analytics firm Teradata acquired in 2014) is discussing totally (hardware) agnostic technology - not just talking about appliances but also a total infrastructure."
While insisting he's "not the full IT man", Wassén has some valuable insights into how best to use Teradata's technology.
"What's important is the way you have a structured approach to this, where [IT and analytics] come from the data itself, with methods to ingest and analyse, and have a structured approach. We're trying to go down the same road. We have alignment on our strategies, but it's a question of applying it in all cases as well, and we haven't necessarily been bulletproof on that," he admits.
"From an analytics perspective I'm very much involved now, so there's no shortcuts being taken," Wassén continues.
"All we know is we need to have access to data - it's the raw material of analytics."
Computing asks whether Wassén considers Teradata, which is seen by some as being rather expensive, good value for money in terms of ROI. Predictably - as often with analytics - the response is nuanced.
"It's a combination of us utilising the data as well [as assessing cost]," he says, "but it gives us a framework to work out from, and make sure we can actually have the data accessible and, with Aster as well, that we can actually have in-database algorithms running.
"So far, from a pure data handling perspective, I think it's giving value for money, but the important data stands in front of us - we haven't done that to a large extent yet. So I'd say that's the journey ahead."
Asked for a specific example of a game-changer for Volvo that would only be possible with Teradata, Wassén cites the number-crunching the firm does on customer surveys, which the company uses to predict who might want to buy a new vehicle within the next six months.
"They responded to the survey and we checked the data, and we'd always thought the ones who had a good response would be more likely to buy a new vehicle, but it turned out it was simply just if they'd answered the survey they'd be more likely to buy again."
Volvo's next major push into analytics is a self-driving car programme dubbed "Project 26", which refers to the average time it takes to drive to and from work, 26 minutes.
"If this feature could free up 26 minutes of time, then come the questions of ‘What could we offer for that 26 minutes?' Being able to have a comfortable situation to take a nap in the vehicle, or to provide time for doing your office work, or consuming infotainment like films and music too," says Wassén.
Big data analytics - making you lazier and giving you more boxset time. At least Volvo seems to have its priorities right.