#192 How diversity is more than an agenda and can build a better business with Michelle-Joy Low, Ph.D, the Head of Data & AI at Reece Group

On this episode of the podcast, we have the privilege to speak to Michelle-Joy Low, Ph.D, the Head of Data & AI at Reece Group. Low talks us through how data is the foundation of the ongoing transformation of one of Australia’s most venerable retail brands, having celebrated 100 years in 2020.

Low explains the importance of diversity in the workforce, and how it leads to better outcomes for the business and better outcomes for the customer. The tech space has been particularly good at recognising the importance of diversity, Low said, and Australia is a great place to work in that regard, but at the same time it’s now important to look beyond participating in the movements, and genuinely build diverse teams that are empowered to speak out about and drive further change.

Low also shares some of the challenges that come from AI, and how she and her team are grappling with them. For example, data is complex to implement and expensive… and the decision makers behind the data projects are distant from the team that builds the applications. The Chief Customer Officer, for example, doesn’t build apps themselves. So how do you tackle that challenge within the data team and deliver outcomes for the organisation?

Tune in for this deep-dive and fascinating conversation about how a business is leveraging data to drive toward better outcomes for all.

Enjoy the show!

Thanks to our sponsor, Talent Insights Group

Discover what the Data Futurology Community is up to...Join here 

“I've held a few roles that are all a little bit different, a different flavour in terms of how analytics and data and the more intelligent technologies were applied. But the common thread through them all is the use of data and intelligent tech, to make something better, whether that's making something work faster, whether that's helping a human make a better decision, or whether that's building a product with some intelligence in it. It's all in service of driving some kind of change.”

- Michelle-Joy Low, Ph.D, Head of Data & AI at Reece Group

WHAT WE DISCUSSED

0:00 Introduction

03:56 It sounds like there is there's a b2b component to the business and some retail side can you tell us about the size and the components of the business?

06:57 How did you find the transition to the industry?

10:05 What are your passions in this space?

18:03 Looking at a whole industry, how do you think we're going on diversity and inclusion?

31:09 What are your frustrations on how the work gets done on the technology, or how it's operationalized?

36:28 Data and technology- where do you see that they're similar and different? And what's the perspective that you take between the two?

43:06 How do you see the differences between a system like an ERP and a data system, like a data warehouse, what do you see as the commonalities and differences when we look at transformation, particularly in cloud migration?

EPISODE HIGHLIGHTS

• But a lot of this portfolio is really centred on the one proposition, something we're obsessed about, which is how is data going to be used to help improve the lives of our customers or our people? And so yeah, it's that's really at the heart of what we do.

• And so I think the common thread then, and the thing I've found over time has been the biggest reward is in seeing the work, change, or behaviour or decision for the better And I think the passion and sustenance in you know, keeping up with it is seeing something through what I care a lot about is being able to see something through because a lot of these things don't happen quickly, you know, building a model can be quick to prototype, or it can be quick to experiment with the research pro, you know, progress or process, you know, is one that's quite iterative and fun and quick to start, but really hard to turn into a sustainable solution that you run and operate and continually invest in. And so to realize a lot, the value of the work we do in this space, I think calls for persistence and a conviction that what you're doing matters and will generate value.

• I guess proving out value, given that sort of uncertainty over the lifecycle of data is a tricky one, there's not it's not a linear path. And so I think something that helped, certain things I've worked in, is to take you to know, fail fast sort of approach to measuring value, proving out the value and constantly checking, and adapting if we need to, you know, and this does, it's a combination of working backwards from an end state, but also having the discipline to periodically check if it is delivering value quite often, you know, we get asked questions like, can we build a smart AI thing into one of our digital products? And the answer actually, is maybe we need to ask that question a bit differently, which is, if we were to build a very smart thing into our digital products, what does that do for our customer? Is that going to change? How is that going to make something better? Are they going to find what they need faster? Or are they going to get a delivery out to where they are fast? So whatever that might be? Let's work backwards from what a great outcome we think might look like.

• The discipline to accept that a failed experiment is as valuable, if not more valuable than a winning experiment. I think it's something that we've certainly found beneficial here at race.

• I love that approach to start with a, with a problem, a business problem, or something that we want to improve something that we know that our customers care about, and work backwards to see if AI is needed. And if so, what it would do, that's, that's great. And then doing different hypothesis testing to see what customers really, really want. Because sometimes what people say and what they want are slightly different. And this is the only way to find it.

• It's a challenging industry to be in data and technology. I won't call these frustrations, no in a real negative sense of there are real challenges and as every industry does, but I think a couple of things are unique to data and AI, in particular, one sort of ties in with what I was earlier mentioning about seeing things through the value of a leader is not easily seen. Data is hard, it's complex, is expensive. It's expensive because a decision maker is usually very far away from how something is built, you know, the chief customer officer, for example, almost will never make a decision about whether to build an app with a free text field to collect customer feedback or to or with a drop-down list. But the chief customer officer will care if it was built in a free text field that they later need a machine learning team to decipher through all the data that comes back to get an understanding of customer sentiment, right. So small decisions can become really consequential later on. So it is expensive, it is complex and hard to then prove the value because usually the time periods between now and the leader are long. And you don't always live to see the end result of a choice made, you know, much earlier.

• I think related to that, is that accountability over those horizons is also really hard to measure, you know, the people who started the transformation, normally don't stick around long enough to see the end result. And new people get on the bus or come on the journey and decisions change. So the short term horizon of a lot of investments leads to probably much longer journeys than organizations might want to be on when it comes to data and transformations. So you know, these are very real challenges. And can be frustrating

• I think that and technology are very related areas of work but have some distinctions. You can't serve up scaled solutions without technology, certainly not scale data solutions, what I mean is, without technology, you know, we have the luxury of cloud technologies, and distributed computing. And so they are complementary complementarian, in that respect, and the technical aspects of architecting, a system, really to bring in an effective technical architecture for delivering data, you know, requires both deep software engineering skills, for example, as well as a deep understanding of information, how you manage it through its lifecycle, which in and of itself, it's undisciplined.

• Getting to a successful outcome requires a learning process. And in many instances, so I think the art is in, at what point have we learned enough that's of value to production eyes, and, you know, I think someone wants it, you know, don't go for perfection, go for good enough. And add to that, you know, and so I think the journey we've done is very much that way in that allowing for some degree of iteration, but wrapping around it, the discipline to go that's got enough value, let's make it available to people.

• Part of our DNA has always been a transformation, you know, we didn't wake up one day and say, we're going to transform and here we go. It's no DNA to improve the lives of our people and customers every day. That's something that's been loud and proud to hear it reads every day. So for wellness transformation is almost just part of how we plan to do things. People expect process improvements, people expect tech improvements, you know, we've got continuous improvements, teams, we've got continuous uplift rhythms.

---

At Data Futurology, we are always working to bring you use cases, new approaches and everything related to the most relevant topics in data science to help you get the most value out of these technologies! Check out our upcoming events for more amazing content.