#190 Part 1: The value of data science, and bringing new innovation to market, with Sarah Dods, formerly Head of Advanced Analytics at AGL

On this episode of the podcast, we are excited to speak to Sarah Dods, one of Australia’s top tech entrepreneurs, innovators, and company leaders.

In a career that has spanned from public organisations like NICTA and CSIRO, to private companies as wide ranging as AGL, Telstra Health, RMIT University and Gerson Lehrman, Sarah has been at the forefront of bringing technology and ideas to market.

As Sarah says, one of the big challenges with data science is articulating its value. Data science costs money to develop, and data science costs money to run. So, why would somebody pay money for what you’re doing? In this episode, she shares some of her proven strategies for justifying to those outside of the data science team how the investments will create and add value.

The other great challenge that the data science team needs to grapple with, Sarah adds, is change management – how do you explain an application or product to people that have not used before?

Stay tuned as Sarah talks about how the data science teams can turn these challenges into opportunities.

Thanks to our sponsor, Talent Insights Group!

We're always striving to further the conversation around the latest AI advancements and we would love to keep you involved in future conversations. Please join our Slack community to keep in touch with the wider data community.

“When you're trying to achieve an outcome, that means that by definition, you're not there yet.”

- Sarah Dods, formerly Head of Advanced Analytics at AGL

WHAT WE DISCUSSED

0:00 Introduction
12:06 Given your background in startups and commercialization in taking things out of the lab and getting them to value, what do you think we're missing as a data science industry? What are we not thinking about? What should we be focusing on more? Where do you see some of the gaps between the worlds from your experience, and now becoming into the data science space.
17:54 What are some requirements or some ways that you've seen in change management that help it go right? What does it look like when it goes wrong? What are the differences between the two approaches?
22:47 How do you help customers get to the point where you can be having those conversations and fill in the gaps around those blank stares?
26:56 Tell me more about the wireframing piece and approaches that can help the wireframe and make it easier.
28:32 How do you think people can close the gap when it becomes a bit more abstract? Or sometimes people might think about interacting with a model and it might not necessarily have a visual component. Are there any things that we can do when people see it from that perspective?

EPISODE HIGHLIGHTS

• Change management, together with some tech work is a big fan of wireframing.

• Your first version is not going to be the right version and you can't get to the right version without the incorrect first version.

• As an industry, we need to be comfortable with being wrong, and comfortable to improve and iterate.

• The one thing that's not okay is not telling anybody that you were wrong, and trying to fix it. Because it's not learning at that point, it's hiding. And generally, the problem gets bigger until it gets out in the open, while we can actually work on it together.

• I prefer to make a much higher risk, a small decision at the start and set off in a certain direction, because it's much easier to change direction than it is to stop and start. 

---

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.