10. The Value of Data for Organizations - A Talk With Lieve Lanoye & Paola Masuzzo (TP Vision)

Listen, and get insights into the Value of Data for Organizations in this talk with Lieve Lanoye, Team Leader Market Intro & Big Data Team,  and Paola Masuzzo, Lead Data Scientist at TP Vision . We talk about when and why TP Vision started with their data program; measuring data value; how data can help your organization move forward; the biggest threats to data value and how to handle those; and we give you some resource recommendations to learn more about the value of data.

Guest Info

• LinkedIn: Lieve Lanoye
• LinkedIn: Paola Masuzzo

Resources

• Big Data at Work: Dispelling the Myths, Uncovering the Opportunities (Thomas Davenport)
• Newsletter: Data Products - Discussing Data Product Development, Semantic Layers, Data APIs, and Modern Data Modelling (Chad Sanderson)
• Storytelling with Data (Cole Nussbaumer Knaflic)
• The 6 most common types of bias when working with data (The Metabase Team)
• Thinking, Fast and Slow (Daniel Kahneman)

Mentoring Program - Women in Big Data
Mentoring is essential to success at every stage of a women’s career, both as a mentee and mentor. The many WiBD mentoring programs are open to WiBD members and cover opportunities for junior, mid-career, and senior women in technology. Not yet a member? No worries. By joining a mentoring program, you automatically become a WiBD member. Both membership and mentoring are free of charge.

Website: Women in Big Data Podcast
LinkedIn:  Follow - Women in Big Data
Contact us: datawomen@protonmail.com

Transcript

Note: Podcast transcription edited to improve readability.

Desiree Timmermans  00:02

Hello, welcome to the Women in Big Data Brussels Podcast, where we talk about big data topics with diversity and inclusiveness in mind. We do this to inspire you and to connect, engage, grow, and champion the success of women in big data. The aim of this podcast is to reveal to you what you can do with big data, how organizations and societies use it, and the potential of big data to create a better future for everyone. 

Paola Masuzzo  00:29

So, if you're saying with a certain expertise in a certain profession: 'and my algorithm does better than you,' you need to show why. And you need to try and explain why that is: you need to take that algorithmic, that bunch of equations and numbers, and make sure that people understand them.

Desiree Timmermans  00:50

In this 10th episode, Valerie and I talk about the value of data for organizations with Lieve Lanoye and Paola Massuzo from TP Vision. Lieve is the Team Leader - Market Intro & Big Data Team. And Paola is the Lead Data Scientist. We cover when and why TP Vision started with a data program, the measuring of data value, how data can help your organization move forward, the biggest threats to data value, and how to handle those. And we give you some resource recommendations if you want to learn more about the value of data.

Let's start.

Desiree Timmermans  01:26

Lieve, Paola, welcome to the podcast. I'm very happy that you're here with us to talk about the value of data for organizations. So Lieve, let's begin: when did you start with a data program in your organization and why?

Lieve Lanoye  01:41

There were some initiatives before, but it really started in 2016. We had a reorganization of the company. And upon that reorganization, management realized that big data was important, then a full Big Data Team was created. And when we started with the big data team, of course, we needed to know what we were going to do. It's not because everybody else was doing it that we could just use that as a copy. So, we took that as an opportunity to think about and reflect on what we wanted to do with big data. Why would we actually bother going through that?

Lieve Lanoye  02:15

Maybe, to mention for the listeners: So TP Vision, we work with TVs - branded Philips TVs. So in that exercise, we went through the entire lifecycle of a TV. We looked into the different stages that were there, and what information was available next to big data. So if we would not have big data at that stage of the project: what were the other sources of information that colleagues would have to draw conclusions upon? And so starting with - of course, the TV - with the plan: which TV will be making which features will be added to that TV? So there is already a lot of information. You have the information from your suppliers. You have the outlines of these bigger companies that are busy with how well our life looks in five years and 10 years. So you have a lot of information already.

Through development as well, there's a lot of information, test results, and so on. Then you put your product in production - there are a lot of reports there as well. But in the end, your products get sold. And unfortunately, in spite of all the testing, not every product works 100%. So you have products at the customer's home that are malfunctioning.

And we realized upon that exercise: that was the area where we wanted to play because if a customer calls the helpdesk, we cannot expect our customer to be technical. So the best they can do is to give a description of what happened, what they saw. But that was not necessarily linked to what was happening inside a TV. It's disconnected: your screen can go blank, but that can have a million root causes. So, we decided to focus on that part to start with. So the big data for customer care for what we call cost of non-quality, so aftercare of the TV. So that's what we focused on first by collecting software problems and then, bit by bit, extending it. Because when you have that, you have all the information there. I have to say, of course: we ask people for permission. We don't upload that just blindly. We have the entire GDPR framework to adhere to. So if they agree to that: we can see if something goes wrong within their TV and fundamentally in the software itself, what actually goes wrong? And with all that information combined, we can solve the problems from the fields directly. So, that's where we decided to get started with our big data program.

Desiree Timmermans  04:32

And how is it evolved from 2016 to 2023?

Lieve Lanoye  04:37

I think we can say it exploded. We started with customer care looking at software crashes. But of course, a lot of other issues were added on top. Not everything that customers see in the end is a software crash. Or sometimes, we work with Android systems and Linux systems. So on the Android systems, people can install applications themselves. Sometimes, it's not the main software; it's the application causing problems. So the modeling of that malfunctioning became more and more complex, the more and more precise we wanted to be in our modeling, and the more and more accurate we became.

And then, of course, it also triggered an entire wave, let's say, of data-driven decision-making inside the organization, for example, preventively in testing. So if we had identified the sequence that would cause problems in the field, the test teams would incorporate that sequence preventively in their tests of other platforms. Or the test suites that we use, so the test scripts that we use are also using data to make sure that whatever the team is spending time on testing is actually cases that are happening with the customers at home a lot of the time. It's really spread throughout the organization at all levels.

In planning, now it's seen if a certain feature is used or not, so it needs to be included in the TV or not: So that you can focus your time on the features that matter. During development, now: the same way we monitor TVs at the customer's homes, we also monitor all our TVs and test them, regardless of where they are, to see how these are behaving to see if the issues that are seen - or issues that we see on other platforms in the field as well: if it's a new issue, if it's an existing issue. So it's now throughout the entire lifecycle. It spread like wildfire, I would say.

Desiree Timmermans  06:25

Okay, well, interesting. Thanks for that. And Paola, we now know when you started with a data program and why, but how do you measure the value of data in your organization?

Paola Masuzzo  06:37

Let me start by saying that I'm very happy to be here today.

I would say that the short answer is that: there is not one measure that fits every purpose and every organization. If our organization would be a business that sells data, then the return on investment - and investment and the value - would be pretty straightforward to calculate and estimate. But like the big majority of companies in the tech environment nowadays that use data: data for us is a tool. It's a framework. It's not the core of the business. But it's something that we use to enhance how we do business. That basically means that whatever we produce: it's mostly a service type of functions, like Lieve was highlighting before talking about product development together with product management and a cost of non-quality.

So sometimes, it's very difficult to have precise metrics that tell you: okay, this type of data pipeline, of dashboard, of analytics report produces exactly this type of value, right? You can't really quantify.

Desiree Timmermans  07:40

I can imagine.

Paola Masuzzo  07:41

Sometimes, like Lieve mentioned before, if you need to understand: are we going to keep this feature inside our TVs for the next generations of product? You would have to do research or investigation in the market, you would have to acquire perhaps data from external agencies, and still wouldn't reflect exactly our user base. So the time to make that decision would be much longer. And also, most likely, you will end up with a decision that it's not exactly compliant with the reality of how things are in our market.

One of the things that we look into, for example, is 'time to decision'. Sometimes, like Lieve mentioned before, if you need to understand: are we going to keep this feature inside our TVs for the next generations of product? You would have to do research or investigation in the market, you would have to acquire perhaps data from external agencies, and still wouldn't reflect exactly our user base. So the time to make that decision would be much longer. And also, most likely, you will end up with a decision that it's not exactly compliant with the reality of how things are in our market.

So with data, the answer to: should we keep this feature, yes or no, is more straightforward. It still requires some work, of course, but the decision can be taken faster and more reliably. Having said that, what happens mostly with data teams is that you need to elbow your way within your organization and say, 'hey, here we are', you need to invest in data because data is good for business. Not only because everybody's doing it, that's never the right framework or the right reason: but because we can create value, make better decisions, get better products, can be there for our customers in a different way.

And once you have elbowed your way in slowly, day by day, month by month, and year by year, then you also have the other stakeholders that are ready to sing your praises. They will be the ones that say: okay, we need now to look into products for the next two to three years; let's go to the big data team because they know the current situation of future usage, know how people are interacting with their televisions, know what people struggle the most with.

So, in short, there is no metric that fits every purpose. If you're doing your job properly and are focusing on the business questions - never just on technology -: why would you need the data to help the organization? In which sectors are those data going to help your organization move forward? Then, proving the value of data, it's something that will become natural. And it doesn't necessarily have to fit into an Excel spreadsheet with numbers and metrics that sometimes are not capable of capturing the complexity of service support, as the big data team provides. I believe Lieve would, and I hope she would agree with this.

Lieve Lanoye  10:23