EP 25 - From Data Abundance to Decision Intelligence
The conversation is no longer about how much data we can collect; it’s about how wisely we can act on it.
In this Episode:
Francois Jacquemin and Neil Wirasinha discuss the strategic shift from data collection to decision intelligence across insurance and entertainment industries, highlighting precision, alignment, and cross-sector learnings.
In my latest conversation with Neil, we delved deeply into this transition. From entertainment to insurance, two seemingly distant sectors are converging on the same truth: the future belongs to those who can translate data into decisive, intelligent action.
Neil’s perspective, rooted in media and streaming, made one thing clear. The entertainment industry thrives on immediacy, knowing what the customer wants before they do, and tailoring every touchpoint to that. Their data isn’t just about volume, it’s about nuance. Which show did you almost finish? What post made you click? Which segment converted?
In insurance, we’re increasingly navigating the same terrain. But our stakes are different. We aren’t just personalizing experiences, we're managing risk, protecting lives, and anticipating costly claims before they materialize. That demands a higher level of precision.
We discussed the importance of controlling your own data architecture. Whether you’re Netflix or a large insurer working through third-party administrators, owning the loop from signal to insight to decision is where real advantage lies.
At its core, this transformation is about something greater than tools or dashboards. It’s about alignment. Fragmented data leads to noise. Precision leads to insight.
Yes, biases exist. Models misfire. But as processing power grows, so too does our ability to interrogate assumptions, refine models, and shorten the path from observation to action.
The future of insurance and leadership in general demands decision intelligence. Not just seeing more, but seeing clearer. And most importantly, acting sooner.
Timecode:
00:00 Introduction to Data Utilization
00:49 The Depth of Industry Data
01:19 Challenges and Benefits of Data in Advertising
02:19 Insurance Industry Data Strategies
05:03 Data Processing and Analysis
09:32 Social Media and Movie Industry Insights
Francois Links:
Apple Podcast
Guest:
Neil Wirasinha:https://www.linkedin.com/in/neilwirasinha
Transcript:
Francois: data has been there for a long time. You're an engineer. I'm an engineer. We started with that even before knowing where we're gonna land. The, um, the amount of data has always been available but never used because although it was there and available, transforming it into a usable availability, was, was very challenging. But, uh, for recently there's been much more ability and calculation power and the ability. To mine data uh, a much better way. for our industry to, to, actually get something meaningful out of it. Before it was nice to have it, there were trends. We could do graphs, but now it could be meaningful. So how deep is the data you use in your industry?
Neil: Oh, the data So it really depends. So when you own your own architecture, like Netflix.you You, you, have a much deeper, richer vein of data that you can see what people are watching, how long are they watching for it, do they rate the shows afterwards?
What else can I serve to them? So it can be a lot more tailored and a lot more, um, uh, customized to your needs and wants. But the downside to Netflix is they don't spend much money in advertising, which is where the studios really benefit because they're in the advertising world of, you know, when they're spending half of their budget in digital.
And you can read a signal from that. And with the right data tracking, you can see which Instagram post led to them then going on to either consume more content or ideally with more data tagging, you can actually see them get to the point of transaction at the cinema chain. That relationship stops when you get to the cinema chain, unless you have some sort of data agreement with that cinema chain that we can see how they've gone all the way through to transaction.
Francois: there's a, little fight for data.
Neil: there's a, there's always There's always, a little bit of a, uh, of a, of a relationship
Francois: Yeah. Yeah.
Neil: And, uh, and as, as, as, and, and that relationship needs to become more normal, more quickly because then it means that you can get your consumer's data much more readily available.
Francois: It's so important for us, it's, uh, we worked with, uh, TPA, so it's basically the insurer is not doing all the work. It delegates part of the work,of the work to a, to a servicing for, and um, the key for the insurer is to get the data,and, and, and, and, and use the data to be able to, to see what the trends are and, uh, in the cost of, whatever we cover and be able to adapt. So you change the, the product, you put limitations or you increase the Price or you reduce the price, which because, of, you know, inflation usually is, doesn't happen. Usually it's trying to go up. But also it allows an insurance company to take action to prevent, as we, we spoke earlier, or to negotiate or, or to limit the cost of some, some reparation or medical. Procedure also, or mutualize them anyway. There's plenty of techniques. that that can
Neil: Yeah.
Francois: Yeah. But it only works when you, get the data and and, and, and, and you work either with a third party administrator or if you're a big group, you have your subsidiaries. But putting all the data together and and, it's always a challenge when you match that you have a competitive advantage over the others because you know what's going on.and being able to, to create proper assessment of those. data. so much data before We're happy when you see, you know, oh, there's that, that company that have 5,000 people in that location, they always go to an hospital, which is uh, you know, very close, but the hospital a bit farther, has better healthcare and is is, cheaper.So you try to diver that. You either Negotiate with the hospital to say, well, you keep on getting all clients, or, you know, our employees is your clients, but we, we, we have a better price and you need to stop charging that ludicrous amount of money. Or we send them somewhere. else and then incentivization to So hat was like 10 years ago. But Now you can start using the data, to actually improve the health of those workers.
Of course, the hospital stuff is, is done and so on. But start,start taking different action, which brings us closer to the client. have maybe a sentiment of
Neil: Yeah.
Francois: but creating something which is,um, helps to avoid that there is a, a claim happening. Um, so I find that the the more we are able to use data, the more we are gonna drive towards, uh, a much extended servicing from the insurance world. And I don't know if that's the same
Neil: I think it's ex, it's, it's exactly the same as what the issues are in the, in the entertainment sector. That, that data and the processing power of that data is really important. The, the, you know, valuing data and understanding which data is important, which is what is a primary signal. What is a signal that feels like the others, but maybe doesn't point to an outcome. A lot of it in the past has been navigating because of the limited processing power. They were always studies over a period of time, how do we do this? You know, is that the same signal? There's always a need, especially when you're working through. Uh, an intermediary, like a media agency, a digital media agency, they will often add a series of metrics, so many metrics to your business because it's, and it's not because they're trying to distract you from the job you are trying to do.
It's because individually you've got some very bright people trying to answer a series of questions of how can we sell a ticket more effectively? But just by asking that one question, the interpretation of that question or the biases of that question to the three or four people that are, are trying to answer it in their way, might mean that you end up coming back with.
A serious amount of data that quite a lot of it, whilst it's complimentary, it's like, well, which are the three things that really make a difference? What's the, oh, we've got 21 pieces. It's like, okay, well let's simplify this, Dan. The question was this, if these, if these signals are genuinely unique, business driving signals, fantastic. If these are, you know, complimentary signals, let's remove them. That's when processing was at a lower level. Now the processing can be significantly higher. Yes, you have to write the prompts to make sure that they're valid and understood, and the key ones you can run the trend analysis quicker and, and, and, and really look for any of those changes and trends. So it's gonna be super exciting. I think that it's, it, it will be exciting. It's gonna put more pressure, uh, I suppose, on a media agency to, to, to prove to a client they understand the business because. Agencies have always been an extension of the client's
Francois: Yeah.
Neil: Because the clients can't go and ask for more headcount.So you end up having more people in your agency team and offsetting it into the marketing budget, right? So there, there've been a parallel, uh, team member, but it won't take long for the right technology partners to partner up with clients to say actually the way that you run your business. And it might not be the major, uh, the major.
Consulting groups, the four big consulting groups. It might be something a bit smaller, it might be something that can be a little bit more tailored. It's gonna be difficult for clients to build that capability in-house. They would normally like to look at a service or a product and, and then trial it for a little while and, and then take it to, to, to full, to, to full utility.
But I think the most important thing is it's about the utility. And I think that clients all want to understand their customer better. What, what they like and what they don't like. And, and it's, it's, the things they don't like are probably really important because it's about then adapting the product to make sure that they have a better relationship
Francois: So You have to also be very, very precise with the data. You know that, uh, if, um, if you analyze a big set of data and insurance, I'm just taking that because it's, you know, you have it, it's this, the data that you select will be,uh, to train your model will, will be a smaller set. So There will be bias in it.and then you will add data and you need to access to more and so on and new development.And you now it can access more data, so et But you can't, if, if, if you, if you get it wrong, you know, and it's the bias, it will just completely flow your model, so you endanger your own profitability or your,own, product line. Or even your company, although,you know, I think that since there is level three lines of defense,
Neil: yeah.
Francois: gonna Be lots of people in there involved looking at this critically.yeah,
Neil: so let's not put everything against it.
Francois: but it could suck up a lot of resources for A no result, or wrong result.which we don't want that. So precision is also important. Yeah.So, but I understand it. It's, it's the same
Neil: It's the same thing. It's exactly the same thing, and I think.If you look at all the social media platforms that are out there and all the data that you can pull from those, I sort of said the movie studios, when they break a trade or any part of a campaign, if it's digital investment, then they, they can draw back a number of signals from that, from the audience or whatever that might be.But it's trying to rationalize that down into how does that help this movie right now? And also, what's the learning that we can take into the next movie? And that's where the dividend really plays out, and it's gonna be, and it's gonna be more compressed. So it means that they can, they can learn from some of those mistakes a little bit quicker or even offset some of those mistakes a little bit quicker.Obviously, audiences across the world are similar and very different all at the same time. And taste and censorship and culture, it's, it, it can play. Huge role. So it's about making sure that the data and those biases, um, you know, are respected or, or, or protocols are set out so that data can con, can be consumed.The, the worst way that it happens is that those models are run in the US looking at the US consumer, when actually the play is, is, is a, is a multicultural, you know, it's a multi territory thing. Even within a market, even within a, in a country like the US where you've got, it's not a singular culture.There's multiple cultures. And so it's about how do you, how do you set out those
Francois: those They know that in the us we don't know that here, but the US thinks that the rest of the world.is mono as well.Right? Yeah.