Don't think data, think people and relationships.

[00:00:00] Loris Marini: If you’ve been following this podcast, you know what I think about data management. We simply cannot run organizations effectively without the right data, the right information, given at the right time. The problem is that without management, data is unusable. It depreciates. Forget for a second about the data and follow the question all the way to the individual. What do we really want? Well, I don't have all the answers, but I think it's safe to say that we want to enjoy our work. We want stability and we want to drive back home to our families and to our kids with a little bit of brain power left so we can enjoy life after work.

And so, you might ask yourself, “what's the point of me thinking about family and work-life balance?” Well, I think it's because one could make an argument that there is a connection between frustration and stress and lack of information flow. When we don't have the right information at the right time, things slow down and eventually, we suffer in a way or the other.

So, what is data management all about? It's about giving the right people access to the right data at the right time. And of course, it's easier said than done. So why is it so hard and what can we do to maximize the chances of success? To explore these questions today, I have the pleasure of speaking with FX Nicolas, VP of product at Semarchy, one of the top companies in the data management space. FX has been in product for more than 50 years; from Sunopsis to Oracle and Semarchy. I'm sure you're gonna enjoy this one.

I'm here with FX. Welcome to The Data Project.

[00:03:36] FX Nicolas: Thank you very much.

FX Journey to Data Management

[00:03:37] Loris Marini: Let's start with your story. How did you end up in data?

[00:03:46] FX Nicolas: It started after college. Initially, I as working with software vendors: on fax servers, the health cost on legacy machines. And I had a little bit of database access at that time. I jumped to data through data integration in a company called Sunopsis that was acquired a little bit later by Oracle. Now the product is named Oracle data integrator. So I started doing product management at that level.

[00:04:15] Loris Marini: Okay.

[00:04:15] FX Nicolas: The data integrator is a data integration tool, ETL, ELT. By the way, we at Sunopsis created that term, ELT. That's a very interesting but it's another story. And then I switched from data integration then master data management and now, more generally, I work in data management.

A day in the life of a VP of product

[00:04:33] Loris Marini: It's been quite a journey. As a VP of Product, can you describe your job? What do you like about the job as well as what you least like?

[00:04:43] FX Nicolas: The part that I like the most is actually creating features. That moment where you speak with the clients, you see what requirements you need, and you just have this spark, this idea that pops into your mind and say, “okay, this is what they need.” And you see the feature that is going to bring you some competitive advantage. And once you have this spark, the rest is just like surfing a wave. Well, it's not very easy because there are some technicalities, of course, but it's very exhilarating.

And the thing that I would say I like less is actually the phase that comes afterwards. Because once you get that spark, you just want things to be. But it's not done yet.

[00:05:26] Loris Marini: Yeah.

[00:05:26] FX Nicolas: And usually there is some latency, but that's reality. Because having smart ideas like, I wish I could have a space ship, but it's not there yet. Come on. It's exactly the same thing. That's just normal life.

So, you've designed it. You've seen people working on it, participating in real big teamwork. And when it goes out and you release it and you give it to clients, you just saved one hour of my every day for me. So it’s like, yay, job done.

[00:05:57] Loris Marini: Man, you'd be my best mate iIf you can save one hour of my job every day, that's it. Tell me where to sign.

[00:06:03] FX Nicolas: That's exactly what we do.

Designing for adoption

[00:06:04] Loris Marini: Yeah, it's funny because I completely agree with the fact that there is latency to creating and solving problems, regardless of the domain. It's true in carpentry, construction, engineering, and it's true in software.

In software, because everything is so immaterial, you can't touch code, we always have this expectation that things should happen quickly, but it takes time. And if you're a software engineer listening to this, you know exactly what I'm talking about. You have when you first write the code, and then you have production quality code.

[00:06:42] FX Nicolas: That's the prototype. And then there's the actual released version.

[00:06:46] Loris Marini: Yeah, but even if you remove the technical barrier, I think there are still other elements to solving a problem that go beyond that. I've been experimenting with a no-code platform for my own data management. So there's no technical barrier. You just build, put together the blocks, and you create functionality out of simple elements, but there's still human behavior and you have to get used to new processes. So, in the end, it's not just technology that’s solving the problem.

Everything looks easy, but is it?

From your perspective, how big is the gap, how much people are expecting things to happen with a snap of their fingers thanks to a magic tool?

[00:07:36] FX Nicolas: I would say that it's a human way of thinking when you compare what you’re doing to someone else. When you do it, it’s harder but when somebody else does it, it's easy. It's our way of thinking.

You just have to put yourself in the shoes of the other guy and see, “If I really had to do it myself, how long would it take?” And be honest with yourself when you say that. And it's difficult to do, because if you don't have the skill of the guy who is working in front of you, well, what he does will look easy.

I'm going to give you an example: Take a great artist and you see them painting. It looks easy. And if you look at the painting, it looks easy. Picasso had a great quote, which is, “It took me four years to paint like Raphael, but a lifetime to paint like a child.” Learning the perfect gesture took him six years. And that's the same thing with software. If you look at the great developers, great designers, what they're doing seems easy. But what we miss is all the processes that they undergo before they accomplish their work.

Either they do it in the backend so you don't see it, or they conceptualize it in their brain. And that's basically the six years of experience. At the end of the day, it's a perception.

And we have the same situation when we speak with clients. Even when I speak with our team, people just see software, they see projects as so easy to complete.

What is the human side of data?

[00:09:11] Loris Marini: There's also a difference between solutions that are predictable and those that need to adjust and change over time in data. And I wonder about your take on this? I feel like the feedback loop has to be somewhat closer, not just in the sense of the teams building the products, the data hubs, or the data platforms.

Once the software is there and once the functionality is available to the team, going from having that functionality to getting to an outcome for the business or the organization is a completely different story. It's the human element, it’s the people, their behavior, their world and incentive systems. Do you think that the human aspect is understood intuitively?

[00:10:14] FX Nicolas: I would say that this is just a drawback from the fact that we treat data as a technical thing. It’s a technical thing because we are talking about IT, about computers and systems and databases. However, the people who will be using and leveraging data are human beings. They think in terms of the business, or as you’ve said, in terms like, “how can I go back home earlier?” People do not think in terms of technical stuff, they think about what the outcome will be for them. What is the result for me? What is the result for the company?

I always tell my clients; we don't do data for the pleasure of doing data. We don't do data quality to have good data quality metrics. We do good data quality either for the people to be happier or for the company to work better. People are happier if they can do their job faster, if they can remove a little bit of friction, or these two or three additional steps. And then from the company’s perspective, you help them improve the processes so that everything runs smoothly in a more predictable way. It’s like a ping-pong action: good companies help good people. People will make the company work better. So, it's going to be a ping-pong effect.

Learning from conversations around data

We try to put the focus on either the users or the company when thinking about a data project. The more I speak with users, the more I see the human aspect: the frustration, this little button that upsets you because it's in your critical working path. This little button is just very, very annoying. Then sometimes it boils down to something that is underneath, something that they don't tell you. It's not exactly the button that blocks them. Maybe it's a change management thing.

And we tend to underestimate that whenever we speak about data governance, which has been a big trend for the past five or six years. Governance is a whole bunch of disciplines, like in change management. There is a certain aspect of psychology in there because as long as you have human beings, you have to do a little bit of psychology. We are not machines, it's not as easy as flipping a switch on or off. So, data governance is extremely important. And underneath that is data processing, then data management. The technicalities should follow but that is difficult because of the changes you mentioned.

The art of managing expectations

[00:12:56] Loris Marini: Yeah. And I heard you say recently in one of your LinkedIn lives that it's important not to boil the ocean, to have that agile mindset in our approach to the problem of managing data with a stance that is centered around agility.

Imagine that there is a problem. You show it to everyone involved, everyone that is affected by the data, directly or indirectly, that there is an actual benefit in having that agile mindset. And that is the propelling force that sustains the next cycle. Do you think this is almost like a general principle or the devil is always in the details / it's a lot trickier in practice?

[00:13:50] FX Nicolas: It's not highly complicated. You have to be a very, very good negotiator. Like with kids: you give them a small piece of cake, they think it’s very good. But then they want something better like a larger piece of cake. It’s the same. Once you have something that looks successful or that looks good, people will ask for more. Normal brain behavior, but what you have to explain to people is, I cannot give you what you’re asking for in the present because I haven't produced this small piece yet.

[00:14:27] Loris Marini: It's in the future.

[00:14:28] FX Nicolas: If you ask me to produce more, it's going to take more time. So, let's stick to the small piece and once you have it and you can taste it, you'll figure out if you like it or not, then we'll be able to modify it. Maybe add some more sugar, or less sugar to make sure that the next piece is even better. You have to negotiate all these things.

But I would say that a more critical aspect is that it is not possible to do it otherwise because of two things. The first is that the more you grow the size of what you want to achieve, the more difficult it is to produce. And so, you're just creating more and more frustration because you create more expectations and you just extend the time to deliver. So, imagine the frustration at the end of the day, very bad practice. And the second reason is while I'm doing that as the time expounds, the probability that there are some changes required on the way expounds too, which is a source of issue because what I'm going to serve you at the end...

[00:15:32] Loris Marini: It's not going to be good.

[00:15:33] FX Nicolas: Is not what you wanted at that moment. Your needs could have changed. At first, you wanted a pizza and then I serve you your pizza, and then it's a different time and now you want a cake.

[00:15:44] Loris Marini: Yeah.

[00:15:46] FX Nicolas: And that's the only way of doing it: by serving smaller pieces at a time. It's the best way to do things. Which doesn't exclude the fact that you still need to have the big picture.

If you miss that big picture while creating the small pieces you're just putting small pieces on the table without having an idea of what the whole meal will look like. I'm using a food metaphor.

[00:16:20] Loris Marini: No, no, no. Actually, I'm hungry now. Thank you.

[00:16:25] FX Nicolas: If you're just putting small pieces on the table, that's fine. But if they don't fit together, you're missing something. That's why you need to have a menu at a certain point, you know? And that's where the governance is going to make sense. You step back a little bit. You're not sticking to one piece. You step back and see if everything works fine together.

Governance means that you have the whole view. It doesn't mean that you have to do everything at once. It's impossible.

[00:16:52] Loris Marini: Yeah. I mean, it could even be possible if data wasn’t what it is: an asset, a resource that touches every single business line, every single person in every part of the organization. It's just a matter of when they're going to be touched by a piece of data, but eventually it's all connected because data is almost a digital representation of what's happening in the world. It's inevitable that everyone, at some point will be affected by a data set. Let's call it information assets in general.

Why Semarchy

But, let's crystallize the conversation a little bit. Can you give me an overhead view of what Semarchy does and what types of customers you work with?

[00:17:47] FX Nicolas: Semarchy actually produces what we call the xDM platform. So xDM is an intelligent data hub. In essence, it's a single piece of software that lets you run your data management initiatives, and I'm using initiatives on purpose, because for me, projects have a beginning and an end, initiatives last forever.

You’re never done when you do a data management initiative. As long as there is more data coming in or data changes, you are never done. So, we have this platform where you can do an end-to-end initiative, which starts by discovering the data: profiling, doing some data governance based on what you found, some data integration because you need to move back and forth data.

[00:18:34] Loris Marini: Yeah.

[00:18:34] FX Nicolas: Data management, which is applying data orders, data rules. What I mean by data rules is qualities security things on the data. And at the end of the day, measuring the outcome of your data management initiative so that you can have this loopback to governance. It's basically discovering, governing, managing, integrating, measuring, and then there's this loop.

So it's a platform that does all these.

Some clients use it just for the purpose of management, some clients use for governance, but once you get the whole platform’s capabilities, then you can drive your whole data management initiative.

For the clients we work with, we have a variety of customers out there. We don't focus on one type of industry or a use case. In the US we have Chipotle. In France, we have the cosmetics brand Rocher. We have clients in the UK, like BAE Systems. So, we have customers across the globe, and the main type of data that they manage on the platform are customer data and supplier data, mainly what we call parties. B2C, B2B parties, products or assets, and then locations. For example, we have a client in the UK called RES and they're managing energy production plants, basically solar energy plants and wind energy plants, these types of things. We have the structure of the plants as well as the various pieces in their factories. This same concept applies for clients who manage housings. So, a house with its contents, every appliance, piece of cupboard that is in the house is part of their data model.

[00:20:33] Loris Marini: Wow.

A relationship-centric view of data

[00:20:33] FX Nicolas: And if you look at it, the house is a location with plenty of stuff composing it. It's like the plant. The plant is a location with plenty of stuff composing it. From our perspective, these two things are almost the same. Instead of naming your house, you name it a plant. And instead of naming your piece, same stuff.

[00:20:54] Loris Marini: Yeah. At some point, everything is obstructed away into objects that have relationships with other objects.

[00:21:04] FX Nicolas: Exactly.

[00:21:05] Loris Marini: And this relationship is really crucial. It’s one of the most beautiful concepts in software, in data, in business.

Nothing happens in isolation. Everything of value, if you actually think about it, flows from something. There could be an exchange. It could be a transactional exchange. I give you $5; you give me a beer. Or it could be the opposite.

Like what we're doing now, we're having this conversation. We’re exchanging ideas. There is no transaction. Maybe it's obvious. But when I think about relationships in that way, I realize that they're everywhere. And so, sort of the job of the data team or data platform team is to encode that into the database.

By encoding I mean to say the representation of the relationship in the real world, in a relationship in the intangible world. And as the real-world changes, those intangible relationships also need to change to follow the reality.

And there's always going to be a gap, right? Unless you have that visibility across the different areas of the business, how do you know what's going on? How do you know how big the gap is between what you know and what is actually happening? There's no way.

[00:22:42] FX Nicolas: What I saw is when you don't have the relationships properly managed, you get into trouble.

The first example is analytical processing. When you don't have the relation between a client and a location. If you don't have all these relations properly mapped from the data perspective, then when you arrive at the analytical piece you're done for because you're trying to figure out what products are sold in this country.

[00:23:18] Loris Marini: Good luck.

[00:23:18] FX Nicolas: And it's kind of fuzzy magic, some numbers appear, and you're not sure about the numbers.

And the second example is when you get into the operational urgency. Say that this client calls me and he's got a problem with a product that he bought. He tells me, “I bought the product at that location, and I'm not able to find it. I'm not able to find the customer in the location and the employee who sold the product to the client.”

And then you have this idea that you mentioned which is interaction. Interaction is not something that's formal like having a chat with someone on your Twitter feed or email, you know, it's interaction. If I send you an email that doesn’t mean that we signed a contract. However, we have some kind of relation. There's a relationship that's created. How you want to use these, like when you look at it in data, for example, reporting, do I want to base it on invoices or transactions?

[00:24:28] Loris Marini: I had a similar conversation with Chris Boys, the founder of Umano. He's been doing great work recently, building a platform for helping teams find their own pace and momentum and velocity, and helping them mix tangible and intangible. So, like what you’ve said about the transactional aspects, things that you can quantify and things that you can't really quantify, you can give a rating but it's completely based on your feelings and not the actual data. And you need both of them, they’re two dimensions.

This relates to another conversation I had with Gilbert Eijkelenboom here on the podcast where we talked about decision-making and he recently wrote a book called People Skills for Analytical Thinkers, inspired by the work of Daniel Kahneman in Thinking, Fast and Slow about the rational and emotional sides of the brain.

Closing the gaps

I see a strong connection here because everybody says that data capture is crucial. We need to learn how to behave around our data sets, but behavior is triggered by emotions. Emotions are not directly accessible from our rational sphere. There's just not direct communication between those two parts of the brain.

So, you could say that you're facing an impossible task. You're striving for an accurate model that puts relationships at the center. It mixes quantitative and qualitative, but fundamentally the very thing that allows us to be and be conscious is the brain. It's designed in a way that those two parts don't really talk to each other. So, who's going to crack that problem. How do you see this?

[00:26:30] FX Nicolas: You know, in science, a metric has no value without this plus minus. In French we call it “incertitude”. It's basically a value that tells you how accurate your metric is. Even when we designed it, there are data architectures where you can have a very inaccurate number, but you get it very quickly.

If I need a decision right now, it's going to be an instinctual one. I believe that the human brain is able to cope with that. You know what's tricky is that even with the most certain facts, like how a number like 75.38 was produced, you have to look back at the production chain and figure out where this number comes from, because if you don’t, you just trust a number that probably is wrong. It boils down to data.

When we think about this number, we think it's very accurate data. Quality is high. If you’re a scientist and you read a scientific article, most of the article is not about the results. It's about how they came to that result. And then you can argue on the methodology because there can be flaws in your methodology. So, the whole methodology is important to the accuracy is data quality, but then the lineage...

Thinking like a scientist

[00:28:04] Loris Marini: Mm.

[00:28:04] FX Nicolas: The lineage is important because it’s everything that brought you to that precise number. And this is where the more you argue about what you see in front of you, the more you think like a scientist, the more it becomes interesting.

I've got this story for you. A friend of mine was in a meeting with a bunch of venture VPs. And people are talking about results from an application and how the application helped in creating a number of new invoices with this amount of money. And this VP is presenting. You know what they forgot in the process? The number that the guy was showing did not take into account exchange rates. They were adding Euros with dollars, with Francs, with whatever currency they had.

[00:28:57] Loris Marini: Apples and pears.

[00:28:58] FX Nicolas: The guy didn't discuss the path. And in the path, there was this little flow that was the currency exchange rate.

As VP of Product, I usually discuss the numbers. Either I trust the guy who gave me the numbers, otherwise we don’t have visibility.

[00:29:11] Loris Marini: Yeah. Or have visibility.

[00:29:13] FX Nicolas: Show me how you got there.

[00:29:15] Loris Marini: Show me how you got that number. This is a big problem. Science itself has been struggling with this for a long time. There are many papers being published which analyzed other papers in the same field to try and reproduce the results. And, I don't know the numbers, but last time I checked, it was shocking.

We should do a hell of a lot better in science, if we want to call it science, sometimes it's a bit sloppy for sure. And there's been a movement in science to try to use Jupyter notebooks in the Python ecosystem or R notebooks, to try to get people to use systems that allow you to redo what you've done from scratch.

Just to make sure that there's nothing going on in the process because the pipelines can be quite long. If you're measuring the rate of reproduction of a bacteria like E.coli, you've got a Petri dish, you got E.coli, you put it with some nutrients there and you're taking pictures. There's only so many things that can go wrong because there's one bacteria. There's one petri dish. There's one camera. Ultimately, you're comparing pictures.

But in data, people often see different things even when they look at the same number, whereas in science, the same thing is defined the same way across the globe.

When you look at what you’re defining, for example a customer, who is a customer in this context? Sometimes definitions can change.

And so, we go back through that iteration and communication and conversational piece that’s really essential. We can have the best lineage system in the world, but if the stuff we're looking at doesn't actually reflect what's happening outside, in the real world we think we have the same understanding then we'll be spending countless hours in the room, arguing whose numbers are right.

[00:31:24] FX Nicolas: You’re completely right. That’s one thing once I spoke to a consultant with. He was consulting at a hospital in the US and he told me that the hospital had seven different definitions of what a patient is.

[00:31:38] Loris Marini: Oh, great.

[00:31:40] FX Nicolas: When you arrive the hospital to get some type of treatment, they have seven ways of defining you. I don't know how they manage it, because for me, the patient is somebody who gets in sick, gets out healed. It's very easy.

And talking about a common definition in order to know what you're talking about when you say client, I know what a client is. If I give my definition and then yours, I'm fairly sure there's a difference because we don't have the same background.

However, if we work in the same company, we have to agree on a common definition. Another example is when I say mouse. You immediately know what I'm talking about.

Towards building common definition

Some people believe that the client is someone that has paid someone. Sometimes it's someone who's just been invoiced. For someone working in sales, it’s someone who signed the contract. For someone who works in finance, a client needs to have made a payment.

[00:32:54] Loris Marini: Where is the money?

[00:32:55] FX Nicolas: Okay. That's different. When is a client a client?

And just having these clear definitions help a lot. And I assume that you may have different definitions but you have to agree on one. And once you have this, you can have a type of common language and the discussion becomes easier.

A lot of people, I would say the vast majority of people, really like to understand what they're dealing with. I believe there are very few people who just take it as magic. But if you just give them a little bit of information, they'd be very happy to understand the things because it involves them. That's the psychological part. If I don't just give you the thing then I tell you that it’s magic, you’ll just take it and use it. That's it. If I give you the thing and give you a little bit of explanation, you make it yours, you take the thing and make it yours. If you think that you understand a little bit something, you like it more.

[00:34:03] Loris Marini: And you start coming up with the ideas, which is ultimately what we want for it. That's the soul of innovation.

[00:34:09] FX Nicolas: Exactly.

[00:34:12] Loris Marini: Ideas don't come up just because someone gave you the ingredients. Ideas come up when you are curious and you start pursuing a line of inquiry, you start connecting the dots and maybe that picture that you end up with sucks and it's wrong, but you've done the exercise and next time it's going to be better. I mean, it's an attitude.

[00:34:34] FX Nicolas: Exactly. Yeah. I would say it's psychology, but also a company vision for doing things. Whether you want to treat people like machines and I'm giving you something, no need to understand it. You're a piece of the whole engine. That’s not your job. When you want to treat people like people and you tell them, I'm giving you this. I'm giving you a little bit more information and if you want to bring ideas, you're more than welcome.

And I believe that the the first way of managing people is over. It's done. It’s last century. And the new way of managing people, looking for creativity and ideas, and making people feel better in their job, that's the better way of doing things.

[00:35:40] Loris Marini: That's the way to go.

[00:35:41] FX Nicolas: I know this is something that's been written in management literature for ages. People don't go to work just for the salary, they go to work because there is meaning in what they're doing. If you remove that, it's shallow.

[00:36:02] Loris Marini: And they will jump ship as soon as they have a better opportunity because it's purely transactional. Going back to relationships there's nothing more than just the money coming in. That's for sure.

Should data governance be enforced?

So, this is actually why every time I hear about how data governance should be enforced, I'm like, “hmm, let's define what we mean by enforced.” If we mean that it's a program that's backed up by the most senior person in the room and boards are onboard, pardon the pun, and they really encouraged this and there's funding allocated to the program, and everybody understands that and there's data literacy initiatives to uplift people to understand why this matters, then I agree.

But if data governance being enforced is to be interpreted as someone just flipping a switch and deciding that starting tomorrow, if you don’t do what you’re told then you’ll be fired or you’ll receive less money, then it’s the wrong reward system.

[00:37:05] FX Nicolas: It's interesting because when people tell me about data governance, I just tell them it's actually like governance in general. I mean, political governance, it works the same way.

Pick your system, whether you want democracy or pure dictatorship, it's up to you. It has to be enforced. And then the question becomes, how would it be enforced? And it's an important piece because the same rules can be enforced in different ways. It's a government's choice how they’re going to enforce it.

For example, if my rule is you shall not steal, then they can enforce it in multiple ways. I can send you to jail immediately or decide maybe we can chat together and see what the problem is before I decide what's going to happen to you.

The way we enforce it is also part of the governance exercise. I give you a rule, like when you create a new customer in a system, this customer must have an email address; when you create a product for the catalog, it must have a price. You're not putting a free product on my catalog for Christ's sake. You're not doing that.

[00:38:35] Loris Marini: Yep.

[00:38:35] FX Nicolas: We are selling for nothing. In France, that’s a legal issue.

[00:38:43] Loris Marini: Absolutely. Must be enforced

[00:38:45] FX Nicolas: And it's not a must because I decided it. We come back to data literacy. If you're able to explain to people the reason why they have to do something because of some legal implications, you bring the whole company in jeopardy and even your own job, you're putting someone else into trouble.

The worst thing that I saw once was from a prospect. They were creating products, new materials in SAP, and they had this spreadsheet. It was passed through various people. And at the end, I was entering the data, copy-pasting from Excel to SAP. Imagine the process. I looked at the Excel spreadsheet. It was left with complications, plenty of fields didn't make sense to anyone. And they were struggling with that. Half of the rules didn't really make sense.

The lack of governance at that level was really blatant because you just stacked up enforcement systems without trying to govern and thinking about the value. Am I making my user's life easier? They did nothing about it. Just stacked up rules and rules and rules. And it was just like a nightmare. And I came back to them and I said, you should go back to the basics. Figure out why you're doing all these things.

[00:40:18] Loris Marini: And filter out terms.

[00:40:19] FX Nicolas: What is the reason? They didn't know the reason because they didn't ever track the governance.

[00:40:24] Loris Marini: That’s very interesting. You raised a number of points.

I guess one thought that came to mind when you were talking about interoperability and agreeing on protocols is examples of immense projects that we successfully completed as a human kind.

Think about the way we figured out that are black holes colliding and gravitational waves, building super sensitive interferometers, the LIGO project. Thousands of people contributing across the globe.

You don’t even have to think far, the International Space Station, you can see it almost with the naked eye. That's an example of collaboration across many, many countries. So, it is possible. We know how to collaborate and define protocols and do crazy, complicated stuff.

But when it comes to data in our companies and our organizations, we don't think that way. There has to be that culture of thinking that these problems are not easy to solve. The reason why we are where we are is because this stuff is hard and only a few percent of companies actually do it really, really well.

One, you have to agree that it’s hard. The second one is let's try to implement some of the best practices. And the third element is most important and that’s to have that curiosity mindset. Every business has nuances. Every team has different ways, different checks and balances. So, you really have to do the work.

A vision for the next decade

How do you see the field going into the next decade? What are you excited about? What do you think is going to be hard? And what do you think you can't wait to happen as soon as possible?

[00:42:36] FX Nicolas: Frankly, it's hard giving you any prediction in the long run because there are so many in the data space. There are so many new trends and technologies popping up. There is one thing that I believe is going to be a trend and that's basically the market trend is going to be consolidation because we have had an explosion of terms and tools and a market niche in the data market.

And it seems to me like every single need or requirement or use case has spun up into a new product. We have a galaxy of different products. You cannot be truly agile if you multiply tools too much, simply because at the end of the day, if these two teams know how to speak together, but the systems that they're using do not speak together then they’re in trouble. Because then the people who are willing to work together have got tools that make it harder.

You walk into a bar, you see this beautiful person, and you want to have a chat with that person and you don't speak the same language, too bad.

[00:44:04] Loris Marini: Yeah. You spend an hour looking at each other.

[00:44:06] FX Nicolas: So you will leave the person at the bar. And you will never speak again. If you have something in common, like a language you speak together, then the discussion is going to be smooth.  And it's exactly the same thing. So, the problem is that this over segmentation of products is leading to a lot of product fatigue, because it's difficult from the technical perspective for the governance to work with the enforcement and then vice versa. At the end of the day, the discussion won't be pleasant.

And I would say that at a certain point, people will have to choose between things that work together and the best things that are only able to work alone.

[00:44:59] Loris Marini: We're almost going full circle. We started with SAP, with Oracle systems, then we fragmented into different apps.

[00:45:14] FX Nicolas: You know, that's something that amazes me about IT. I've been working with computers since I was 12. And the thing is that we like in enterprise, it is a revolution. You know what a revolution is?

[00:45:34] Loris Marini: No, what is it?

[00:45:35] FX Nicolas: You go in a large circle and then you come back.

[00:45:39] Loris Marini: Yeah.

[00:45:41] FX Nicolas: A large circle and then you come back to your starting point. But it's a long, it's a long circle. We'll revolve a lot. We like to do these things. At the end of the day, we usually end up at a different position, but in computers, we love to do this type of revolution.

This whole trend of creating a lot of new stuff, spitting out a lot of new stuff, I believe that's probably a way for us to innovate.

[00:46:17] Loris Marini: Yes, for sure.

[00:46:18] FX Nicolas: You take this little piece that you think maybe I can do it better. You spin it out. You create something that's 10 times better because you're alone with that little piece and the whole thing explodes and then collides again.

[00:46:31] Loris Marini: Collied into something that is way, way better than the best you could've come up with without that process. Yeah. I love it.

I think I'm going to leave it at that. Perfect. FX, thank you so much for your time and for being here for The Data Project. I'm really looking forward to catching up. Hopefully the borders all will open and we'll be able to fly and perhaps work and see each other at some conference.

[00:46:56] FX Nicolas: Thank you.

‍