Understanding the World: The Power of Data

Transcript

Understanding the World: The Power of Data

Soundbites:

Tom (00:03):

In our data-driven world, using data correctly is a competitive advantage for businesses.

(00:09):

It's about the speed in decision-making because fast companies win. Governments also have to leverage data to inform policy decisions.

Clare (00:17):

If we started using that, Rachel Reeves would be able to get information today on all the transactions that took place yesterday,

Tom (00:26):

But Some complexities exist. We need to spend a lot more time on making sure that the data we are producing is a lot cleaner.

I'm Tom Parker and welcome to the next five podcasts, brought to you by the FT Partner Studio. In this series, we ask industry experts about how their world will change in the next five years and the impact it will have on our day-to-day lives. In this episode, we explore the importance of data to our modern world, to the future of business and to the strength of our political and economic systems. We'll discuss the need to see data, to understand it, to believe it, [00:01:00] and to take action on it. If money makes the world go round, then data tells you how fast it's spinning and when it might stop. 90% of all data was generated in the last two years and every two years the volume of data doubles with 11 billion devices connected to the internet today, the annual global data generation in 2025 is expected to be 181 zetabytes.

That's 181 trillion [00:01:30] gigabytes. To put it in context, we use about three gigabytes to stream Netflix in high definition for an hour. In this digital world, data is king and one must understand data to receive the key to the kingdom. And a kingdom it most certainly is the big data industry is expected to be worth 229 billion in 2025. That's just 5% shy of Qatar's GDP this year. [00:02:00] In other words, the modern world can't live without data nor can business. In today's business world decisions are complex. Analysing disparate smaller data sets takes time and could miss the subtle correlations the big data visualisation can afford.

Alex (02:17):

When we started at Verity, we needed to educate the market a lot.

Tom (02:20):

This is Alexander Gelsberg, co-founder and CEO of Verity, a marketing data platform.

Alex (02:26):

We need to educate the market a lot about how important's to have [00:02:30] all the data business ready at your fingertips and really base decisions on data and insights the market change to now there is a clear understanding of that organisations need to have a data ownership strategy that they need to have this data. Of course, all these different organisations are at the different stage of a maturity curve in terms of where they are in the expertise and discipline of leveraging the data and creating insights. But the market has changed. The awareness is there, nobody disputes the relevance of it. [00:03:00] Sure, everybody's maybe a bit more advanced or has a different aspect to it, but now the importance and the technology, the tool stack is there to really work with data and put data to work.

Tom (03:13):

Data from every business function is required to get a clearer picture on future growth and having accurate information at your fingertips to pivot your organisation is a competitive advantage

Alex (03:25):

If you as a business want to accelerate and to be ahead of your competition and data [00:03:30] enables you to have that competitive edge by being data driven and being faster than the competition. If you leverage data the right way, if you have a data analytics discipline, if you have strong data democratisation in your organisation, you will ultimately be faster. And if you automate the right aspects of it, get the right tools that in place you obviously create the insights needed for your business. So insights can be in the analytical aspect if it depends on the question. You as an organisation need to be [00:04:00] able to ask the right question and then the analytical discipline will help you to find the answers and the questions are different from organisation to organisation and obviously the questions are different in terms of how sophisticated you are. And so answering this question can be in terms of sporting a shift in consumer behaviour when suddenly covid hits and swimming pools are much more important than booking a plane trip to the coast.

(04:25):

So this sport of shifting consumer behaviour super early is an example. [00:04:30] Knowing where you invest to reach new audiences and new customers is only possible if you put the right analytical framework in place to analyse the investment and the yield being generated. And that can say a few big percent of your marketing budget and create more impact at the end of the day. And the other example made was really identifying new products and new customer opportunities. So it ranges really from cost saving, time saving, where you put your precious resources to better [00:05:00] use and it ranges to seizing new revenue opportunities by potentially placing a new product or marketing a different product to a different audience. So I think at the end of the day it's all about speed. It's about the speed in decision-making because fast companies win.

Tom (05:17):

Speed is key as is the way you see data gone are the single static format bar charts of old that advised boards on purchasing or investment decisions. Now every part of the business, including [00:05:30] the C-suite, must have real time interactive data sets to make informed decisions

Florian (05:36):

A city in the marketing function, we look after a brand, we look after the way we engage with consumers.

Tom (05:42):

This is Florian Jackie, global head of data consumer engagement at Nestle.

Florian (05:47):

First we always refer to marketing as a not in the science and the science bit for us is mainly the data. Part of it is how much we can leverage data to be more strategic, be more predictive, and then be more real time as well in the decision we are making. [00:06:00] So as an example, we use a lot of econometrics modelling to understand the output of all of the marketing activities. And the traditional way of doing it is you spend a lot of time collecting the data and then you have data scientists working on building the model and that model will explain where to invest why, and try to have the right simulation. Right now with machine learning, we are able to just update the data in the background and then the model will recalculate automatically. So a lot of that technology is helping us [00:06:30] data faster and building more data pipelines to always refresh.

(06:35):

In the past, I think if I look at it, you typically have business executive or brand manager that was really good business people with a lot of good business sense and experience and they were just using a lot of that to understand like, well now this year I'm going to put 2 million or 3 million in TV and that's how I'm going to grow my business. So I think the change here is really about how we can go from experience to intuition plus insight coming from that data. And then I think [00:07:00] the biggest change that we have now is because we have this awareness that this is what we can unlock in terms of decision. If you ever go back to the beginning of the value chain, we need to spend a lot more time on making sure that the data we are producing is a lot cleaner so we can make sure that there is no question around, well is the data is right or good or we can basically trust it. Is the data right? Can you check that first? I think that's a big, big change as well.

Tom (07:24):

The clearer the picture, the better informed the business and better visualisation requires clean [00:07:30] and complete data, then it's what you do with that data to drive the business in the right direction. Peter Sonder guard, SVP at Gartner said in 2011, information is the oil of the 21st century and analytics is the combustion engine.

Alex (07:47):

I agree with the broadest aspect of the picture he wanted to paint. If I'm not wrong, I think it's about 13 years ago, more or less he made the statement time flies. I would describe it a bit differently. I would say information [00:08:00] is rather the energy in its purest form like an electricity where analytics is basically the light bulb, shedding light and obviously powered by electricity. This is how I would describe it a bit more today, but at the end of the day, it's a raw energy form that needs to be put to use and can create a lot of power. But same in analytics. It doesn't need to be complicated like a potentially combustible engine. For many it can start simple and already create a lot of impact. [00:08:30] But ultimately the combustion engine example, what does it do? It creates torque, it creates power and that inside creates power. And if you take this whole argument and perspective from the other side, if you don't do it, you're blind, you don't see anything, you're acting blind. And this is why I like the light bulb example a lot as well. If you are not analytics driven, you're just simply flying blind.

Tom (08:55):

As with businesses, governments can't fly blind with all the data [00:09:00] available. Policymakers looking to address ESG challenges for example require fast evolving and easy to understand data-driven insight, if you want to save the rainforests, you need to see the wood from the trees. Are Conan Doyle's famous detective Sherlock Holmes mused that it is a capital mistake to theorise before one has data, data facts and truth inform ideologies. Therefore a flourishing society and the political administration of one [00:09:30] will rely on how well we see learn from and action data

Clare (09:35):

Governments are relatively mature with their use of data.

Tom (09:38):

This is Dr. Claire Walsh, director of education at the Institute of Analytics.

Clare (09:44):

They are well on top of issues around quality and making ambitious assumptions based on what they find. It's really dangerous for you to find a pattern and then extrapolate all sorts of things from that. What you see is all [00:10:00] there is. They've also been really good at sharing their data publicly since around the Tony Blair years, most governments have had the policy of open data sharing and that's brilliant because something like weather data or traffic data adds billions to the economy just by making that data freely available to anyone who wants it. So in terms of promoting business governments, they know what they need to do in terms of using data to make decisions and inform policy. [00:10:30] They are pretty good, but then they have the same kind of limitations that we have to gather and collect accurate data. I have a very unpopular opinion that if we move to digital currency, we would be in a whole new world.

(10:44):

So digital currency is not the same as the money in your bank account. You have to have a bank, you have to have an agreement, a contractual agreement with a bank. There is another form of currency that is entirely digital run by the government. It's aligned to the UK [00:11:00] sterling value. It's completely under their control. So it's not Bitcoin, it is not all over the place. If we started using that, Rachel Reeves would be able to get information today on all the transactions that took place yesterday and she would have an incredibly complete view of where we were spending, how we were spending, what we were doing. Obviously it's very unpopular because it's data privacy and there's also the question, would it make things better? I don't know. I [00:11:30] honestly don't know. Mitchell Weaves would not be able to look at a data set this size on her laptop.

(11:35):

She'd have to be in a physical room, probably a chamber. Let's convert one of those old ballrooms with screens all around that would give her the kind of overview that we need. And when you have a huge data set like that, then that level of visualisation that size is really useful actually. And of course you can drill down into it. Then there's the assumption, well, she's going to make [00:12:00] better decisions on the basis of the insights she gets. Maybe she would make reactive decisions and that would be a problem and she would ignore long term policy. We don't know until we try

Tom (12:13):

Data policy governance, transparency and use of data is a critical area for governments, businesses and consumers, one that will require ever more scrutiny as we move further into a digital data-driven future

Clare (12:27):

Data policy is really important. It's quite [00:12:30] challenging. So many laws touch on what we do with data. We've got some very obvious new ones that have come in, like the EU AI Act or Colorado has also passed an AI act this year. A lot of those bigger apps are aimed more at governments, for example, to make sure they're not overreaching in their policing strategies or how they run their elections. A lot of them are aimed at [00:13:00] the developers of these foundation models, a foundation model. You and I are not building one, not those. A tiny group of people in the world are building one of those, so we don't really need to worry too much about it. But one of the great things about these acts is that they are encouraging more transparency, which is really important when you as a business buy a third party algorithm and if you're sensible, you'll be buying it third party, you won't be creating it yourself. [00:13:30] You have to ask so many questions and in the past these companies perhaps might have hidden behind international property. I can't tell you what I did, I can't tell you what data I used. It's just not right. You need to know that information. So I'm really hoping we're entering a new era where we have much more communication between third party suppliers and the businesses and we can all start making better decisions.

Florian (13:56):

So I see it's in the marketing environment and marketing [00:14:00] is about collecting brands to consumers. So a lot of the focus on policy for us is relevant consumer privacy. We have been very proactive endlessly to go after consumer engagement and use their personal data to serve them better, to create better experiences. But it comes with the responsibility of using the data responsibility transparently consented with the right value exchange. So it's a key one for us. I think we're spending a lot of time and effort with our legal team as well, but to make sure that the [00:14:30] practice and the way we use consumer data is very transparent and always consented because that's for us the biggest reality at least in my space. That's where there's a lot of focus, both because that's one of our principles, but also because there's also regulations that are changing very, very often. So we need to make sure that we keep up to date with them.

Tom (14:48):

Being able to see in real time insights from disparate systems correlated together is the power of big data visualisation, but understanding the consequences is important. Subtle correlations may [00:15:00] exist that aren't at all meaningful when given human context from 2004 to 2019, the number of Google searches for the term how to build a lightsaber directly correlated to the number of pest control workers in the District of Columbia. It would be unwise to suggest either had an effect on the other. If data is absent of careful supervision, then it is more a problem than a solution. But this isn't the only challenge in a data-centric world.

Alex (15:30):

[00:15:30] Well, there are obviously many different aspects to challenges when it comes to data, not just in terms of sheer amount, in terms of velocity, et cetera. It's a full data value chain in an aspect. There are challenges in terms of having a strong technology bedrock in place that enables accuracy, quality, recency, et cetera, enables the different business teams to really become data driven and utilise the data for the [00:16:00] decisions complexity. But then there are obviously also challenges in terms of definitions and standardisation again that everybody bases their decisions on the same set of KPIs, on the same updated recency of data, et cetera. So it's a multidimensional problem at the end of the day and the issues then you could easily drown in the complexity and in this amount of data. And then one issue happens then from time to time when the pendulum swings too much to the side of, Hey, we [00:16:30] want to be data driven, that everybody asks for more data that nobody dares to make a decision. Just ask for more data, more data. And this is also then a really, really bad symptom. So it is this right balance of having a good tool stack and governance framework in place and really enabling teams to succeed. But at the end of the day, it's all about fast and speed in decision-making

Tom (16:52):

The amount of data companies must manage at speed is always a challenge, but also with such a rapid technological advancement, [00:17:00] we are leaving some data in the dust.

Clare (17:03):

There's a website that I love that's dedicated to spurious correlations and it correlates things like margarine and winning beauty pageants totally spurious. And so we need humans who are looking out for that. I think a more likely thing that we fall for is the problem of extrapolation. And this of course collapsed the entire world economy. Back in 2008. They had all the data they needed to feel [00:17:30] confident that changing mortgage laws would still be manageable, that they would be able to contain any problems and absorb them because of the data that told them that this would happen. But what they didn't take into account was that there was a new variable, the change in their behaviour as a result of having the data and that affected everything that went on after that point. And so extrapolation collapsed the entire global economy. [00:18:00] Another challenge is that we all think that data is going to be there forever.

(18:05):

We all have this idea that data is permanent, but we know that technology changes and with each change, unless somebody is monitoring and maintaining the code behind that dataset, it's going to become very challenging to access it in the future. We know the problems of accessing data in the past from, for example, journalism that [00:18:30] we've lost completely. The Rocky Mount in high newspaper went out of business, and so they stopped maintaining their data set, which contained pulitz, the prize winning journalism, the kind of thing we want to pass on to future generations. And that particular one is a good example because the effort that was needed to revive it and to get it back on the internet was just so much. And so we are entering a new dark [00:19:00] age where we think we are leaving all these records, but they're all digital and they will deteriorate. The machines they're stored in will rust and die. The code will age and will not be accessible anymore. And we've got this false image that it's permanent and even things from 2013 are lost

Tom (19:22):

For large multinational companies. There are other challenges that must be addressed.

Florian (19:26):

So in a company like Nestle, when we look at the complexity [00:19:30] around data, I think we need to start from what Nestle is, which is one of the biggest CPG in the world. Hundred 80 countries is a very, very decentralised culture. So we thousands of skews, sometimes very different. And on the good side, a good degree of ization, especially on ERP or on some of the tool in MarTech for instance. But I think the three big things for us to tackle, and we've been working on it and we are going to continue to work on it, is access. So unlocking access data that are buried into market, into function that we really need to [00:20:00] partner together to make sure that we strategize, we make that available and we create an asset that the company can use multiple time in multiple use cases. The second sterilisation, I'll give you a very simple example.

(20:11):

If I look at a campaign, take a campaign in KitKat, that was kind of the last launch of the KitKat campaign with lot of the rings as a sponsor, it's been called 20 times different. Our three names different across 20 markets. And then you can think, yeah, VI will help me connect back and then make sure that this is all aggregated [00:20:30] into one. And for the realities, I think that's still a very, very simple discipline that whenever we create that in an organisation, if you really want to use it beyond the very small activation in one market, we need to make sure that we have that discipline across the board. It is simple, it's campaign name, it's product name, range brand, et cetera. Things that are very simple all the way to things that are more complex, like calling a format on Facebook platform the same everywhere.

(20:54):

So I can know that this format is actually working much better for this category and not that category. [00:21:00] So that's the second big challenge. And the third one for us is always about utilisation of that data. What do we do with it and what capabilities do we need to bring in the company to produce, maintain, and then use that data to take better decisions? That always be going to be the biggest challenge right now and the biggest race as well in talent because everybody's competing in the same space. So that's one of the key elements for us is getting that access to talent in multiple geography, multiple location is quite complex.

Tom (21:26):

All of this leads to an interesting next five years when it comes to data.

Clare (21:30):

[00:21:30] One of the big things I hope for the next five years is that we get clarity on the legality of these generative AI technologies. We do have a slight problem at the moment that even SAML one has been in deposition of saying, well, yes, we stole the data. We couldn't train it by any other means. What did you expect us to do? We do need a legal decision on that. Historically, this has been decided by judges rather than governments. And so judges will decide that a company [00:22:00] has to do algorithmic disgorgement, which literally means they've got to delete everything, the data, the algorithm that they've spent billions training, and I'm assuming we're not going to do that with these generative AI tools, which have cost so much money and taken years. But we do really need some decisions on that. I think for businesses, they're going to have an amazing five years identifying efficiencies and removing some of the losses in their process.

(22:30):

[00:22:30] Another thing to look forward to I think is a reversal of a pattern. For the last couple of decades of monopolisation we had in Europe a thing called the Digital Markets Act a couple of years ago, and it went over people's heads under the radar, but it does an amazing thing. It regulates data to make it possible for all other smaller companies to interact with the data sets that [00:23:00] some of the large social media companies have. So for example, just this year Apple announced that they're going to allow other companies to sell through an app store on their phones. So what this means is that it's taken away their monopoly over how much money Apple are able to charge to host the app. And it just opens up opportunities. It means that you can't be prevented from selling through these [00:23:30] major gatekeeper, we call them organisations, and it just opens up the possibilities for equity

Alex (23:38):

Out of my perspective. It is impossible to predict what will happen in the next five years. Technology advancements are so drastically accelerating. Nobody would've seen what will happen with LLMs and Gen EI today five years ago. It's almost impossible to do that. But what I am absolutely certain of is that the relevance of data [00:24:00] information will remain to be a key foundational bedrock for everything on top because as we say, data is new oil or data is the energy for the combustion engine that will remain this bedrock. Having everything in place to power your next AI application to really automate, to accelerate, to really increase your speed, that bedrock will need to be in place for all the organisations in the world. So that [00:24:30] will be for sure a constant in the next five years. But great and crazy applications on top of the data will happen.

(24:38):

It's hard to predict, but the most important thing is that the companies have everything in place to seize the opportunities when they are there. But I think the technology will get to a point where the systems and the analytical systems become a bit more goal aware so that actually you use an organisation, teach the analytical systems the [00:25:00] direction, the goal, where you want to go, and the analytics systems will find in the data, the needle, in the haystack in terms of what you can adjust to get there. So that is something that is being worked on and I hope that will bring the impact everybody hopes for in terms of accelerating businesses. So that is for sure some aspect that will have malls on the back of ai, but then beyond that much more obviously will happen.

Florian (25:26):

So if I look at the next five years, for me, there's [00:25:30] two trends that will impact us as a company and specifically the function I represent in marketing. The first one is I think this data for decision support will continue to be a strategic battle. We'll continue to have many digital transformation programmes with a large enabler called data and a lot of data products and capability that constitute a lot of what needs to be deployed at scale to be able to power our organisation and really become what we call a data-driven marketing organisation. That's pretty key for us. The third piece I [00:26:00] think is more on the consumer length. We see a big shift. It's been a big shift already for the last two to three years since Google announced the depreciation of third party cookies. But we are moving from an area of abundance of data collected from you and I in a very grey manner to something which is much more transparent.

(26:17):

So a lot less data and a scarcity of data when it comes to your behaviour online. And what is the value for me to use to better serve you, or better engage with you? I still see a very simple example where you talk [00:26:30] to your wife about buying a new kitchen and then you get an ad that comes there. I think those platforms are still the thing that we feel are completely off from a pure privacy standpoint, but I still see the happening in the landscape and the ecosystem. I think that's one thing I hope for is that some of those kind of creepy strange targeting between even a conversation to an ad is ding. So I think for consumer, it would also get a lot more confidence that their data is there for whether it's by the platform or by anybody who collects them like us

Tom (27:00):

[00:27:00] French writer and lecturer, earnest Diner, author of the Art of Thinking in the 1930s said too often we forget that genius depends on the data within its reach that even Archimedes could not have devised Edison's inventions. This is true of the business setting. Without data, success is limited. Using data to unlock business opportunities is a competitive advantage. [00:27:30] Knowing your data and actioning it correctly will give you an edge. You can't afford not to take data seriously. But as the 20th century economist and statistician w Edwards Deming understood, you can't run a business on visible figures alone. Nothing becomes more important. Deming said, just because you can measure it, it becomes more measurable. That's all. Therefore, to continue to thrive and survive in this digital world, we humans must better understand the data [00:28:00] we are creating. And this means in every part of life be you a consumer, a marketeer, a policymaker, or CEO.

(28:08):

People still need to be an important part of the puzzle. Human led decisions need to be made based on data that we leave behind. A century ago, there was a dearth of data on individuals limited to birth, marriage, and death. Tombstone was the cooling card of their life. Now there is an abundance of data collected upon every interaction we have with the modern digital world. [00:28:30] In five years time, one hopes we will be using understanding and leveraging data to the best of our and our machine's capabilities so that the digital tombstone doesn't read Heal Lies data. And it was sorely missed.