In this Salon Debate, Louis-David Benyayer, Associate Professor of Digital Transformation and AI Initiatives Coordinator at ESCP, and Professor Marie Taillard, Dean of ESCP’s London Campus, consider whether data-driven organisations make worse decisions by over-trusting analytics, the advantages of ignoring data, and why leaders must still place human instinct and understanding at the heart of everything.
The Choice: Let’s begin with the central question. Are data-driven organisations actually making worse decisions by over-trusting analytics?
Louis-David Benyayer: I would add one important word: sometimes. Data-driven organisations sometimes make worse decisions by over-trusting analytics.
My view is not that we should forget about data and analytics. Quite the contrary. I’m a professor in business analytics, so that would be a bit surprising. But there are clear situations where a data-driven approach is probably not the best option.
We tend to think data is neutral and objective – purely facts. But data embeds decisions made by organisations: what to measure, how to measure it, and how frequently to measure it. All these decisions are made by humans during data collection.
If we start by saying there is no such thing as purely objective data, we begin to see why it can sometimes lack relevance. Not because of the data itself, but because of the way it has been collected.
Marie Taillard: I agree there is a caveat. Yes, it is possible to make worse decisions if you over-trust analytics. But we also need analytics. We cannot function anymore as marketers – or, more generally, as organisations – without them.
Analytics gives you the power to really understand many customers and to trust that a trend you are observing is happening, at a large enough scale to invest. But the caveat is exactly that: it is about large numbers. It does not go into the complexity of human behaviour.
Louis-David Benyayer
Associate Professor of Digital Transformation
AI Initiatives Coordinator
Marie Taillard
Dean, ESCP London Campus
Professor of Creativity Marketing
The Choice: So the danger is not data itself, but mistaking the measurable for the meaningful?
L-DB: Right. An organisation is probably not making a bad decision or a worse decision because they collect data, but because they do not collect the right data, or they do not collect it in the right way.
For example, Meta manages Facebook in a very data-driven way. Everything is measured, but there are some hypotheses. Many things are organised to maximise engagement, such as how regularly users interact with each other or with the content. And probably in the minds of people at Meta, engagement is a proxy for value. But it’s well documented that maximising engagement has negative consequences.
No one can blame Meta for making engagement a KPI. But it’s clear that this single KPI doesn’t capture the platform’s full impact. Every proxy is limited.
MT: That is where the human aspect really matters. There’s a well-known story in marketing about how Clay Christensen at Harvard Business School developed his “jobs to be done” theory as a framework for understanding why customers make certain decisions.
He had been asked by McDonald’s to help rescue their milkshake business. His researchers observed that many people bought milkshakes in the morning and took them into their cars. The data, or at least the observation, revealed a pattern. But the insight came when they interviewed those customers and discovered that they had a long drive ahead and wanted breakfast on the go.
That is the kind of insight you don’t get from analytics. You need to understand the functional job – breakfast while driving – as well as the emotional, social and cultural dimensions of customer behaviour. What problem am I solving? How does this help someone feel better, fit in, or reinforce their identity? Those elements will not come from numbers alone.
The Choice: Does data become less useful when organisations face something new?
L-DB: When companies are trying to disrupt markets, innovate radically, or create entirely new categories, they usually don’t rely on data. They’re relying more on vision or intuition.
There was no unmet demand for the iPhone. If you had surveyed people beforehand, they would probably not have described that product. Or consider hotel chains in the 2010s. They were using data very effectively for yield management and automated pricing. Then Airbnb emerged with a completely different view of hospitality. The hotel chains were optimising what they already understood, while missing a much larger shift.
Some of the most valuable business decisions are impossible to justify with existing data.
The Choice: Does generative AI change this discussion?
MT: I think the same principles apply. Generative AI gives you the ability to make broad statements. But you still need to check whether those ideas make sense, and you do that by talking to real people.
The creative aspect of generative AI still relies on large quantities of data. What really matters to customers – what creates value for them, what gets into their hearts and minds – comes from spending time with them. AI does not give you that.
The Choice: If leaders suspect their organisation over-relies on data, what should they do?
MT: Spend time with their customers and get to understand them. So yes, look at the data. But also ask: have we talked to real people? Have we done a reality check? Very often, the small human detail is where differentiation and competitive advantage begin.
L-DB: They need to become clearer about two things: the future they want for the organisation, and the assets, resources and capabilities they have.
Because if you’re clear on those two aspects, it’s easier to identify signal versus noise and determine what to act on. This aligns with what we want to do in the future and what we can do.
It’s not data or-, it’s data and
Louis-David and Marie see data as essential to modern management – this debate was not a rejection of analytics – but both warn against treating it as a complete account of reality.
For Louis-David, the risk lies in proxies, assumptions and signals misread without strategic context. For Marie, it lies in losing touch with the human motives that sit beneath behaviour.
The challenge for business leaders is not whether to trust data. It is about knowing when not to.
Actionable intelligence
Three key takeaways from Louis-David Benyayer and Professor Marie Taillard:
- Good data sharpens decision-making, but over-reliance on analytics can narrow perspective, reinforce flawed assumptions, and blind organisations to emerging change. Organisations should regularly challenge the proxies and KPIs they rely on, asking whether they reflect genuine value or simply what is easiest to measure.
- Effective leaders combine quantitative insight with qualitative understanding. Analytics may reveal patterns at scale, but human observation explains why those patterns exist and whether they matter. There is no substitute for spending more time directly with customers, employees and users to understand behaviour, emotions and unmet needs.
- Perhaps most importantly, historical data is best at explaining what already exists. Breakthrough ideas, disruptive shifts and new categories often emerge from imagination, intuition and interpretation rather than from data alone.