Playing To Win

Tackling Mysteries

How to Avoid Abandoning the Top of the Funnel

Roger Martin
7 min readOct 23, 2023


Source: Roger L. Martin, 2023

I have been pleasantly surprised by the great reception for my last two pieces, Heuristics, Management & Strategy and Ways of Understanding. I worried that they were too esoteric — but apparently not so. There has been much great commentary and many questions, including one from the always thoughtful reader, Dan. I have decided to continue what has become a de facto mini-series with my 47th Year III Playing to Win/Practitioner Insights piece: Tackling Mysteries: How to Avoid Abandoning the Top of the Funnel. You can find the previous 157 PTW/PI here.

The Broader Question

In the previous two pieces, I focused how the approach to knowledge development in the business academy produces a bias toward the refinement of heuristics and away from the tacking of mysteries. Readers were very quick to draw a parallel to biases within corporations — and adopting my metaphor of a funnel as their bases.

Some commented on the parallel with the modern tendency of corporations to spend resources on ‘performance marketing’ rather than ‘brand marketing.’ That is, they spend advertising dollars on attempting to convert customers already in the sales funnel with online advertising for which they pay only for specific performance (typically a click), at the expense of bringing new customers into the top of the sales funnel with brand advertising. (Interestingly, my co-authors and I talk about that very tradeoff in our forthcoming Harvard Business Review article (Jan-Feb 2024).)

The aforementioned Dan analogized to product development and the tendency to work down-the-funnel on incremental improvements, based on collection of extensive quantitative customer data, rather than on up-the-funnel on breakthroughs. I quote his point of view here: “While this can be very effective in driving incremental improvements and/or achieving certainty with respect to a specific concept/question, it’s unlikely to drive breakthrough products and risks producing products where you give customers what they want, which is not actually what they want…So, I’d be interested in your thoughts on how an organization can nurture great product development, without engendering total anarchy. I.e., how do enable breakthrough innovation while also delivering the sort of predicable results investors demand.”

I think there are three components of the necessary approach to tackle mysteries and avoid abandoning the top of the funnel in business.

A Different Relationship with Data

The tackling of mysteries requires a different approach to data. In fact, I think of it as a different relationship with data. The current relationship is largely passive and servile. We think of data as something that exists out in the world that we should collect, analyze, and conclude on its basis. Since, as I have pointed out numerous times, the only data that exists is (at the point of its analysis) in the past, collecting and analyzing it can’t tell you anything about the future — only what exists today. Yet the entire business world is taught to collect and analyze data to determine what to do in the future.

But you can’t tackle a business mystery based on this relationship with data. You can only hone and refine your understanding of a heuristic, whereby you already have a solid understanding of a given phenomenon — e.g., customers like these five attributes of products in this category and let’s try to better understand which they put first versus second, etc. The data will not tell you how they would react to something they have never been offered and that you had never conceptualized.

To tackle mysteries, you must nurture an active steering relationship with data. The task is to create data — that is to take steps that produce new data to guide your thinking. As I have argued before, this is why the key question is not What is True? Rather, it is What Would Have to be True (WWHTBT)? To create a breakthrough, we must make something true that isn’t true now. If we are cowed by what is true, we will never create something that isn’t true now.

That is why rapid iterative prototyping (as popularized by IDEO co-founders, designers David Kelley and Bill Moggeridge) is so important. It helps you create data — and to build your confidence as you create data. I always argue that the bad thing about the next (say) six months is that there is no data about it yet, but the good thing is that in six months, there will be lots of data about it. That is why being rigorous about building future data is key to tackling mysteries. You can’t just do things and then observe. You have to put forward a theory that says: if we do the following new thing, we will expect this reaction. Then you can see whether the data that you have created is consistent or inconsistent with your theory.

That is an active steering relationship with data, not a passive, servile one. And it is necessary for tackling mysteries.

By the way, the late and terrifically great Clay Christensen had analogous thoughts about the importance of creating, not just collecting data. We even had discussions about writing a book together on the subject. Sadly, due to his untimely passing, that isn’t going to happen.

A Fine Balance

A second requirement is a fine balance between exploitation and exploration, as I have written about before. No company can spend all its energy and resources at the top of the funnel for long. Mysteries unravel when they are ready to unravel — not when a company orders them to do so. That is why so much mystery work is done in non-profit environments.

Even Steve Jobs found this out — the hard way. He was fired from the company he founded because he was spending too much of its resources exploring mysteries and not enough on exploiting heuristics. However, firing him actually didn’t solve the problem because Apple was within one quarter of running out of cash when Jobs was brought back to save it.

I would argue that Jobs learned his lesson well and in his second stint as Apple CEO — during which time he created immeasurably more value — he balanced exploration with exploitation. He made sure that the exploitation machine was consistently spinning off the resources necessary to continue to explore — a sea change from his first tenure as CEO.

The price of exploration is pushing what you now know further down the knowledge funnel toward algorithm and code so that those activities (whether performance marketing or product refinement) can spin off the resources to tackle the next mystery, the solution to which you will be exploiting some years later.

Desirous Leadership

Those first two things — an active relationship with data and a balance between exploitation and exploration — will only happen if leaders are desirous of tackling mysteries. Most aren’t at all keen. Most love to have a passive relationship with data and live in the world of exploitation — honing and refining what they have been given. That is, in fact, the route to the top of most widely held, publicly traded companies today. They make the analysts happy with their consistency — until it becomes consistently drab.

Fortunately, a minority of leaders focus on making true what is currently not and keeping a productive balance between exploitation and exploration — and their organizations do extraordinary things for the world. Some of them are famous for doing so, and I have had the pleasure of watching from close up, like AG Lafley of P&G, Jørgen Vig Knudstorp of Lego, and Issy Sharp of Four Seasons. But I have watched others do so similarly without as much fanfare including Ken Thomson of Thomson/Thomson Reuters, Ted Rogers of Rogers Communications, Jim Hackett of Steelcase (and Ford), Steve Voorhees of WestRock, and Michael Kneeland of United Rentals.

All experienced spectacular company success because they were desirous of tackling the mysteries of their industries and were completely unsatisfied with honing and refining what they inherited.

Practitioner Insights

Regardless of your level, you have agency over your relationship with data and the degree to which you balance exploitation and exploration. You can spend your time collecting the data that exists or on creating new data to shape the future. You can spend all your time doing only the things for which you are already expert or spend some of your time on new things that you don’t yet know how to do.

It is a choice — as is everything in strategy. In this case, we know that the opposite isn’t stupid on its facebecause most people in business do the opposite. That is what they are trained to do in their formal education and encouraged to do in their business environment. It won’t be an easy choice — no good strategy choice is — because you will be rewarded more consistently in the near-term for analyzing data and making incremental improvements. But you will be selling yourself short.

Remember that you are the master of your own domain. You can operate it in a way that optimizes performance in it. You don’t need permission to do that. That is where you should start.

It brings to mind an executive with whom I have worked and am highly impressed. She runs a business but is not the CEO of the whole company. She inherited two big brands that had gone into significant decline and was determined to reverse that. She quickly tired of the incremental ideas that arose out of her organization and declared that she would henceforth refuse to ratify any recommendation or decision that was supported solely with quantitative data. She wanted to see thinking from her organization that was powered by imagination, by trying new things that generated new data, both quantitative and qualitative. She knew that to turnaround her brands, she needed to create the future, not just optimize the present. And she has. Several years later, her brands are thriving because she insisted on a different relationship with data and on balancing exploration and exploitation. Importantly, she didn’t ask the permission of her boss. She just did it — and to be sure, her boss didn’t complain!

Be like her! Don’t start by attempting to change anything outside your domain. And don’t start by asking anyone’s permission. Like her, you don’t need permission to be effective. You just need a bit of courage to be different.



Roger Martin

Professor Roger Martin is a writer, strategy advisor and in 2017 was named the #1 management thinker in world. He is also former Dean of the Rotman School.