by Rita McGrath
A final pivot was to work with advertising agencies to place ads against relevant web content. The product Invite Media eventually sold to Google was called Bid Manager, and its main benefits were to allow agencies and media buyers to buy more efficiently across exchanges. Turner and Weinberg thus ended up launching the first universal buying platform for display media. That has subsequently become a critical part of the advertising-supported ecosystem that pays for the Internet as we know it today.
In a controversial decision, Google bought the three-year-old company in 2010 for $81 million and brought its two cofounders on board to integrate their technology with its DoubleClick offering. The two knew that this wasn’t going to be a long-term gig, however, and they started scouting around for new business ideas that could be their next opportunity.
With the demonstrated success of Invite Media behind them, they had credibility, access to funding sources, and an attractive lure for future employees. For ideas, Turner and Weinberg engaged in angel investing. As Turner later said, “It was selfish, I admit. Zach and I learned by being involved with the entrepreneurs in these startups and we got a lot of insight into the health care industry.” Notice again the behavior of making small investments to open up new perspectives—but not committing to a single point of view about the future. And that led to Flatiron.
From Advertising Software to . . . Healthcare?
Actually, the real impetus for the firm that became Flatiron Health was personal. Turner’s younger cousin was diagnosed with leukemia. The boy’s father said to Turner and Weinberg, “A few hundred kids get this every year—what drugs do they get, and do they work—I can’t find any information.” Misdiagnoses, lost time due to travel, and a host of other problems with his cousin’s care convinced them that automating information flows in the cancer care system was a big, important problem that would be potentially worthwhile to address.
That simple question from his cousin’s father caused the duo to drop all the other ideas they were flirting with in the healthcare space (insurance, medical malpractice, and others) and to focus on the idea of putting together a platform that could provide a unified view of all data flowing through the treatment system for cancer patients. As Turner said later, one of the characteristics that both he and Weinberg share is a relentless curiosity: “We ask a million questions and we want to know: why is it that way?”
The insight that sparked Flatiron was that by bringing together two worlds—digital capabilities and medical training—they could create a different perspective on the entire treatment experience. The two entrepreneurs went through a period of being “obsessed” with this idea, following doctors around, meeting with “20 people a day” and trying to learn as much as they possibly could. You can see the emphasis on learning in the guidance Turner offers to other would-be entrepreneurs:
Turner recommends meeting with as many people in the industry as you can (Turner and Weinberg met at least 500 people before getting Flatiron off the ground). “Take notes and pitch ideas and prototypes and get your idea in front of a lot of people who know their stuff—physicians, hospital administrators, insurance companies and clinics—to get their feedback early.”
As they discovered, you are unlikely to find a solution to a software problem by relying on people steeped in a medical background. What Flatiron has now created is the ability to take in both structured and unstructured data from community oncology offices. It can use that data to perform sophisticated analyses in order to better determine a course of treatment for a given patient, as well as to find larger patterns in the information.
Flatiron is one of the beneficiaries of regulatory changes on the part of the FDA that allow data extracted from electronic health records to be used in support of clinical trials. As FDA commissioner Dr. Scott Gottlieb said in a 2019 speech, “Digital technologies are one of the most promising tools we have for making health care more efficient and more patient-focused. This isn’t an indictment of the randomized controlled trial. Far from it. It’s a recognition that new approaches and new technologies can help expand the sources of evidence that we can use to make more reliable treatment decisions.”
Another major issue that concerns the FDA is that clinical trials are often not representative of the population, potentially distorting the results that come from them. With Flatiron’s technology, smaller and more tailored approaches using vastly more information than was typically available are now possible. The company, which was acquired by Roche in 2018, will be pursuing the analysis of what it calls “regulatory-grade” information. For example, in 2018 Pfizer and Flatiron presented evidence at a breast cancer conference that data from a cohort of Flatiron patients matched data from a group of patients in the control group—meaning those who had not received treatment—in an advanced study. The intriguing idea is that if reliable data reflecting what would happen to a group not treated with a promising therapy could be obtained from a database, perhaps it would be unnecessary to assign patients randomly to receive treatments or not. This would, for example, allow all trial participants to receive promising therapies and reduce the cost of conducting such trials.
How Entrepreneurs See Around Corners
Turner and Weinberg’s story has many similarities with those of other successful entrepreneurs who are also able to “see around corners.” In particular, over the years I’ve studied what are sometimes called “habitual” entrepreneurs—people who have started many businesses. The reason they are so valuable from a research point of view is that it is highly unlikely, if they’ve done it over and over again, that their success has been a result of mere good luck. Instead, what we find with this population is that they have a set of methodologies they use to gather lots of information, detect patterns, test assumptions, and bring together resources.
They have vast and nonredundant networks that they can turn to in order to generate ideas and solutions. My colleague Ian MacMillan calls this “webbing”: habitual entrepreneurs are tireless in connecting with others, particularly those who do not overlap with their own knowledge. They are also incredibly curious. It is such an ingrained part of their personality that they may not even be conscious of it—they are simply really interested in why things work the way they do. They are also resourceful. As Mac is fond of saying, they “spend their imagination” to validate assumptions rather than buying their way into an answer. They also do things quickly and are unafraid to change direction when new information comes in. They look for patterns in what they see. Are there underserved segments? Unmet needs? Places in which key stakeholders are working around a process because it doesn’t work? Are there areas of surplus? Scarcity?
Consider serial entrepreneur Steve Blank, a colleague of mine at Columbia and a legend in the startup world. He’s guided four companies to IPO status and mentored many more. To find opportunities, Steve says, an entrepreneur has to be endlessly curious and to recognize patterns that no one else does, by, as he puts it, “showing up.” To validate ideas, he advocates “getting out of the building” and learning how a potential innovation might change a customer’s life. He calls this “customer development.” After taking an idea to market, in what Steve calls the “build” stage, he recognizes that new information always requires adaptive behavior. And finally, he realizes that what a serial entrepreneur enjoys is starting businesses, not necessarily running them once a scalable, repeatable model has been discovered. But such entrepreneurs can leave behind immense amounts of value once they have moved on. As Steve put it in 2012:
I loved everything until it got big. My job was great from search to the beginning of build. When it started to feel like I became the HR department, then it was time to go. In my second to last company I first lost $35 million, then raised $12 million and eventually returned over a billion to each investor. Only in entrepreneurial clusters is there a special word for failure like this: It’s called “experienced.”
Creating a plan for fast learning is something successful ser
ial entrepreneurs do almost by instinct. The rest of us can learn to do the same. It does, however, require a different mindset than we bring to the planning situations of business as usual.
“Fall in Love with the Problem . . . Not with a Particular Solution”
I first heard this saying, which has popped up in any number of stories, from Kaaren Hanson, who at the time was vice president of design innovation at Intuit, when she was speaking at a Columbia Business School BRITE Conference. Although many seem to recognize its wisdom, it is very hard to remember it or implement it in an organization that is focused on planning in the conventional or traditional way.
To that end, a mistake I often see people in organizations make is to link, in their minds, the outcome they are seeking with a particular solution. It is entirely possible to have appropriately identified a desirable outcome, but to be completely wrong about the best way to get there.
For example, Procter & Gamble initially attempted to commercialize PuR, a chemical that allows users to create safe drinking water from dirty water. The chemical worked as advertised, but it failed as a commercial product because the target customer was not conditioned to trade off scarce resources for cleaner water. It also required behavioral change, which meant commercial uptake of the product was painfully slow. The company considered canceling the product altogether in 2004.
Greg Allgood, a senior P&G executive at the time, was a huge believer in PuR and sought a different path to making it available to the people who needed it. Coincidentally, right around this time, a devastating tsunami hit Southeast Asia, and P&G donated over $3 million in aid, including 13 million packets of PuR. This experience, according to Allgood, was a turning point in the business model for the product. Instead of trying to make it a commercial consumer offering, the company made it a social venture.
P&G partnered with international aid organizations that paid for the product on a cost-recovery basis and distributed it for free. With this model, P&G also realized that the program was a public relations bonanza, that it could help them understand consumers in emerging markets, and that it gave them a position of strength in discussions with governments, NGOs, and other strategic partners. P&G eventually sold off its other water purification initiatives but kept hold of (and branded) PuR. This illustrates the idea that you can get the basic “job to be done” part of the arena right without first figuring out what the best solution might be. This is a classic case of seeing around the corner, and doing so due to unusual and unexpected circumstances.
Define What Success Looks Like
To begin a discovery-driven planning exercise, articulate what would make a particular initiative worthwhile. This could be couched in terms of finances and numbers. It could also be couched in terms of opening opportunities or expanding the reach of an organization. As we saw with the Flatiron Health founders, they were looking for a large problem that had not yet been adequately solved, essentially following Nassim Nicholas Taleb’s guidance that when uncertainty increases the upside of an opportunity, it can create valuable opportunities.
The next step is to specify the benchmarks that suggest whether the initiative is realistic or not in terms of key comparisons. Then spell out how, operationally, things would need to be done to make it a reality. As you do this, you’re going to find yourself making a great many assumptions. Write them down and then think of how you might validate or, even better, invalidate the assumptions.
The practice that brings this all together—and that differentiates a discovery-driven plan from a conventional one—is planning around critical learning moments, which I call “checkpoints.” A checkpoint can be a naturally occurring event (the regulation passes or it doesn’t). It can be part of an experiment. Whatever the case, quickly moving through these checkpoints is the key to mobilizing an organization that is facing potential strategic inflection points.
Discovery-driven planning is well suited for those situations in which the change (potentially ushered in by an inflection point) increases the uncertainty that decision-makers face. It can provide structure, discipline, and thoughtful resource utilization.
Discovery-Driven Thinking in an Entirely Different Setting
Lest you think that the methodology applies only to high-tech startups and other big business initiatives, let’s examine the thought process by considering the evolution of a real project that I encountered on Kickstarter.
The toy industry has been through its share of inflection points, with hundreds of retailers closing their doors as toys increasingly were sold by Target, Walmart, and other large retailers; children aged out of toys earlier; and screen-based entertainment began squeezing out traditional toys. Despite strong growth fueled by a number of blockbuster movies and tie-ins to toys, Toys “R” Us, the largest dedicated retailer in the sector, shut down its US operations in 2018.
Further, the digital revolution has not bypassed toys. Many innovative startups (and some incumbents, such as Mattel) are adding a digital component to their toys.
So let us consider one product (it’s real, and on Kickstarter) whose team is trying to benefit from the collision of physical toy properties and advanced digital intelligence. It is called Octobo, produced by Thinker-Tinker. Octobo is a physical plush toy crammed full of sensors that allow it to respond to touch and movement. Its intelligence comes from a tablet inserted inside the toy.
The vision of its creators was that it would provide an educational and interactive experience for kids (and parents), and because the intelligence behind the device is tablet-based, the toy can “grow and develop” with children, bringing about essentially an upgradable object.
The project won an OpenIDEO award, and its evolution is outlined by Thinker-Tinker’s founder, Yuting Su, in the documentation submitted for the prize. Calling Octobo a “plush companion robot,” Su says she was inspired to start the business when she was pregnant with her own child and initially developed the plan as part of her master’s thesis for her degree in game and interactive media design from the University of Southern California in 2015.
She describes the problem she was trying to solve as follows:
Toys and learning tools today fail to engage young children to learn in a safe and enticing way. Although children develop rapidly, many times toys cannot keep up with a child’s learning curve and have only limited replayable value. Most toys that integrate technology today are actually detrimental to the amount of time a child spends with their family, with other children, or even reading a physical book. This and the ability of Octobo to grow with the child are examples of digital’s ability to fix these problems.
What Su intuited, and the corner she was seeing around, was a gap. Conventional toys are used for one or two things only, and children can quickly tire or grow out of them. Digital interfaces, on the other hand, aren’t suitable for very young children, aren’t interactive, and just encourage more passive entertainment. She recognized that parents could become enthusiastic about a toy that promoted interaction and learning rather than the addictive game playing of conventional digital apps. Her eureka moment was the inspiration to conceive of a way to create a toy that had all the cuddliness of a plush stuffed animal, but with the engaging interactivity of a tablet computer.
Su observed that as digital devices become unavoidable, children aren’t being given enough opportunities to play in the physical world, which is essential for motor development and other skills. “Physical play trains us and helps us grow,” she said in a 2018 interview. “Our hand-eye coordination. The training of our littlest muscles. The different textures actually educate our bodies by helping us sense surfaces. We learn not just on an intellectual level but our bodies learn through tactile objects.” While it is still early days for Octobo and Thinker-Tinker (Su described it in the interview as a “roller-coaster”), the combination of digital and physical play is likely to create a significant inflection point for the conventional toy business.
Among the checkpoints the Octobo
venture went through were:
A Facebook ad campaign to collect data about potential users and customers
Identifying potential manufacturers for the scaling of the project
Realizing that content development could proceed faster by going beyond in-house production to connect with communities of content creators
Identifying educators to both provide content expertise and evaluate the design of the product
Identifying key metrics for success and how they might be measured
Lots of interactions with the prototype product
Each of these checkpoints represents a moment in time at which critical assumptions made by the project team could be tested, with the learning being used to inform subsequent choices. As of this writing in early 2019, the Kickstarter campaign has nearly reached its goal, with the next major uncertainty for the team being how they will scale up production capacity while maintaining the level of quality the product requires. The toy itself has received many enthusiastic reviews and is in prelaunch.
Beyond Innovation Theater
To benefit from the upside of an inflection point that opens real opportunities, organizational leaders need to develop a continuous flow of innovations to replace advantages that have faded away or to respond to challenges that mean that old business constraints have changed in meaningful ways. Unfortunately, many organizations are still dealing with “innovation theater,” in which a lot of lip service is paid but the process itself suffers from a lack of rigor, corporate-level practices, metrics, and other essential elements that would make innovation a proficiency.