Most boards fail by design, not by talent. Draw the architecture before you fill the seats.
In December 1975, a 24-year-old engineer named Steve Sasson carried a strange machine into a conference room in Rochester, New York. It weighed eight pounds and looked like a toaster with a lens bolted to the front. He pointed it at the people in the room, pressed a button, and asked them to wait. Twenty-three seconds later, a grainy black-and-white image crawled off a cassette tape and onto a television screen.
Sasson had just shown Kodak the first digital camera. The company owned it, patented it, and had a lead of two decades on every rival. The executives asked why anyone would want to look at photographs on a television. Then they sent him back to the lab.
In January 2012, Kodak filed for bankruptcy.
Most people tell this story as a failure of nerve by managers. I read it as a failure of architecture. Through the 1990s, when the choice between film and digital still lay open, the outside directors at Kodak came from heavy manufacturing, traditional finance, and packaged goods. One person held both the chair and the CEO role. Staggered three-year terms slowed any change in who sat at the table. Every director had earned the seat. Nobody had designed the room to ask whether the company should cannibalize its own film business.
Good directors cannot rescue a bad design
Timothy Rowley and Stanislav Shekshnia use the Kodak case in the November-December 2026 issue of Harvard Business Review to expose two design flaws. Owners and CEOs build boards that echo their own instincts, and even strong boards stop changing while the company keeps moving. Their remedy treats the board as a counterweight to whatever force dominates decisions in a given ownership setting.
I take their argument one step further. In this article I lay out five design moves that any chair or nominating committee can make in the next twelve months. My thesis is simple: architecture comes before appointments. A board should be drawn, then staffed, then redrawn on a schedule.
If you think Kodak belongs to another era, consider what directors say about themselves. Dominic Barton and Mark Wiseman reported in Harvard Business Review that only 34% of 772 directors believed their boards fully understood company strategy, and just 16% claimed a strong grasp of industry dynamics. A decade earlier, David Nadler found in the same journal that 56% of boards evaluated themselves regularly, yet only 16% had any plan to act on what they learned. Your board may have a Sasson in the building right now. The question is whether anyone at the table could recognize him.
Move one: name the force the board must push against
Every ownership structure bends decisions in a predictable direction. A founder overweights conviction. A family blurs the line between owner, director, and manager. Public markets pull attention toward the quarter and bury directors in compliance. A private equity owner optimizes for the exit.
WeWork shows what happens when nobody names the force. Adam Neumann chaired his own board, held super-voting shares, and sat among directors with close ties to him. The 2019 public offering collapsed.
Start your design by completing one sentence: "This board exists to counterbalance ___." If your directors cannot agree on the blank, you have found your first agenda item.
Move two: write the mandate before the skills matrix
Most boards own a skills matrix. Few own a one-sentence mandate that states whom the board serves and what it will leave alone. Reverse the order. One founder-led industrial firm in the Rowley and Shekshnia research wrote a purpose that committed the board to long-term development and explicitly excluded quarterly performance management and product decisions.
Copy that discipline. Draft the mandate in a single sentence, then list three things the board will refuse to do. The exclusions free hours for the work only a board can perform.
Move three: recruit for the gap, not the resume
Kodak recruited accomplished leaders whose experience matched the company it had been. Rowley and Shekshnia put the lesson in five words: build "a star team," not a team of stars.
So map capabilities against the strategy three years out, not against the company you run today. Mark the two largest gaps and reserve your next two seats for them, whatever prestigious candidate appears in the meantime. Tell every candidate the true time commitment in days, and decline anyone already stretched across too many boards.
Move four: engineer the meeting
Here is the finding that should change your next agenda. Rusty O'Kelley and PJ Neal surveyed 750 directors for Harvard Business Review and found that the best boards spend no more time than the rest. They spend it differently. Only 11% of top boards ranked financial statement review among their main uses of time, against 28% of all directors.
Structure produces that difference, and structure can be copied. Send materials at least seven days ahead. Label every agenda item as inform, discuss, or decide. Cap management presentations so deliberation claims most of the meeting. Require three alternatives for every major decision. Hold a session without management every time. Before the vote, ask what would have to happen for the decision to fail. And if you chair, speak last.
Move five: stamp an expiry date on the blueprint
Kodak ran board evaluations. They did not help, because directors answered through the assumptions they already shared. Add a harder test. Once a year, ask: if we built this board from scratch for the organization we expect in three years, which seats, committees, and rules would survive?
Then set triggers that force a redesign without waiting for a crisis. A new owner, a new CEO, a new strategy, or a technology that threatens the core business should each reopen the drawing. Bring one outsider into that review, because insiders cannot see the frame they sit inside.
The toaster on the table
Picture that conference room again. The machine works. The image glows on the screen. The future of an entire industry sits on the table, built by an employee, paid for by the company, and protected by patent.
Kodak never lacked the camera. It lacked a board built to see it.