Most chairs I work with inherit their boards, with all directors already appointed, the committees in place and the routines set. Their job is to adjust and nudge, and keep the company on an even keel.

Gilbert Ghostine walked into an empty room. Actually, not even that. In February 2023 he was named chair-elect of Sandoz, the generics and biosimilars business Swiss pharmaceutical giant Novartis had decided to spin off and list as a public company. 

Sandoz had US$9.6 billion in sales and a history going back to 1886. But it had no finance function, no supply chain organisation, no investor relations team, no IT support of its own and no office. It also had no directors. "Everything we had was sales," Ghostine told us.

In eight months, he and his partners had to fill nine seats, write the governance rules and take the company public. Elizabett Yashneva and I document how they did it in a new INSEAD case, Assembling the Sandoz Jigsaw. 

The board as the first signal

The board matters especially in a spin-off because investors look at it before anything else, not least in the generics sector. Teva Pharmaceutical's mid-2010’s acquisition spree had destroyed billions in shareholder value, and US investors in particular were wary. Governance was therefore a strategic priority and part of the investment case for Sandoz.

The process started with a meticulous hunt for the board chair. Tripti Jha, then Novartis's Chief Talent and Transformation Officer, ran a search that took almost six months. It covered candidates from the US and Europe, inside and outside the industry. Finalists met the Novartis CEO, CFO and CHRO, the governance and nomination committees, and the board chair.

Ghostine, a former CEO of fragrance maker Firmenich, impressed them with his sense of purpose. "I'm looking for a mission," he told the recruiters. Before accepting the position, he flew to Zurich to have lunch with Richard Saynor, the future CEO, who had never run a listed company. Both left the meeting convinced about the other. Ghostine accepted his appointment as Sandoz chair shortly after.

Novartis gave him a broad mandate, albeit with one non-negotiable condition: no Novartis executives on the board. How much freedom should a parent give the chair it has chosen? What role should the parent play after that? My case leaves both questions to the classroom.

Boards that fit, not impress

In the November–December 2026 issue of Harvard Business Review, Timothy Rowley and I argue that good governance means building a counterbalance to the forces that shape a company's decisions. Those forces depend on the company's ownership structure and context. On the latter, we observe two recurring design flaws: Some boards are built to reinforce the instincts of whoever controls the company; others work well at first but don't change as the company does.

WeWork shows the first flaw. Its 2019 IPO collapsed largely because the board wasn't built to challenge Adam Neumann. He held super-voting shares and chaired the board himself, and the directors close to him tended to ratify his decisions rather than question them. 

Kodak shows the second. In the 1990s, its board was built for a film manufacturer that competed on engineering and scale. One person was both chair and CEO, and directors served staggered three-year terms. Together those features made the board respond too slowly to the digital market that replaced film.

The remedy, Rowley and I argue, rests on four interdependent elements: purpose, people, process and the chair. At Wright Technologies, a €100 million industrial firm, the founder and a new chair wrote down the board's purpose: to serve both the founder and the company. They then recruited directors to fill the founder’s knowledge gaps and challenge him rather than mirror what he already had. 

At Empire Group, a 200-year-old family company with more than 500 family shareholders, the board's first job is to keep shareholders from interfering in management. Its chair requires at least three alternatives for every major decision and always speaks last.

In contrast to the other firms, GeoTech never defined its board's purpose. The private-equity-backed company’s board was shaped by ownership stakes rather than a map of the capabilities the company needed. As a result, the board was reluctant to debate the strategic bet that could decide the company's future.

Assembling the jigsaw

Sandoz puts the four elements of good governance to the test under real pressure. Ghostine spent his first two months learning rather than recruiting. He met with each member of the management team and concluded that Sandoz sat "at the intersection of pharma and FMCG (fast moving consumer goods)". That diagnosis produced six anchor profiles for Sandoz’s director, two of which – supply chain and IT – were unusual for boards. The company still relied on Novartis for operations, so it needed directors who could help build those capabilities, not just oversee them.

Ghostine and Jha deliberately chose a different search firm from the one that had headhunted him for Sandoz to avoid any perception of bias. A first list of 150 candidates was cut to 72, and 28 of them were interviewed, in person wherever possible since Ghostine believed “the body language and the interaction reveal cultural fit”. 

The process was like assembling a jigsaw puzzle, balancing gender, nationality and experience while managing practical constraints such as candidates’ availability. Several times the puzzle had to be reworked because a preferred candidate's employer would not give approval.

The final board had nine directors from seven nationalities. Forty percent were women, and 60% had been a CEO or CFO. Yet the outcome is less interesting than the choices made along the way. For example, Sandoz’s board has no strategy committee, and the whole board was inducted together instead of the chair first.

Sandoz's board defined its purpose as supporting the executive team through the transition and establishing governance practices without disrupting operations. The board's process followed from that purpose. 

The chair and CEO prepare each agenda together with the corporate secretary, and each meeting has a theme, from year-end results in February to science and technology in December. Papers arrive seven days ahead and are capped at 20 slides. Each comes with a one-page summary stating whether the item is for information, discussion or decision. Presenters keep it brief, so 70–80% of each meeting is dedicated to questions and discussion, and decisions are taken by consensus. Every meeting ends with an in-camera session without management, and once a year the board evaluates itself. 

My case asks readers: What worked, and what would you do differently? How should work be divided between the new chair, the parent's board, the parent's CEO and the talent function? And what competencies should the board of a company that doesn’t yet exist have?

These questions are not just for spin-offs or new firms. Founders who hand over the chair and family companies that professionalise their board face a version of them. Get the design right and you build a board that helps the company thrive. Sandoz is literally the case in point: Since debuting on the SIX Swiss Exchange with an opening trading price of CHF 24.00 in October 2023, its stock has surged over 240%.