Fifty-five percent of public company directors said in 2025 that at least one of their fellow board members should be replaced, the first time a majority has said so on record, according to PwC's Annual Corporate Directors Survey. Every one of those directors sat across a boardroom table from someone they no longer wanted there. Most said nothing and voted for reelection anyway, because the alternative meant admitting the selection process that put both of them there had failed.
That is the decision now facing anyone offered a board seat, and anyone building a board: whether the seat is being filled because of what a name signals, or because of what the person in it will actually do with the information in front of them.
The résumé stopped being the qualification
Boardrooms in 2026 are being assembled around strategic priorities rather than reputation, a shift away from the older pattern of filling seats based on title, tenure, or prestige, according to BoardSi's Boardroom Trends 2026 report. That sounds obvious until you look at how seats are actually filled today. More than 70% of board seats still go to candidates sourced through personal and professional networks, per The Veblen Director Programme. A network hire is not automatically a bad hire. But a network is a filter for who you already know, not a filter for what a board currently needs, and those two things only overlap by accident.
The old logic assumed a former CEO, a retired regulator, or a familiar name from another company's proxy statement would transfer competence by association. The new logic asks a narrower question: what specific gap does this person close, this year, on this board. That is a harder question to answer at a dinner party, which is exactly why it took a governance crisis to force it.
What directors are paid to ignore
Board work has also gotten considerably more lucrative, which raises the stakes on getting selection right. Average total compensation for public company directors passed $325,000 a year as of 2026, according to TechCXO. Median director pay rose 3% year over year in 2025 across 1,400 public companies spanning 24 industries, per the NACD and Pearl Meyer Director Compensation Report, cited via BoardCloud.
| Metric | Figure |
|---|---|
| Average total director compensation (2026) | $325,000+ per year |
| Median director pay growth (2025, YoY) | +3%, across 1,400 companies |
| Companies with combined CEO/chair role (2025) | 35%, down a decade running |
That last figure matters more than it looks. Only 35% of public companies combined the CEO and board chair role in 2025, continuing a decade-long decline, according to BoardCloud's analysis of NACD and Pearl Meyer data. Splitting the roles was supposed to fix oversight by putting someone other than the CEO in charge of evaluating the CEO. It has not, on its own, fixed who sits in the other chairs. The people who benefit most from title-based selection surviving are incumbent directors whose seats depend on relationships rather than a current skills gap, and the search firms and nominating committees who find network hiring cheaper and faster than a defensible skills-matrix process. The people who benefit from strategic-fit selection are shareholders, and directors who built a specific, demonstrable competency rather than a long tenure.
The information nobody hands you at the door
Even where boards get selection right, they are running on bad information. Seventy-two percent of directors cite the quality and flow of information from management to the board as one of the top three drivers of board effectiveness, according to the Harvard Law School Forum on Corporate Governance. A director with the right skill set and no reliable data from management is still guessing. That gap shows up in how directors rate each other: 87% of respondents report at least one ineffective board member on their board, and the average share of ineffective directors ticked up to 37% in 2026 from 36% the year before, per OnBoard's 2026 Board Effectiveness Survey.
Executives sitting outside the boardroom notice. Only 41% of C-suite executives rated their boards as excellent or good in 2026, according to The Conference Board and PwC, the highest share ever recorded in that survey. Read that number twice. The best year on record for board effectiveness is a year in which fewer than half the people managing the company think the board is doing a good job.
The AI test case
The clearest evidence that competence beats title comes from a controlled comparison rather than a survey of opinion. Companies with AI-savvy boards outperform their peers by 10.9 percentage points in return on equity, according to a 2025 MIT Sloan study cited in the Global Board Institute's 2026 Report. AI fluency is not a credential anyone had ten years ago, and it correlates with almost nothing on a traditional director's résumé, tenure, prior chairmanships, prestige of prior employer. The gap in performance is a gap in whether the board actually understands the decisions in front of it.
The case for the old system, made honestly
The strongest objection to skills-based selection does not come from people defending prestige for its own sake. It comes from governance researchers, including at least one McKinsey veteran with direct board experience, who argue that governance experts pushing rigid skills matrices often have not sat in a boardroom under real pressure, and that their frameworks add procedural weight, additional disclosure, additional committee review, additional formal skills audits, without improving the substance of what gets decided. On this view, continuity itself is a form of competence. A director who has sat through three CEO transitions, two activist campaigns, and a restatement carries institutional memory that a skills matrix cannot quantify, and swapping that person out for someone who checks a cybersecurity or AI box on paper can leave a board more fragile, not less.
This case deserves a direct answer rather than a dismissal, because it is often right in isolation.
Where the opposing case is right, and where it fails
The research shows the opposing case is correct about one thing: a skills matrix bolted onto a board as a compliance exercise, disconnected from actual strategic priorities, produces paperwork rather than performance. That is the failure mode BoardSi's own research warns against when it frames the shift as one toward strategic priorities specifically, not toward credentialism in a new form. A board that replaces a tenured director with someone whose only qualification is a fresh label, digital, ESG, AI, without asking what decision that person will change, has not fixed anything. It has just changed which kind of title gets the seat.
Operators report that the transition costs are real too: a board that turns over too much of its membership at once loses the shared context that lets twelve people make a fast decision under pressure without re-litigating history each time. That risk is highest in founder-led companies, family businesses, and boards mid-crisis, where trust built over years is doing work that no interview process can replicate on short notice.
Our view is that the fix is not tenure versus skill as a binary choice. It is holding tenure to the same test as any other credential: what specific, current gap does this person's experience close, and can anyone on the board name it. Fifty-five percent of directors could not say yes to that question about a colleague in 2025. That is not an argument for keeping the newest director out. It is an argument that continuity has to earn its seat the same way a first-time director does.
The seat, not the seat's title
For a director candidate, the practical question is not whether to accept a board seat. It is which one, and why they're asking. A seat filled because a nominating committee needed a specific gap closed, cyber risk, AI governance, capital allocation in a specific sector, comes with a mandate and a way to measure whether you delivered on it. A seat filled because your name looked good next to the others on the page comes with neither, and it is the version most likely to leave you as the 37% average that another director quietly wants gone.
The recommendation fails under specific conditions: early-stage boards where founder trust substitutes for formal process, boards mid-crisis where turnover itself is destabilizing, and any board that adopts a skills framework without linking it to an actual strategic priority. Outside those conditions, the math is straightforward. The pay is now over $325,000 a year. The information you'll be working with is, by directors' own account, often incomplete. The question worth asking before accepting the seat is not what the title will say on your bio next year. It is what specific problem the board hired you to solve, and whether you can still name it eighteen months in.