Why score the chief executive separately

Ask a board and its executive director the same question and you will often get two different answers. The board says it understands the organization's financial position well enough to govern it. The person who assembles the financial packet every month, sits through the discussion, and notices which questions never get asked has a different view. Both groups are answering honestly about the same organization.

That difference is the most useful thing a board assessment produces. It is also the first thing destroyed by the obvious design choice, which is to include the executive director in the survey and average everyone together.

What folding the executive director in actually does

Take a nine-member board plus one executive director. The board's mean on a domain is 4.2 out of 5. The executive director's is 3.0. Blend all ten answers and you get 4.08, which displays as 4.1.

So a difference of 1.2 points, the largest signal in the report, becomes a rounding artifact of 0.1. The executive director is one voice in ten, and the arithmetic does exactly what arithmetic does: it dilutes the minority view until it disappears. On a larger board it's worse. On a 15-person board that same 1.2-point disagreement moves the blended mean by less than 0.08.

You lose twice over. The executive director's view is gone, and the board's own number is now slightly wrong, contaminated by one respondent who isn't a board member and wasn't answering the same question in the same role. Neither figure describes anything real.

This is why the executive director's answers here go in their own column and are never merged into the board average. Not as a configuration option. The merge isn't a setting, because there's no version of it that produces better information than keeping them apart.

The gap is the finding

Once the two are separate, the number the report is actually built around is the difference between them, domain by domain.

The threshold used is 0.4 on the 1 to 5 scale, after rounding each side to one decimal place. Below that, the report says the board and the executive director see the domain about the same way and the domain doesn't appear in the divergence section at all. A section titled "where the board and the executive director see it differently" that lists places where they don't is noise, and the report treats a non-gap the same way it treats a withheld section: it shows nothing rather than something thin.

Above the threshold, the report states the direction plainly and shows both numbers next to each other. "The board rates financial oversight higher than the executive director does" is a sentence generated from the actual arithmetic, in whichever direction the arithmetic came out. It is not a pre-written narrative applied to whatever numbers arrive.

Three questions in the board instrument are asked only of the executive director and never of the board, including "I can raise bad news with this board without it damaging me." Those aren't averaged with anything. They're one person's answer, reported as one person's answer, because that's what they are.

One thing to be plain about, since it's the exception to how everything else in this assessment works: the executive director has a personal link and their answers are reported separately, so their column is identifiable by design. They know that when they answer. Board members' responses are not identifiable, and there is no mechanism to make them so. Both facts should be stated out loud before anyone answers anything.

What a gap actually means

The direction of the gap changes the reading completely, which is the whole reason for keeping the two figures visible rather than reporting a difference in the abstract.

The board rates financial oversight high and the executive director rates it low. Invented example: the board comes in at 4.4, the executive director at 3.0. The board believes it is governing the finances. The person preparing the numbers doesn't think the board is reading them closely enough to catch anything. In practice this usually means the finance report has become a ritual. It gets presented, nobody asks a question that would change a decision, and the board reads the absence of alarm as evidence that things are fine. This gap is the closest thing an assessment produces to an early warning, because the executive director is the only person in the room who knows what the board isn't asking about.

The reverse: the executive director rates the board's support higher than the board rates itself. Boards are often harder on themselves than their executive is, and sometimes that's just conscientiousness. But read it against the confidentiality point above. The executive director's column is attributable, and an executive who feels insecure has an obvious reason to rate the board generously. A flattering executive column from an organization where several other signals are strained is worth a private conversation, not a celebration.

A gap on executive support and evaluation specifically. If the board rates its support of the executive well above how the executive rates it, one common explanation is that the board counts encouragement as support while the executive means something more concrete: a written review process, a decision made rather than deferred, someone answering the phone during a crisis. Two definitions of the same word, showing up as a number.

None of these are verdicts. A 1.2-point gap doesn't tell you who is right. It tells you two informed parties are describing the same organization differently, which is a question worth an agenda slot and not something either party would have raised unprompted.

How to use it without overreading it

A few working rules.

• Read the gap, not either score. A domain score on its own means little without knowing how comparable organizations answered. The gap is internal to your organization and doesn't depend on any cohort at all.
• Don't ask the executive director to explain their column in front of the board. They answered honestly on the understanding that the answers would be reported. Turning the meeting into a defense of them teaches everyone what to say next year.
• Start with the largest gap and ask what each side was picturing. Most gaps come from two reasonable people using one word to mean different things. That conversation is short and often solves it.
• Watch the same gap over time. One year is a snapshot. A gap that widens is the actual finding, which is one more reason to assess on a yearly cycle.

The same logic extends outward. A board's view against the community's view produces the same kind of gap for the same reason, which is why the community and staff questions exist alongside the board ones and are reported separately rather than pooled. BoardSource has been making a version of this argument for years in its sector research: the board's own opinion of the board is the least reliable measurement in governance, and the fix isn't a better question, it's a second point of view.

The assessment here is free, takes about eight minutes per board member, and reports the executive director's answers in their own column with the domain gaps called out. If you'd rather run it yourself, the mechanism is simple enough to copy: ask both parties the same questions, keep the two sets of answers apart, and look at where they disagree.

A board scoring only itself is one point of view describing itself. The executive director sat through the same year in the same room, and is usually the only other person positioned to say whether that description holds up.