How to Hold Your Executive Team Accountable
Holding an executive accountable is a different problem from holding a team member accountable, and the standard advice does not transfer. There is no manager above them to notice. The outcomes take quarters rather than weeks to resolve. They are frequently more expert in their domain than you are, which makes their explanations hard to evaluate. And the relationship is peer-adjacent enough that a direct conversation feels like a rupture rather than a routine part of management. The result is that most founders can name an executive they have had doubts about for two or three quarters and said nothing.
What’s inside
Why this is harder than team accountability
The prerequisite: outcome plus decisions
The cadence
Assessing someone in a domain you do not know
Four causes of executive underperformance
The removal you are delaying
Where Blomma fits
Why this is harder than team accountability
Four structural differences, each of which defeats a standard accountability practice.
No one else is watching.A struggling team member has a manager who notices within weeks. A struggling executive has you, and you see a curated version.
Long feedback loops.A VP Sales’ real performance resolves over two or three quarters. By the time the trend is unambiguous, you have lost a year, and the intervening quarters each had a plausible explanation.
Asymmetric expertise.They know their function better than you do, which means you cannot evaluate their choices directly and have to rely on proxies. It also means a confident explanation is difficult to challenge, and confident explanations are what senior people are good at.
Peer proximity.They are part of your leadership team, possibly a friend, and often someone whose arrival was a milestone for the company. Raising performance feels like a betrayal of that, so it gets deferred — which is the same dynamic that makes cofounder performance conversations so hard.
None of these makes accountability impossible. They mean it has to be designed rather than left to arise naturally, because naturally it does not.
The prerequisite: outcome plus decisions
Before any accountability conversation is possible, two things must exist, and their absence is the reason most executive accountability fails.
A written outcome with a measure.One sentence, agreed. Without it, every performance conversation is your impression against their impression, and their impression comes with more domain expertise attached. You will lose that argument even when you are right.
The decisions required to move it.If they own an outcome but three of its five drivers sit elsewhere — with you, with another function — they cannot be held to it, and attempting to is unfair and unproductive. Trace the drivers before you hold anyone to anything.
If either is missing, that is the first conversation, and it is a much easier one:“I don’t think we’ve been clear about what you own and what you decide, and I want to fix that.”Founders who skip straight to performance when the outcome was never defined are addressing their own omission as though it were the other person’s failure — and the executive will know that.
The cadence
Accountability is a rhythm, not an event. Four layers.
Weekly one-to-one, mostly not about performance.Blockers, decisions, what they need. This is where the relationship gets maintained so that harder conversations are possible. If your one-to-ones are status reports, you have no venue for anything else.
Monthly outcome check.Against the written measure. Short, factual, and consistent — the consistency is what makes it unremarkable rather than an event. “Where are we on activation, and what’s your read?”
Quarterly assessment.A real conversation about how the quarter went, what they learned, what changes. Written, briefly, by both of you. This is the layer almost nobody has, and it is the one that prevents a two-quarter drift from becoming a surprise.
Annual, backward and forward.Whether this is still the right role for the company’s next stage and for them.
The monthly check is the load-bearing one. Its function is to make the performance conversation routine rather than exceptional, so that raising a concern is a normal Tuesday rather than a signal that something has gone badly wrong.
Assessing someone in a domain you do not know
The genuine difficulty. Five proxies that work without you needing to be the expert.
Are they hiring people who are better than the last set?The most reliable single proxy. Strong functional leaders raise their team’s ceiling; weak ones hire people they are comfortable being better than.
Do their people develop?Name what each of their reports can do now that they could not two quarters ago. If nobody has grown, that is a signal regardless of the numbers.
Do they bring you problems early, or explanations late?A strong executive flags a miss before it lands, with a plan. A struggling one produces a well-constructed account afterwards. The timing of information is more diagnostic than its content.
Can they tell you what they are wrong about?Ask what they have changed their mind on this quarter. Strong senior people answer easily. An executive who has never been wrong about anything is either not learning or not telling you.
What do their peers think, unprompted?Not a poll — but if two other executives independently work around someone, that is data. Peers detect capability faster than you do because they interact with the work rather than the reporting.
Use the proxies together. Any one can mislead; the set rarely does.
Four causes of executive underperformance
Same diagnostic discipline as anywhere else, because the remedies differ completely.
The seat is broken.They own an outcome without the decisions, or the role is genuinely two roles. Run the attribution question: would the best possible person succeed here, given the authority this seat holds? If not, replacing them reproduces the outcome. This is the most common cause and the one founders check last, because it points at their own design.
Never told.No defined outcome, no feedback, and eighteen months of accumulating doubt you never voiced. Common, and it means you do not yet have the information to judge — people frequently change when the expectation is finally stated.
Stage mismatch.They were right for the function at twenty people and the job is now different. No villain, and the resolution is a role change or a real development attempt with a defined horizon, not a performance process.
Genuinely not right.After a defined outcome, honest feedback, and real support, the gap remains. This exists and it is a conclusion to reach at the end of the list rather than the start.
The removal you are delaying
Worth addressing directly, because nearly every founder has one and the delay is the expensive part.
The cost of keeping an executive who is not right is not primarily their own output. It is that they set the standard for everyone below them, they hire in their own image, and your other executives are watching what you tolerate — which tells them what the real bar is, whatever you say about excellence.
The most common regret founders report about their executive teams is not a hire they made. It is how long they waited after they knew.
Before acting, do two things. Confirm the seat is sound, using the attribution question. And confirm you have actually said it plainly — not hinted, not raised it as a general observation, but told them clearly what is not working and what needs to change by when. A surprising number of executive removals happen without the person ever having been told directly, which is unfair to them and makes the exit worse.
Then, if it holds, move faster than is comfortable. Handled early and generously, an executive departure is a normal event that the company absorbs. Handled after four quarters of visible drift, it is a referendum on your judgment.
Where Blomma fits
This is one of the loneliest parts of the CEO job. You cannot discuss an executive’s performance with their peers, with their reports, or with the person themselves until you have worked out what you actually think — and your board will draw conclusions about your judgment from how you raise it.
Blomma is an always-on AI career coach with no stake in your team. Use it to run the five proxies on an executive you have doubts about, which converts a vague unease into something specific. Use it to check the attribution question honestly — whether the seat you designed is the actual problem, which is the most uncomfortable and most frequently correct finding. Use it to establish whether you have genuinely told them, or only hinted. And use it to prepare the conversation, including the version where they push back with more domain expertise than you have.
The second application: several of your executives are doing a job they have not done before, and the honest reason some of them are struggling is that nobody is developing them — you are their only manager and you are stretched. Coaching for your executive team is what turns a stage mismatch into growth rather than a departure..
Executive accountability does not happen naturally, because every structural condition works against it. It happens when there is a written outcome, a boring monthly rhythm, and a founder willing to say the thing in quarter one rather than quarter four.