Earkeep
The Eisenhower Matrix for smarter meeting decisions
Last updated: ยท Reviewed quarterly
Most people accept a meeting invite for one reason: the slot was open. The Eisenhower Matrix replaces that with an actual question — is this urgent, is it important, and does the answer make it mine? Applied to a to-do list it sorts tasks. Applied to a calendar it does something more useful, because a meeting you decline gives back an hour to everyone in the room, not just to you.
The four quadrants, translated to invitations
Eisenhower's distinction is between urgency, which is about timing, and importance, which is about consequence. Crossing the two gives four quadrants, and the classic labels — Do, Decide, Delegate, Delete — map onto meetings more cleanly than they map onto tasks, because every meeting already has a time attached and a set of people who could go instead of you.
| Quadrant | What it looks like on a calendar | Action |
|---|---|---|
| Urgent + important | Decision forums: incidents, escalations, launch go/no-go, a customer about to churn. Something breaks if it slips a week. | Do — attend, and protect it against everything else |
| Important, not urgent | Strategy, planning, architecture, 1:1s, retrospectives. Nothing breaks this week if it moves. Everything degrades if it always moves. | Decide — schedule it deliberately and defend it |
| Urgent, not important | Status syncs, FYI reviews, meetings where you are one of nine and speak twice. Urgent because someone put a time on them. | Delegate — send someone, or read the notes |
| Neither | Recurring meetings nobody remembers starting, standing invites with no agenda, updates that are already in a document. | Delete — decline, and propose ending the series |
The second row is where meeting-heavy calendars actually fail. Quadrant two is important precisely because it is never urgent, so it loses every collision with quadrant three: the strategy block gets moved for the status sync, week after week, and each individual trade looks defensible. The matrix exists to make that trade visible, because the only thing pushing back on it is a rule you decided in advance.
Important to whom? The question that makes the matrix work
The matrix fails when "important" is treated as a property of the meeting. It is a property of the meeting and you. A launch review is important; whether your presence is important depends on whether a decision needs your input or whether you are there to stay informed. Those two are constantly confused, and the confusion is what fills the calendar.
The sharper version is three questions about your own attendance:
- Is a decision being made that needs me? If yes, this is quadrant one or two regardless of what it is called.
- Am I here to be informed? Then the meeting is a delivery mechanism, and there are cheaper ones — a document, a recording, a summary. That is quadrant three.
- Would my absence change the outcome? If the honest answer is no, the invite is a habit, not a requirement.
Being informed is a real need. It is just not a need that justifies an hour of synchronous attendance, which is the whole content of the delegate quadrant.
The 80/20 view of a calendar
The Pareto principle — roughly 20% of inputs produce roughly 80% of outputs — holds uncomfortably well against a meeting calendar. Look back at a quarter and a small number of meetings produced nearly everything that mattered: the decision that unblocked the project, the 1:1 that stopped someone leaving, the review that caught the flaw before launch. The rest maintained a state of shared awareness that a written update could have maintained for a fraction of the cost.
Pareto and Eisenhower answer different questions and are stronger together. Eisenhower is prospective: this invite arrived, what do I do with it? Pareto is retrospective: over the last quarter, which meetings actually produced the outcomes? Run the second one and you get a calibrated definition of "important" drawn from your own history rather than from how the invite was worded. A structured version of that exercise is in the meeting audit.
One warning about the arithmetic: the 20% that produces the outcomes is rarely the 20% that felt busiest. High-urgency meetings feel productive because they are eventful. Quadrant two meetings feel slow and produce the compounding results, which is exactly why they lose the collisions.
A decision tree for the next invite
- Is there an agenda or a stated purpose? If not, ask for one before accepting. Half of quadrant-four meetings resolve themselves at this step, because writing the purpose down reveals there isn't one.
- Is a decision being made, and does it need me? Yes to both → accept, quadrant one or two. Yes to the first only → ask to be sent the outcome.
- Would something break this week if it didn't happen? Yes → urgent. No → not urgent, whatever the invite implies. "Someone wants it soon" is not the same as urgent.
- Could someone closer to the work go instead? If yes, that is a delegation and usually a better meeting — they know more detail than you do.
- Could this be a document? If the meeting is one person presenting to eight, the answer is almost always yes, and proposing it costs a sentence.
- If none of the above, decline with a reason and an alternative. "I can't add value here — send me the decisions and I'll follow up if I disagree" is accepted far more often than people expect, and it leaves the door open.
What to do about the meetings you keep
Sorting only pays if the surviving meetings produce something. A quadrant-one meeting that ends without a recorded decision has to be held again, which converts it into quadrant three the second time around. The discipline that makes the matrix stick is that every meeting you accepted as important leaves with its decisions and owners written down — that is what never leave a meeting empty-handed covers in full.
It also solves the delegate quadrant's real problem. "Read the notes instead of attending" only works if notes exist, which is why people attend meetings they don't need: attendance is the reliable way to know what happened. Earkeep changes that reliability. It records mic and system audio continuously while it runs and transcribes locally on your device, so a colleague who did attend has a searchable transcript by the time it ends, with no bot invited and nothing to start or stop. Meetings are defined afterwards by selecting a span on a timeline, or automatically from a connected calendar; audio is processed only in memory and never written to disk. Declining becomes cheap when the record does not depend on your having been there.
Where this fits
The matrix is the filter at the front of the system. What to do with the time it gives back — batching, buffers, defended focus — is in the pillar guide on time management for meeting-heavy roles. To calibrate the filter against your own quarter rather than your instincts, start with the meeting audit.
Make declining safe
Download Earkeep free for macOS, so the meetings you skip still leave you a record you can search.
14-day full trial, no account, no card required.