Digital minimalism for meeting-heavy knowledge work

No one decides to use fourteen tools. Each arrives with a reason: a team adopts a chat app, a project needs a tracker, a client insists on their video platform, someone trials a note-taking app and it becomes load-bearing. Every addition is locally justified, and the total is a working environment nobody designed.

Newport's central claim is that the costs of these tools are cumulative and mostly invisible, while the benefits are individual and obvious. That asymmetry is what makes accumulation the default. At work it produces three specific costs:

• Attention residue at the tool level. Each tool is a surface with unread state, and checking one is a switch out of whatever you were doing. Fourteen surfaces means an unbounded number of small switches per day.
• Fragmented records. The decision is in the meeting; the summary is in one app; the follow-up is in a chat thread; the document is somewhere else. Nothing is wrong with any piece, and the whole is unsearchable.
• Maintenance work. Tools require configuration, integration, permissions, and periodic migration. That is unpaid work performed by the person who least wanted the tool.

The last one compounds with meetings, since coordination overhead is a meeting generator: more tools produce more places where state diverges, and diverging state produces syncs.

Newport's method is not gradual reduction. It is a clean break for thirty days, followed by deliberate reintroduction of only what earns its place — because gradual reduction lets each tool argue its own case, and each case is individually persuasive. The work version needs adapting (you cannot stop using the company's ticketing system for a month), but the structure holds.

• Days 1–3: inventory. Every tool you opened last week, and every recurring meeting. Both lists are longer than expected, and that is the first result.
• Days 4–30: suspend the optional. Anything not required by someone else's workflow — turn it off. Not deleted, suspended. Newport's point is that the discomfort of the first week is information about habit rather than about need.
• Also suspend notifications, all of them. This is the higher-yield half. Notifications are the mechanism by which fourteen tools reach into a focus block, and turning them off is reversible in a click.
• Day 30: reintroduce with a rule each. Each tool that comes back needs a stated job and a stated time — "chat, checked at 11:00 and 4:00." A tool without a rule is a tool that will resume checking you.

Expect two or three suspended tools to be forgotten entirely by day 30. That is the point of the exercise, and it is not an argument you could have won on day 1.

Newport's declutter targets apps. The work version has to include the calendar, because a recurring meeting is a subscription with a weekly cost that nobody re-evaluates — and unlike an app, it charges other people too.

Apply the same three questions to every recurring meeting: what does it produce, would anyone notice if it stopped, and could a document do it? Then run the actual experiment rather than the debate: cancel the doubtful ones for one month and see what breaks. Almost nothing does, and where something does break, you have learned what the meeting was for — which is more than the calendar entry ever told you. The structured version is in the meeting audit.

Notification decluttering interacts with this directly. A calendar full of meetings plus a desktop full of alerts produces a day with no interval long enough for anything demanding, and the two are usually addressed separately by different people, neither of whom sees the total.

The standard response to a fragmented stack is integration: connect the tools so information flows between them. It is worth being clear-eyed about what this buys. Each integration is a moving part that breaks on API changes, duplicates records into a second place, expands the set of services holding your data, and adds a maintenance obligation. Ten tools with twenty integrations is a more complex system than ten tools, not a simpler one.

The alternative is fewer tools with plain outputs. A tool that writes files to a folder you chose needs no integration to be useful to the next tool: everything on your machine can already read a folder, including things that do not exist yet. That is a different bet from the integration one — it is a bet on formats rather than on connectors — and it is the argument in plain files beat platform lock-in.

The same reasoning applies to the number of services holding your working conversations. Every cloud tool in the stack is another processor, another privacy policy, another account that must be revoked when someone leaves. Local-first tools reduce the count structurally rather than by policy: there is no server to audit if there is no server. See private by design and local-first AI at work.

The part of Newport's book that gets skipped is the one that makes the rest sustainable: decluttering creates a vacuum, and a vacuum fills itself with whatever is nearest. Remove the tools and add nothing, and the freed attention goes to scrolling, which is why most digital declutters last three weeks.

His prescription is demanding leisure — activities with skill, effort, and often other people in them. For meeting-heavy work there is a specific reason this matters: a day of meetings is a day of continuous low-grade social performance, and the recovery it needs is not more screen time in a different application. Something physical, something with your hands, or something requiring real concentration restores capacity in a way that passive consumption does not. Newport's stronger claim is that solitude — time free from other minds' input — is itself a requirement, and a meeting-heavy calendar with a podcast in every gap supplies none of it.

A page arguing for fewer tools should be direct about a tool it recommends. The minimalist test is not whether something is useful — almost everything is useful — but whether it supports something you value enough to justify its cost, and what that cost actually is.

Earkeep's cost is deliberately low on the dimensions that matter here. It has no notifications and nothing to check; it does not become another surface with unread state. It records mic and system audio continuously while it runs and transcribes locally on your device, so there is no bot to invite and no account to manage — audio is processed only in memory and never written to disk. Meetings are defined afterwards by selecting a span on a timeline, or automatically from a connected calendar. Its output is plain files in a folder you pick, which means it does not need integrations to be part of your stack and does not hold anything hostage if you stop using it.

Whether it earns its place is a question about your week, not about the tool. The honest version of the minimalist test: does having a searchable record of what was said support something you actually value, or would it be one more thing to maintain? For most people in eight to fifteen meetings a week, the answer is the first — but it should be answered rather than assumed, which is the whole of Newport's argument.

Decluttering frees time; time management for meeting-heavy roles covers what to do with it, and deep work on a meeting day covers the practice the freed attention is for.

Download Earkeep free for macOS and judge it against your own week before it earns a permanent place.