The tools you don't adopt

The tools you decline shape your working life more than the tools you keep. Every adoption charges rent — set-up time, a learning curve, notifications, maintenance, and an eventual migration — and the rent falls due whether or not the tool earns its keep. The most valuable productivity decision an operator makes is usually a refusal.

I say this as someone who builds systems for a living and enjoys software rather more than is healthy. My personal operating system has a deliberately small tools layer and a much larger rules layer, because rules compound and tools depreciate. This piece is the rule I use to keep it that way.

What does a new tool actually cost?

The subscription is the smallest line on the bill. The real costs never appear on an invoice:

• Learning. Typically several hours before the tool pays anything back, spent in the weeks when your enthusiasm is highest and your judgement worst.
• Configuration. Workspaces, integrations, permissions, templates — work that feels productive and produces nothing a client would pay for.
• Attention. Another icon, another badge, another place to check. Checking costs are paid daily, indefinitely, and never itemised.
• Data gravity. Once your notes, contacts or tasks live somewhere, leaving costs a migration. Commonly people stay with mediocre tools for years because the exit fee is a lost weekend.

Tools are infrastructure, and infrastructure has running costs. It is the same lesson cold email teaches: a new sending mailbox is not usable on day one — it takes weeks of warm-up before it earns a deliverability score worth sending on. Software is no different. Nothing is plug-and-play; everything is plug-and-maintain.

Why do operators over-adopt?

Because buying a tool feels like progress. It is a purchase-shaped promise that the underlying problem — unclear priorities, an undesigned week, a habit that does not exist yet — has been handled. It has not. A tool cannot install a behaviour, for the same reason willpower cannot substitute for a habit system: both mistake the artefact for the mechanism. The app is downstream of the behaviour, and downstream of nothing it does nothing.

There is also a professional hazard. People who work in systems — founders, operations people, consultants — enjoy evaluating software, so evaluation quietly becomes a hobby that reports itself as work. I have lost more hours to comparing note apps than I have ever lost to a bad note app.

When is a tool worth adopting?

This is the mechanism I run, in order:

1. When a friction recurs — three appearances in my friction log — then I write the job down in one sentence. "I need X so that Y happens without me." When the job cannot be stated in one sentence, then it is not a tool problem, and no purchase is made.
2. When an existing tool covers roughly 80% of the job, then I extend rather than adopt. A mediocre feature inside a tool I already run beats an excellent feature inside a new one, because the new one arrives with its entire cost stack attached.
3. When nothing covers the job, then I trial one candidate for 30 days, with the review date already in the calendar. One candidate, not a bake-off. Bake-offs are the hobby again, wearing a lab coat.
4. When the review date arrives and I cannot name what the tool did for me last week, then it goes. No sunk-cost appeals, no "but I might need it".

Most candidate tools die at step one. That is not the system failing; that is the system working.

What survives in a minimal stack?

Think in jobs, not brands: a calendar that runs the week, a notes system that actually gets queried, a CRM as the single source of truth on pipeline, one communication channel per audience, and accounting. One tool per job, and every tool holding a written job description it could be sacked against.

If a tool in your stack has no job description, it is not a tool. It is a lodger.

How do you retire the tools you already have?

Quarterly, I list every subscription and ask one question per line: what did this do in the last 90 days that nothing else could? Anything without an answer gets its data exported and its account closed the same day. Immediacy matters — "I'll cancel it later" is how lodgers become tenants.

What I notice after every cull is the same thing: nothing is missed. The work was never in the tools. The work is in the rules, the calendar and the follow-through — none of which has a pricing page.

---

Keep the thinking coming: The Format — a fortnightly letter on business and personal systems. Subscribe by email.