On the Time We Save

And what actually happens to it.

Every productivity tool sold in the last century has made the same promise. Save hours. Reclaim your day. Get your evenings back.

Oliver Burkeman's Four Thousand Weeks takes that promise seriously and then dismantles it. His argument is simple and uncomfortable: the average human life is around four thousand weeks, and the productivity industry has spent a century convincing people that if they just optimize hard enough, they can somehow beat the math.

Burkeman's position is that we cannot. Every hour saved through better tools, better systems, better habits gets absorbed by more work, more expectations, more meetings. The productivity we generate does not accrue to us. It accrues to whatever institution or ambition sits above us and is happy to take everything we produce.

This is a problem for anyone building productivity software.

Our own included. We are building a piece of software whose core promise is that it gives people back their time. If Burkeman is right, what are we actually selling?

The productivity claim

Every SaaS company in 2026 sells the same story. Save your team ten hours a week. Reduce reporting time by 80 percent. Automate the busy work so you can focus on strategy.

We have used variations of these lines. Most competitors have too. It is the industry's default value proposition and has been for decades. When outcomes are hard to quantify, hours saved becomes the fallback currency.

The problem is that nobody in the industry checks what happens to those hours after they are saved. The assumption is that the user gets them back. That they go home earlier, take more walks, spend more time with their kids, think more clearly, do more strategic work.

The assumption, in other words, is that time saved equals time recovered.

Empirically, this is not what happens.

Parkinson's Law, restated

In 1955, Cyril Northcote Parkinson published a satirical essay in The Economist observing that work expands to fill the time available for its completion. The observation was originally about bureaucracies growing regardless of workload, but the deeper principle has proven durable across almost every human system.

Give someone eight hours to write a report and they will write it in eight hours. Give the same person two hours and they will write a comparable report in two hours. This has been replicated informally across every knowledge worker any of us has ever met.

The productivity industry treats Parkinson's Law as a tool for personal discipline. Shrink the deadline, sharpen the output. But it has a second implication that rarely gets discussed.

Saving someone six hours by making a task more efficient does not give them six hours back. They do more work in those six hours.

The savings do not accrue to the person. They accrue to the workload.

The Jevons Paradox for time

In 1865, William Stanley Jevons observed something counterintuitive about the industrial economy of his day. As steam engines became more efficient, British coal consumption did not decrease. It increased. Dramatically.

The reason: efficiency reduced the cost of using coal, which made coal-consuming activities more attractive, which expanded the range of uses coal was applied to. Efficiency did not conserve the resource. It multiplied demand for it.

This is now called the Jevons Paradox and it has been documented across dozens of technologies. More fuel-efficient cars lead to more driving. More energy-efficient buildings lead to larger buildings. Faster internet leads to heavier applications. The savings get absorbed by expanded consumption.

Time behaves the same way.

When a report that used to take eight hours now takes thirty minutes, teams do not do the report in thirty minutes and leave. They produce sixteen reports in the same eight hours. When email became easier to send, nobody spent less time on email. They sent and received more of it. When meetings became easier to schedule through calendar automation, calendars did not empty. They filled up faster.

Efficiency of a scarce resource, whether coal or time, does not preserve the resource. It expands demand for what the resource is used to do.

Productivity tools that promise to save time run into this every time. The time saved does not become free. It becomes more expensive, because now it is expected to produce more.

What actually happens to saved time

Over the 6+ years of managing Amazon brands at e-Comas, we have watched this play out at close range hundreds of times.

Give a client a better dashboard. Their weekly reporting time drops from four hours to twenty minutes. They do not leave four hours earlier on Friday. They now review the same numbers three times more often, ask deeper questions, run more experiments, request more variations of the report for different stakeholders.

Save a brand manager six hours a week on advertising analysis. Those six hours do not become strategic thinking time. They become time to manage more campaigns, more marketplaces, more product lines, because leadership now expects proportionally more output for the same salary.

The people who buy efficiency do not get freedom. They get scope creep.

This is not a criticism of them. This is the water we all swim in. Any hour saved gets absorbed by whatever ambition, expectation, or institutional pressure is sitting above us waiting for available capacity.

The uncomfortable question

If users do not actually get their time back, what are we really selling?

The honest answer: not time. We were never selling time. We were selling something else and calling it time, because time is easier to quantify in a sales deck than the thing we were actually delivering.

We were selling attention. We were selling cognitive load reduction. We were selling the ability to hold fewer things in working memory. We were selling confidence in a decision. We were selling the removal of a nagging feeling that something might be wrong that has not been looked at.

Those are all real and valuable. They are also very different from time.

A tool that takes a task from eight hours of anxious spreadsheet work to thirty minutes of confident scanning is not saving seven and a half hours. Those seven and a half hours will get absorbed regardless. The confidence stays.

What we should measure instead

If time saved is a proxy for something else, and the something else is what customers actually value, we should measure the something else directly.

For an analytics tool, the honest metrics are things like: how many decisions did the user make this week that they would not have made otherwise? How many surprises did we catch that would have become expensive next month? How many hours of unproductive worry did we replace with five minutes of confident scanning?

These are harder to quantify. They do not fit on a landing page as neatly as "save 10 hours a week." But they are what people actually pay for once they have used the tool for six months. The renewal conversation is never about hours saved. It is always about confidence gained.

Building for protected attention

There is a version of software that respects this reality. It does not promise time. It promises the return of a specific cognitive resource.

For an executive with twelve calls a day, that resource is not hours. They have plenty of hours in the day, technically. What they lack is uninterrupted cognitive capacity to actually see what is happening in their business.

Software that gives them a five-minute weekly scan is not saving them time. It is protecting attention from being drained by a task that would otherwise consume a full afternoon of scattered checking, second-guessing, and tab switching between tools.

The five minutes get absorbed like any other productivity gain. The cognitive protection is what stays.

This is a subtle reframing, but it changes what gets built. A tool that optimizes for hours saved tends to get faster. A tool that optimizes for attention protected tends to get calmer, clearer, more decisive. The two are not the same product.

What stays

Ask an executive what they actually want from a better tool and they will usually say "more time." Ask them what they would do with more time and the answers get vaguer. Read more. Exercise. See the kids. Think.

The vagueness is honest. They do not really want hours. They want something the hours were supposed to buy but never did.

They want to close the laptop and mean it. They want to be at their kid's football game without half their brain running Q4 numbers. They want to have dinner with a friend without the low background hum of unread dashboards. They want the Sunday evening to stop being about Monday morning.

None of that is time. All of it is the absence of a specific kind of low-grade cognitive load that follows a person from the office into the car, into the house, into the bedroom, into the sleep. The load that says: something might be wrong and you have not looked. The load that Burkeman's four thousand weeks are actually stolen by, one background loop at a time.

Software that protects attention gives that back. Not hours. Presence. The ability to be in one place at a time. The ability to trust the system enough that you can stop checking, and mean it.

Four thousand weeks becomes livable not when we have more of them but when we are actually in them. A tool that helps a person be more present in their own life is doing something time management never could.

That is what should be sold. That is what should be measured.

What we believe

A few positions.

The productivity industry has sold time savings for a century and mostly delivered work expansion. The failure is not the tools. The failure is the framing.

Parkinson's Law and the Jevons Paradox are the two most under-discussed principles in productivity software. Every serious product person should understand both before writing a value proposition.

Time is not a scarce resource that can be conserved through efficiency. Attention is. Confidence is. Cognitive load is. These are the things a good tool actually preserves.

The most useful business software of the next decade will stop promising hours back and start promising clarity, decisiveness, and attention protection. The market will eventually notice the difference.

We should design accordingly.

Why this matters for you

If you are building software, the question is not "how many hours do we save?" The question is "what cognitive resource do we protect, and does that resource compound over time for our users?"

If you are buying software, the question is not "will this save my team ten hours a week?" The question is "will this let my team make better decisions, with less worry, in the small amount of time they actually have?"

The difference sounds semantic. It is not. It changes what you look for, what you buy, what you build, and what you get.

Four thousand weeks is not a lot. If a piece of software really wanted to respect that math, it would stop trying to give us hours we cannot actually keep, and start protecting the attention we can.

Read On MCP → · Read the manifesto →