Small Things Compound

The most underrated force in software, and in almost everything else, is compounding. A 1% improvement, made once, is invisible. Made daily for a year, it is a different product. Made for a decade, it is a different company.

This is not a new observation. Books have been written about it. Spreadsheets have been built around it. The problem is not that people don't believe in compounding — it is that compounding is psychologically unsatisfying in the short term, and almost everything about modern work is optimised for the short term. We reward shipping. We reward launches. We reward the visible hero moments. We do not reward the quiet 1%, which is why it so rarely gets done.

I want to make the case for doing it anyway.

The Shape of Compounding

The thing that makes compounding hard to reason about is its shape. For the first long stretch, it looks exactly like linear growth. Sometimes it looks worse — a 1% daily improvement, in the first week, is indistinguishable from zero. If your metric for success is "did this week feel better than last week," you will conclude that compounding does not work, and you will stop.

But compounding is not linear. It is flat, then flat, then flat, then suddenly vertical. The vertical part pays for every flat week that came before it. The question is whether you have the patience, or the conviction, to get there.

Most people do not. Most teams do not. This is why compounding is such a reliable moat — it is not that competitors cannot understand it, it is that they cannot bear it. They quit during the flat part and go chase something that grows linearly but caps sooner.

What Compounds in Software

The obvious candidates are the ones you already know: codebase quality, team knowledge, user trust, search rankings, content libraries, internal tools. Each of these has a 1% daily decision — did we leave the codebase a little cleaner than we found it, did we write up what we learned, did we respond to that support email with care — and each of those decisions is trivially easy to skip.

The less obvious candidates are more interesting. Reputation compounds. The second time someone hears your name, it lands differently. By the tenth time, they are a user. Taste compounds. Every time you force yourself to look closely at something good, you build a slightly better internal reference library, and the next decision you make gets easier. Relationships compound. The person you checked in on last year, for no reason, is the person who refers you two years from now.

What ties these together is that they are all investments in the future disguised as small acts in the present. The cost is paid now. The return arrives much later, often from an unexpected direction.

The Enemies of Compounding

Compounding has two natural enemies: urgency and ambition.

Urgency pulls you away from 1% work because something on fire always looks more important than something that won't matter for a year. This is usually true in the short term, and almost never true in the long term. The unglamorous ongoing work — the documentation, the small refactors, the support triage — is what determines whether your product is still here in five years. The fire is forgotten in five days.

Ambition is more insidious. It tells you that 1% is not enough — that you should swing for 100%, that small is for small people, that if you are going to do the work you might as well do it at scale. Ambition is what convinces a team to rewrite the codebase instead of cleaning it up, to pivot the product instead of refining it, to hire aggressively instead of training the people already there. Ambition does not trust compounding because it cannot see the vertical part of the curve from where it is standing.

The irony is that ambition and compounding can coexist perfectly well. The people who compound the hardest over decades are usually also extremely ambitious. They just know that a small thing done a thousand times beats a big thing done once, and they plan accordingly.

A Small Practice

Here is the habit I try to keep, which is the only concrete thing I have to offer on the subject: once a week, I spend twenty minutes doing something that will not pay off for at least a year.

That can mean writing down what I learned that week, in a form I could share with someone in the future. It can mean sending a note to a person I admire, with no ask attached. It can mean fixing a bug that is technically not my problem, but that annoyed me. It can mean reading something outside my field, with the vague hope that it will be useful someday.

Twenty minutes is the point. It is short enough that I cannot justify skipping it — there is no week that is too busy for twenty minutes. And the effect, over time, is disproportionate. Half of the best things that have happened to me professionally came from something I did in one of those twenty-minute windows. The other half came from someone else doing the same thing and including me.

I mention this only because 1% work tends to sound abstract until you have a concrete slot for it. Once it has a slot, it gets done. And once it gets done consistently, the compounding starts.

Patience as a Bet

Compounding is ultimately a bet. You are betting that the world will still be around, that your field will still matter, that your own attention will hold long enough for the flat part of the curve to become vertical. These are all real uncertainties, and some of them will not resolve in your favour.

But the alternative — refusing to make the bet, and only doing work that pays off immediately — is also a bet. It is the bet that the future is irrelevant, or that someone else will worry about it. That bet has its own costs, and they tend to show up later, which is exactly the problem: by the time they arrive, the window for compounding has closed.

The cheapest version of this bet, and the one I would recommend to anyone starting out, is to pick something you believe will matter in ten years, and spend twenty minutes a week making it a little better. Do not measure. Do not announce. Do not wait for validation. Just keep going.

The first year will feel like nothing. The fifth will feel like luck.