Marginal gains: In favor of continuous improvement

We overestimate what we can achieve in a short period. At the same time, we underestimate a long-term improvement process.

Adrià Fontcuberta
3 min readSep 23, 2021
Photo by Isaac Smith on Unsplash

It’s pure maths

First, think of any process you’d like to improve.

Say… time-to-market in a given company. Or % of churn. Or lead time for a given team. Or average speed in a 100km bicycle ride.

Imagine you want to improve any of them by 20%. 20% faster, 20% safer, 20% better, whatever. Any process, any activity, any outcome.

Nope. It ain’t easy.

Compounding effects

What would it take to improve by 1% instead?

Can you think of a 1% improvement for a bicycle ride?

It sure looks more doable to me.

For instance, go for a ride today. Ride for 40km, and that would help you improve. Or read about a better training plan. Or try a new nutrition schema. Or buy a more expensive bike (ok, now I’m justifying myself. Sorry about that).

Tomorrow, improve it by another 1%. Then do it again. And again.

Imagine you kept doing that by a year.

That’s right. By improving a 1% daily, that process would become 37 times better better/faster/stronger in a year.

That right there is a compound effect: any future change builds upon the previous ones. So do not underestimate that 1%.

But but but

I can hear you screaming, “but…!” . We assumed a couple of interesting facts here:

First of all, we assume you’d keep improving every single day. It requires commitment, and you need to find daily small, viable improvements.

It sounds daunting because it is.

Yet, what’s the alternative? Good look finding a great, actionable, realistic good idea that improves anything 37 times… at once!

Also, it’s easier to find those small improvements. You know some of them already because you face them every day! So finding and identifying those things could help you get closer to the desired outcome.

Remember that they should be small. You should be able to put them in place in a matter of hours or days.

Better over perfect

Let’s face it: a 1% improvement is not that exciting. It might feel like nothing. Sometimes it isn’t even noticeable.

But it pays off. It’s maths! Maths do never lie.

The key here is to nail that 1% every day. Continuously.

A real-life example

Let’s say you want to improve the cycle time of a engineering team.

You could start by having a better understanding of what affects Cycle Time.


…improve the tracking of the said metric. Maybe you need to create a new dashboard or some alerting.

…suggest using the daily meeting to focus on the rightmost task (the one that’s about to get to Done).

…experiment with focusing on finishing stuff before pulling new work.

…add an automatic report that shows how the Cycle Time changes over time so that people can reflect on their efforts.

And so on. You know the drill.

In short

First of all, you need to have a clear outcome. I know it’s obvious, but still.

Then, think about small changes. Remember — they need to feel small. It’s better to take a small step in the right direction than two steps towards the wrong one.

Keep on improving with a short cadence. Remember that compounding effects will make the most out of each improvement.

Originally published at on September 23, 2021.



Adrià Fontcuberta

Words matter – Software product development, Front-end, UX, design, lean, agile and everything in between.