One Percent Better
Often the margin between winning and losing in sport can be incredibly small. Races are regularly won by fractions of a second and it's rare to find an athlete or team that truly dominates their sport for years without the slightest hint of competition. Software however, is dominated by a few giants — Facebook, Google, Amazon, Apple, etc — that are so unreachable that they aren’t really playing the same game as the rest of us. 1
Many athletes, even at a professional, level will never win a big competition; but there is always the opportunity for a shot at the podium. Some of the best stories in sport come about when a weird twist of fate creates favourable conditions for an outsider to turn the script on it’s head and steal glory from the big names.
Whilst the software industry — and its investors — love the idea of disruption and the plucky David character toppling a Goliath; it doesn’t happen all that often and the reality is messy. It’s likely that these seismic changes will become ever rarer as the large players cement their position using the resources they’ve acquired.
Imagine a scenario where a lone computer scientist makes a breakthrough in Artificial Intelligence that doesn’t require any of the resources of a company like Google. They discover a new technique that is a complete revolution of current thinking, and atop this technique build a business that comes right out of left field and casually sweeps aside all the big players.
However it's more likely that Google, with it’s huge pool of resources — both hardware and people — will get there first. Machine Learning 2 as it’s currently practised relies heavily on volume and luck, effectively brute forcing a solution with little understanding for how the solution actually works.
Google’s process is one of small incremental improvements; one of evolution, not revolution.
Despite these differences, pursuit of success in any arena can be approached in the same way. When Dave Brailsford joined Team Sky (the British Cycling team) in 2010 he took the simple approach of finding lots of small, incremental improvements in a bid to turn the team around.
He called his approach the “aggregation of marginal gains”, and explained that it required...
“...the 1 percent margin for improvement in everything you do.”
He believed that these tiny improvements would aggregate over time into something substantial and something that could transform a losing team into a dominant force.
The changes he implemented weren’t just in conventional areas such as training and nutrition either. One change that could easily dismissed as being too small to be worth addressing was the colour of the floor in the mechanics truck. Brailsford noticed some inconsistencies in the aerodynamic performance of the bikes, which he discovered was caused by dust accumulating in the truck and compromising their maintenance. The floor was subsequently painted white so that any dirt was easier to spot and clean up.
Brailsford recognised that his approach would take time, and he estimated that it would be five years before Team Sky would have a chance at winning the Tour de France. However, two years later British cyclists had become totally dominant at both the TdF and the Olympics taking the yellow jersey and 70% of the available gold medals. His approach had worked!
"Growth-hacking" isn't the answer
As developers and business owners, we love the myth of the “overnight-success” and we let the few exceptions prove the rule in our mind. I hope we can all agree that quick-wins don’t really exist and that shortcuts don’t take you to the same place as the long slow road. “Growth-hacking” is just investor bro-speak for morally grey actions that tie you in knots and prevent you from seeing the bigger picture.
Much like Dave Brailsford, I’m of the firm belief that success takes time and comes from a thousand smaller improvements. The notion of “1% better” can be applied to almost anything.
Many others before me have written about this approach. Harry Roberts talks about short "Refactoring tunnels" that aim to keep the scope of changes small and iterative rather than requiring large rewrites of a codebase fraught with unknowns. Jerod Santo at Changelog illustrates how slowing down can help our software grow faster.
If your software is your business, then this approach is even more important. Software is an unruly beast that's often hard to fully comprehend if the codebase is large with many people working on it. It's both intangible and unpredictable despite being perfectly logical.
The developers mentioned above understand that to manhandle and strong arm software with large sweeping changes is to risk losing control over it. When your business depends heavily on your software, you can inadvertently stop all progress dead and sacrifice future progress by pushing for a quick-win.
It's important to understand that by failing to embrace the notion of "1% better" and chasing glory through large pivots you not only pin all your hopes on a risky strategy — otherwise known as gambling — but forsake the small incremental gains too. If you shoot for the stars, you won't necessarily land on the moon3 — you'll probably crash and burn.
This is all really abstract. How about an example...
Your website isn't getting enough traffic, how do you fix it?
The “quick-win” solution might be to devise a targeted advertising campaign on Facebook. You identify the demographic you’d like to approach, draw up some artwork and put it out there at the cost of a few hundred (or thousand!) pounds.
Fortunately, the campaign is a huge success, and you get hundreds of new users visiting your site! Unfortunately, because they are only curiously responding to an advert rather than actively searching for a site like yours, they pass by quite quickly. Even worse, the funding for the campaign runs out and the surge in traffic dies down to similar levels as before.
Alternatively, if you took the same time and looked at your ranking for certain keywords, you might have seen that your site actually ranks very well in the search engine results, but that doesn’t seem to convert to traffic. You dig a little deeper and discover that the meta description of some of your pages is a little ‘off’ and could definitely be better. You make some improvements, and sit back and wait.
After a few weeks, you check back on your click-through-ratio and find that it’s gone up a few percent. That’s certainly not enough of an improvement that you can call it a day and retire, but it results in the same couple of hundred extra users passing through your site.
The difference is that this improvement, whilst less dramatic and requiring a little lead time before it's visible; is permanent, and the traffic is of much higher quality and more likely to become customers. Unlike the advertising campaign, it didn’t cost anything other than your time, and it will continue providing that little lift whilst you continue to find other small improvements.
This concept is effectively that of compound interest which Einstein once allegedly4 called “the most powerful force in the universe”. Whilst website traffic, bug density or conversion ratios do not always compound as money does, it’s obvious how these small improvements can aggregate into substantial changes.
Putting it into practice
I've been taking this approach with both my client work and side projects for the past 6 months or so, and I'm pleased to say that it has been really beneficial. Much like Jerod's sentiment of "slow down to go faster" I've found that focusing on smaller, incremental releases to have resulted in greater progress being made.
Side-projects that I've had on the back-burner for months have been built and shipped within weeks. Clients have seen continuous improvements in both the performance and finances of their sites and we're both spurred on and motivated by the regular "wins".
An illustration of this that if Apple's AirPods were a business unto themselves, they would have an annual revenue of $12 billion which is less than companies like Spotify, Adobe, Nvidia & Twitter
"Machine Learning" is not "Artificial Intelligence"
This "shoot for the stars" adage was a favourite of a CEO that I once knew. He would always reel it off whenever I disagreed with his latest idea for a pivot and instead advocated seeing our current plan through before writing it off. I wondered at the time if perhaps he was right and I was being too conservative, risk-adverse and lacking in ambition. Unfortunately for both of us, the long slow road would have been the better choice.
Whilst Einstein is widely credited with this phrase, there's no record of it until well after his death, making it unlikely he ever said it.