Marginal Gains in IT: One Step at a Time

When thinking about how to improve most systems, you won’t go far wrong following the approach popularised by Team Sky general manager Dave Brailsford – ‘the aggregation of marginal gains.’ Brailsford’s methods have proved incredibly successful in his field, propelling the Sky and British cycling teams to the pinnacle of the sport in just a few years. However, they are not necessarily unique.

Stay current on your favourite topics


Formula One race teams have long lived by the same philosophy. They seek to get the most out of every part of their system – from the fuel, engine and gearbox that give it thrust to the aerodynamics and tire composition that give it traction. They also need to understand the relationship between those parts to ensure the right ‘balance’ and accommodate environmental factors – making sure the right set-up for the track and weather conditions in question.

In the world of information technology there are many examples where the philosophy behind “the aggregation of marginal gains” can be applied, not only to improve system performance, but other characteristics like security, reliability or cost efficiency. Unfortunately, many people only see change as meaningful when there is a significant visible outcome, or a step change to improve the system in the short term. But these changes tend to require large, complex and costly change programmes. Meanwhile, many small and simple incremental changes may be overlooked, but could be just as meaningful or impactful when compounded in the long run.

So how do you apply Brailsford’s approach to achieve marginal gains in IT systems? You can start by building a detailed map of the entire ecosystem, breaking down a complex system into its smallest component parts (or processes), and understanding the dependencies and communications channels between them. Opportunities for improvement are found by systematically assessing and measuring the characteristics of each component to determine their impact on the overall goal for improvement. When this approach is applied continuously over an extended period of time, the benefit from compounding small gains over time, can result in a much larger aggregate improvement – the aggregation of marginal gains.

This is a very simple concept and by no means a new idea, but thinking about IT system improvement in this way can act as a reminder of some simple rules of thumb to improve the chances of success.

  1. Foundations – Start with strong foundations; you’re unlikely to succeed in building a race winning formula one car if you start with a golf buggy.
  2. Commitment – Make sure the entire organisation is committed to the approach. Without the promise of the dramatic, short-term step change, you need long-term commitment to give the approach enough time to deliver success.
  3. Deconstruct – Breaking down a complex system into its component parts gives a greater understanding of how the system works and can simplify the assessment of the overall system, and deliver achievable improvements.
  4. Systematic – Being systematic when assessing complex systems helps ensure every aspect of each component, and every potential opportunity for improvement is investigated.
  5. Evidence – Decisions for change should be evidence-based.
  6. Assess, Measure, Improve – Assessing to discover potential opportunities for improvement; measuring to quantify the potential opportunity; and improving to realise the opportunity.
  7. Continuous – Small but continuous improvements over time can result in greater overall gains when they are compounded in the long run.
  8. Relentless – You must be relentless in your search for opportunities to improve and in delivering improvement
  9. Thorough – All areas that could have an impact should be investigated, including peripheral factors that may not at first seem relevant

There are countless practical use cases where these simple rules can be applied. In my own experience, I have used them to guide projects ranging from trading system evaluations (where the goal was to optimise for the lowest latency), application availability assessments (where the goal was to identify and remedy IT risk factors), through to optimising the provision of IT services (where the goal was to drive greater customer satisfaction).

We all face growing complexity in almost every aspect of business systems. So simply being able to deconstruct complex systems into atomic or modular pieces – helps cut through that complexity and ensure we stay focused on making achievable improvements, on a regular basis, with the knowledge that in the long run, small gains can compound to have a much bigger impact on the overall system.

Even so, while marginal gains is a pragmatic approach to improvement, we also need to keep an eye on revolutionary advances – or paradigm shifts – that cause wholesale changes in the way people approach a challenge. When the motor car was invented, even the most ardent follower of marginal gains wouldn’t have got their horse drawn carriage to compete. Spotting those ‘revolutionary’ changes is not always easy. But that is the subject for another blog.

Would you like to know more about our work?

The author

David Sewell

David Sewell

CTO & Partner, London

An innovative technologist with more than twenty years experience creating IT strategies and building solutions for eTrading applications, market data and shared platforms, including hybrid cloud across the capital markets industry.