Is your data leading you astray?

In an increasingly complicated and data-filled world, how can we be sure we are making the right decisions? Abi Beaumont, Head of Data Strategy at AtkinsRéalis, shares her process for effective, unbiased decision making.

We live in a world full of noise.

In the next three years we will collect more data than we did in the last thirty. That’s a lot of information. And if you don’t know how to use it, it can easily lead you astray. In fact, IBM estimates that 3.1 trillion dollars are lost annually to poor data management and outdated sources.

The decisions we make now are also vitally important. Net Zero carbon by 2050 is a deadline we cannot afford to miss. And more than that, for the sake of ourselves, for our planet, and the future generations to come, it is our responsibility to make decisions that don’t just maintain the status quo but will help us leave our world in a better state than we found it.

So how do we make the right decisions?

Choosing the right data

In a world full of information, the first step to making well-informed decisions is picking the right data. With so much to choose from, we can’t possibly take all the data into account, and trusting the wrong information can badly skew your results.

Instead, we need to carefully seek out key data from trusted sources and interrogate our results to ensure we are not misrepresenting them.

For example, when we help clients decarbonize their estates and asset portfolios with our Decarbonomics™ service, the first step is to accurately assess their current carbon footprint. This can be an incredibly involved process, with many client estates encapsulating hundreds of buildings and structures. But by carefully selecting data points which are both particularly valuable, and particularly reliable, we can rapidly begin measuring and making informed recommendations for decarbonization.

Aggregated large-scale data can also be a powerful tool when leveraged correctly. We worked with experts to develop robust carbon baselines for different building types, depending on location, climate, age, purpose and so on. These baselines can then be used to supplement client data and interrogate unexpected results.

Asking your client the right questions

In 1985, Coca-Cola launched a new recipe for their flagship beverage – ‘New Coke’. The resulting consumer backlash rapidly became infamous as an example of when companies get it wrong.

Essentially, Coca-Cola hadn’t asked their customers the right questions. The new beverage outperformed both classic coke and Pepsi in blind taste tests and focus groups but didn’t have the same nostalgic and emotional significance for their customers, and when the recipe changed people became upset and angry.

There are a couple of things we can learn from this.

The first is the power of an agile innovation cycle so that you can adapt and change a product as you go along, to make sure it is well-tailored to meet your needs.

The second is the importance of trialing changes in the environment in which stakeholders will actually experience them, so you can be sure you have all the data. Many changes which seem like a great idea on paper translate poorly to the real-world, and unless you understand all the use cases for an asset or product, it’s easy to be blindsided by the impact of a change.

Challenging your instincts

There are 180 known cognitive biases in humans. This means that every single decision we make is being influenced by a selection of biases we are largely unaware of. We can’t discuss all 180 biases here, but it's worth drawing attention to two here – confirmation bias and selection bias.

Confirmation bias is when you interpret data to support your existing beliefs, sometimes at the cost of ignoring data which challenges this. The 2015 Volkswagen emissions scandal was a great example of this, in which upper management failed to challenge the false readings on their emissions tests because they wanted it to be true, and as a result they lost over $30 billion in fines and damages and added an estimated million tons of pollution to the environment.

Then there is selection bias, where the statistics have been based on an unrepresentative sample, allowing you to produce conclusions which do not reflect the truth. Or as the famous saying goes, there are ‘lies, damned lies, and statistics.’

We must be particularly careful to avoid these biases as we make decisions for the future.

The data often returns surprising results. One of the easiest and most common suggestions we make to clients running large estates is to replace their old halogen bulbs with LED lightbulbs, a cheap fix with an outsize environmental impact. But there are instances where this isn’t appropriate, such as when the lights are rarely used, so that it would take a long time before the new bulbs saved enough energy to justify the embodied carbon.

And that’s why it’s so important that we all continue to challenge our expectations, our biases, and even our data.

We have the power to make incredible decisions. Decisions that will help us build a better future for our planet and its people. So let's make sure we make the right ones.

This article was based on a talk given at TechFest in London on the 23rd of November 2023