How to Be Data Smart

These days, we encourage everyone, be it start-ups, corporates, or governments, to use data to make decisions. Given our ever-increasing ability to measure things and the advent of big data, it is natural to use numbers to reinforce the rationale behind key decisions. Enter the domain of analytics, commonly assumed to spit out insights by deploying complex algorithms that crunch vast amounts of data. So, how much influence should data have on a decision? Should one follow the data blindly? Or should one allow it to loosely shape intuitions? There is no set answer to this, which is why the three thresholds: ‘data-driven’, ‘data-informed’ and ‘data-inspired’ have been established. 

The most commonly used term is ‘data-driven’, which means data is the end-all and be-all of everything, and all decisions boil down to what the cold hard facts suggest. ‘Data-informed’ is an approach where data is considered along with experience and insights. In this case, data acts as a check on intuition. ‘Data-inspired’ is an exploratory approach that focuses less on statistically sound methods and more on blending data from various sources to identify common trends.

A better way to understand these thresholds is to map them out on a consumer journey based on the SERVQUAL model. The data-inspired approach is best for building up a hypothesis by identifying trends and works well for consumer and knowledge gaps that focus on understanding the market context and gauging customer needs. The data-informed method works within a generic hypothesis that is tweaked and tested and therefore meshes with the communication and design gaps that focus on designing products and building communication strategies. Finally, the data-driven practice works best with the execution gap, where measurement strategies identify potential issues with routine functions – for example, problem resolution of customer complaints.

Product leaders and innovators tend to leverage all three thresholds but at different times. In the podcast Dear Brown Parents, the founder of Atoms shoes explains how they were initially inspired to produce formal shoes based on their observation of people in Pakistan. They then pivoted to sneakers when they noticed that trends were more informal in the West – akin to the data-inspired stage. They then moved to the data-informed threshold when they began hypothesising what made existing sneakers successful. Their research showed that people liked the look of laces but disliked tying them; this led them to develop laces that needed to be tied only once. The data-driven part was figuring out which materials to use to make these special laces – seven vendors failed before they found the right combination. 

Because product ideation focuses on qualitative data, such as observing general behaviour, it is generally associated with the data-inspired stage. However, we often confuse it with the data-driven stage, which led to Steve Jobs saying: “Our job is to figure out what they’re going to want before they do. People don’t know what they want until you show it to them. That’s why I never rely on market research.” Jobs was right to say that traditional market research is not ideal for product ideation; it is data-driven and people can rarely articulate what they want. However, Jobs and Apple do focus on how people interact with their product – which is why Apple continues to send feedback surveys.  

Asking consumers what they want doesn’t lead to many insights. Taking their feedback by “showing” ideas, as Jobs says, is far more meaningful. A/B testing is not helpful when creating a product but is vital to its continuous improvement and evolution. It intersects the data-driven and data-informed thresholds – you test various hypotheses but focus on specific data points. Airbnb and Netflix arrived at their current designs not only by running millions of A/B tests but by hypothesising which variants would be the most user friendly for their customers. This is how Airbnb came up with their map-based listing interface and Netflix implemented left to right scrolling.

These thresholds apply to advertising and marketing. Old Spice’s ‘The Man Your Man Could Smell Like’ campaign was data-inspired by a relatively straightforward insight: 60% of body wash purchases are made by women. The agency leveraged this insight (data-informed) to provoke a conversation between couples about body washes. A remarkable accomplishment of the campaign was how, by diversifying platforms and integrating TV, online and social media, they were able to increase the number of customers who received their message. Ultimately, this data-driven campaign not only revitalised the brand, but also helped double Old Spice’s sales.

Moving beyond the data-driven threshold also forces one to look at other sources of data and although big data is extremely important, we sometimes overlook the relevance of small data. Economists not only track key financial and economic indicators, but they also measure the ‘feelings’ of a population to help predict recessions and inflation – this is what led to the development of the consumer confidence index. Although many people feared inflation in the US when the government decided to send out stimulus cheques, people did not immediately spend the extra money due to the uncertainty that arose from Covid-19. Inflation happened more than a year later once people felt confident enough to spend the extra money.

Each of the three thresholds has its limitations. Since the term data-driven implies answering a very specific question and ignoring the bigger picture, it should not direct new strategies or design thinking; instead, it should be used to validate a solution or product. The data-informed approach allows one to understand past failures and successes to drive new strategies – but one has to be wary of confirmation biases, as it is easy to cherry-pick data that reinforces a narrative. Similarly, the data-inspired method should never be referred to as concrete data, as the trends showing up could be seemingly related interactions but are not statistically correlated.

It is likely that you already use these concepts in your line of work without explicitly stating them and you may even argue these are old concepts dressed up in new clothes. The key takeaway is that there is no linear path that leads to the ‘Aha!’ moment of insight and discovery, rather it is a circle. Sometimes you may have an intuition or hypothesis, other times you may observe a behaviour or trend and then there may be occasions where you come across an indicator or data point that may lead you down the path of an insight. 

Ans Khurram is an insights professional working in the telecommunications industry. anskhurram@gmail.com