Dr Linda Robinson warns that we must remember to view data and analytics simply as a tool to help address well-defined problems.

On a recent holiday I was devastated to learn that, somewhere in London Heathrow, I had lost my faithful Fitbit. The device that had been securely attached to my wrist for the past 18 months was gone, and I no longer had a way to monitor how many minutes of precious sleep I enjoyed each night or compare the number of steps involved in a shopping trip to Selfridges to a wander around the British Museum.

The despair I felt was enlightening. I had become so consumed with measuring my activity each day, but that was all I was doing; I wasn’t actually consuming the data. Without my Fitbit I took the same number of stairs and walked the same distance, just without the digital commentary on my wrist. The previous 18 months of constant activity data had encouraged me to form some better habits, but without the internal motivation to actually engage in physical activity each day the Fitbit wasn’t going to make me a healthier person just by having it on my wrist.

This is the same message I deliver to my marketing research classes each semester: data and analytics are tools that can help marketers make decisions with less risk and uncertainty. They should not, however, be the driving force of your research or decision-making. Data and analytics by themselves cannot ensure the success of a marketing campaign, no more than a Fitbit can ensure you complete that half-marathon you signed up for.

This is not to say data is not important – very important – to marketers. Data allows marketers to measure variables more effectively. It allows us to better predict the promotion consumers will respond best to, or what web page layout will motivate a customer to buy our product. Data allows us to make better predictions, and better predictions lead to better outcomes. But before rushing to acquire more data, we must make sure we know why we are collecting the data and how it will be used. In the midst of acquiring, analysing data and creating infographics, it is very easy to lose sight of the problem driving the research. He may not have been thinking of marketing research, but Albert Einstein encapsulated the challenge perfectly: “The formulation of a problem is often more essential than its solution.”

Thus, data and analytics should be purpose-led, a tool that will inform the solution to a well-defined problem.

This message should be nothing new. The industry is well-versed in the promise of big data and predictive analytics, with examples such as Target’s pregnancy prediction model now standard practice. The buzzword ‘Big Data’ itself is virtually passé, and today many firms have business intelligence strategies in place, with data scientists delivering the right information to the right people at the right time. So why are we still talking about data, big data and analytics? Gary Marcus and Ernest Davis in their New York Times article, ‘Eight (No, Nine!) Problems With Big Data’, illustrate why the conversation is ongoing. Big data and analytics need to be seen as resources and tools, not a ‘silver bullet’ solution, and we should be pragmatic about what they can – and can’t – do. The prevailing message is that we need to be smarter about how we use data and interpret the analytics while recognising the limitations.

One example of the smarter use of data is New York City’s Mayor’s Office of Data Analytics (MODA). The success of MODA is being put forward as evidence that a similar approach could benefit other global cities such as London. Created by former NYC Mayor Michael Bloomberg, who wanted to prove that the data-driven analytics techniques used by the financial sector could be used to enhance city management, MODA aggregates and analyses data from across city departments to more effectively address crime, public safety and quality of life issues. A chief example of the potential for purpose-driven analytics in city management is NYC’s fire prevention model. Veteran fire-fighters instinctively know what criteria to look for in determining a dangerous building. They know what variables are most frequently correlated with serious fires, and previous iterations of the city’s fire risk model utilised focus group discussions to weight these variables. MODA, however, could complement the gut instinct of these fire-fighters with datasets from other departments (age of the building, type of business etc) and data from actual fires to better calculate the relative importance of each variable. The result was a data-driven model able to predict the buildings most at risk of a dangerous fire with far greater accuracy.

The practical implications for this analytical modelling are immense, such as the immediate benefit to public safety when the most dangerous building can be prioritised for inspection. The success of this approach was clear: using the old modelling approach, the first 25% of inspections revealed 21% of the most severe fire code violations. The MODA model, however, resulted in the first 25% of inspections uncovering 71% of the most severe fire code violations.

Much like Brad Pitt’s character in the movie Moneyball used evidence-based data analytics to assemble a competitive baseball team, Super Rugby team the NSW Waratahs use big data to not only make player selections, but importantly to predict and prevent injuries. With over 9000 data points collected in a single game, and similar numbers from each training session, the Waratahs are able to model the workload of each player. Such information can be delivered to strength and conditioning coaches in order to prescribe individual training programs and, when combined with a medical screening database, player workload data can be imputed into predictive models to ascertain each player’s injury risk profile.

Simply put: New York City used data more smartly to model fire risk, and the NSW Waratahs use data more smartly to tailor injury-prevention training programs for each individual. Both of these examples use passive and active data in their predictive modelling, and both require interventions and actions to achieve their purpose. By itself, the use of big data and analytics will not prevent fires in New York City nor stop a rugby player from sustaining an injury; rather they are tools for the fire inspectors and rugby coaches to use, just as a Fitbit is a tool for me to use in monitoring my daily activity.

Becoming seduced into measuring everything can create blind spots for marketers and take up valuable time, so much so that we may miss an opportunity to create a marketing campaign that would deliver exceptional value to our customers. With so much data available to marketers today, we must be careful not to die of thirst in the middle of the ocean. View data and analytics simply as a tool to help address a well-defined problem. Learn to consume data rather than be consumed by it, and use analytics to illuminate the path you want to be on rather than wander blindly looking for a solution.

“He uses statistics as a drunken man uses lampposts – for support rather than for illumination.” – Andrew Lang

Dr Linda Robinson is a marketing lecturer at RMIT University where she aims to inspire in her students an appreciation for marketing research. While her research interests are primarily centred on services and social marketing, she has a special interest in marketing education and graduate employability.

RMIT University is a Marketing content partner – collaborations with leading organisations on content for the magazine (like this article) as well as exclusive benefits for Marketing Advantage Members. Find out more here.