A five-step guide to becoming data-driven

In today’s hyper-competitive retail market, successful decision making needs to be founded on data. Now that readily available tools can assist in making data-driven decisions, retailers can no longer afford to rely solely on instinct and experience. And yet, many organisations feel that they are lagging, and consider the shift to becoming more data-driven a massive and near impossible step.

The good news is that most organisations can benefit from some straightforward first steps. In this article, I’ll explain why becoming data-driven is important and then walk you through five steps to becoming more data-focused as a business – regardless of where in the process your current operations sit.

Related Articles

Customer experience is – of course – a key brand differentiator to retailers. Improved use of data can improve the customer experience. Data informs your organisation about the products and services that are favoured by your customers. This helps predict what they might need or want in the future. Studies have shown that 91% of customers are more likely to purchase from a brand that recognises, remembers, and provides them with relevant offers and suggestions.

Taking this into account, retailers are discovering innovative ways of using the data they already possess, such as personal information, customer feedback, or purchase history, to transform their customer experiences. As a result, data-driven businesses are 23 times more likely to outperform their competitors in acquiring new customers and six times more likely to retain those customers, with a 19 times greater likelihood of achieving above-average profitability as a result.

However, it’s not just about improving your business’ offerings; data-powered tech can significantly improve in-store experiences through ensuring that supply of any kind (product stock, customer care levels, etc) are sufficient for customer demand. In a study we conducted at Rotageek, we found that 41% of retailers often wrongly predict the number of staff needed in-store. That is a forecasting and demand-matching problem that data-driven tech can solve.

Despite high levels of interest in becoming more data-driven, for most retailers it remains a long-term goal rather than a reality; only 31% of companies have restructured their operational processes.

For those that want to take action now, here is a step-by-step guide on how to start (or improve) your data-driven approach.

Prioritisation of data

If you are like most retailers, you are already collecting more data than you can use. Initially, you’ll need to prioritise the data streams that are most important for your business. Where datasets already exist, you will need to identify any quality or access challenges that prevent these datasets from being useful. Where data are not currently being collected, you’ll need to figure out how to begin collecting that data, how to store it, and how to ensure it remains high-quality. Identifying which data are most important to satisfy business demands will focus data efforts, reduce the amount of time it takes to deliver tangible benefits, and prevent resources from being spent on valueless data.

Clean data

Data quality is viewed as a key challenge by 50% of companies. In order to start creating valuable analytics, it is essential for you to clean up existing data and ensure it is only stored in high-quality form. High-quality data is accurate, complete, timely, valid and consistent. Any breakdowns in data quality can greatly reduce the value derived from reporting, analytics, or data science. To ensure that your data remains high-quality, each collected dataset should have an identified maintainer and should only be collected if it is going to be actively used. Otherwise, it is all too easy for quality issues to occur without anyone noticing, rendering years of data unuseable for meaningful analysis.

Build automated reports and analytics

Once you ensure that the data are high-quality, the next challenge is to use that data to make decisions. When this happens, data become a key asset that underpins the objectives and strategies of an organisation. By the time you have reached this stage, the data storage processes and frameworks are already established.

When you initially prioritised data in Step 1, you decided that some datasets were important to drive decisions about their business. At this point, it is time to automate the reporting process, so that decision-makers can receive real-time insights from these datasets to influence their decisions. This will enable your company to effectively grow, create, optimise, and protect value across an array of operational areas.

Leave space for creative analytics

By this stage, you have answered some of your highest priority questions using data. But, your data use needs to keep evolving. You need to continue to search for new data patterns in the datasets you currently collect and identify additional ways that these data can inform decision-making. This requires stepping back from usual analysis once in a while, thinking more holistically about the data you have, and maintaining a creative approach to analysis. These exploratory projects should be hypothesis-driven, which will ensure that the focus of each project is small and that it will quickly deliver value without using many resources.


The most important step of all is to continue to iterate. By focusing initial efforts on something small, you can derive a tangible benefit much sooner. But, once those benefits are realised, you should already be thinking about the next most important datasets to analyse. By doing this process iteratively, you can develop a consistent and efficient process for collecting and storing new data and ensure that you don’t stagnate when it comes to improving operational efficient. It is important to move quickly and constantly evaluate data, repeatedly.


To conclude, it is not too late to become data-driven, and it is not as large a task as it seems. It is important to remember that this is a step-by-step process that needs to remain agile. How do you balance keeping the projects small, while ensuring that you collect enough data to deliver value? Start off with a plan to build automated and creative reporting on a small set of data that answers questions currently asked by decision-makers. From there, you can track the benefits of using data in your decision-making and iteratively expand your reporting to include additional data.

By Daniel Chamberlain, senior data scientist at Rotageek

Back to top button