COVID-19 has changed the consumer and retail forever. Traffic will be lower, traditional services like returns and the fitting room will need to be completely re-thought, and full omnichannel capabilities will be the baseline expectations. Stores must be more efficient, smarter, and agile if they are to survive and overcome the challenges of this new operating environment.
In order to do this well, store leaders will need to identify the right data, organize it, interpret it, find patterns, and then make the best decisions to create the right labor budgets, schedules, and daily action plans needed for stores to operate at peak performance during each shift.
Retailers can choose one of three analytic paths as they look to improve their store performance.
Descriptive analytics provide static, backward-looking data and answer the question of “what happened.” Reports, dashboards (mobile or otherwise), flash sheets, etc., are examples of descriptive analytics. The advantages of this approach are that the analysis is fairly simple to produce and can be standardized. In addition, they’re very familiar with leaders at HQ who are used to numbers, reports and analysis. As a result, this is how 90% of retailers operate. The core weakness, however, is that this approach requires the end user to interpret the data correctly, draw the right conclusions, and take appropriate actions. At the store level – where managers are balancing untold priorities and whose interest in retail does not come from a love of Excel – this breaks down completely.
Predictive analytics is the next level of analytic maturity. Here, retailers take the output of descriptive analytics and look forward to predict what they think will happen. Trend reports and re–forecasted sales targets are examples of this. Certainly, updating forecasts is a good thing to do and helps take some burden off of the end user to interpret past results. Predictive analytics is an important step in the right direction, but there are still major weaknesses. In most cases, this analysis is neither broad enough, nor frequent enough to impact actions. The analysis typically only looks at sales and sometimes traffic. Rarely does it look at all the other key metrics that drive the business (conversion, basket size, labor hours, sales productivity, etc.). As a result, the actionability of the new forecasts can fall short. This is doubly true when the analysis is only refreshed quarterly or monthly. What a store operator needs is a real–time update on what is expected to happen today – specifically on my shift – and what I need to do about it. Only with this do you have real data driven decision making at the store level.
Prescriptive analytics offer a new way for retailers to look at data. This type of analytics takes the burden off the store operator to read, understand, interpret, and make the right conclusions from all the reports they’re being sent. Instead, this analysis is fundamentally forward-focused, offering a detailed view of what’s likely to happen. The big distinction from predictive analytics is that prescriptive analytics then identifies the implications of a forecast and specifically suggests the actions needed to best respond to what is likely to happen in the future. Essentially, prescriptive analytics standardizes what a great store operator already does – looks at the data, sees what it can tell you about the future, understands the implication to your business, and then devises a plan to avoid risk and capitalize on opportunities. There is a catch this approach, however – it’s hard to do! You need sophisticated AI skills to model the data, and sophisticated IT skills to organize the data and have it operate at scale and in the timeframes you need (aka near real–time). Importantly, this is also not a one-time project. The AI needs constant attention to learn and improve, as does the supporting IT infrastructure. Retailers are not well-suited to either of these capabilities, and so the only viable option is to partner with companies that are.