Partner Amanda Astrologo recently led a discussion based on a Parker Avery point-of-view titled “Future-Proofing Retail: Building a Solid Foundation for Tomorrow.”

Amanda stressed that retailers must ensure foundational systems and capabilities, as well as a solid change management program, are in place before embarking on new initiatives and innovations like advanced analytics and artificial intelligence.

In this post, Amanda shares her perspectives on what it means to ‘future-proof’ and be prepared for retail’s shiny objects, as well as implications for retailers.

What does ‘future-proofing’ retail mean?

It means making sure you are foundationally ready. There are so many ‘shiny objects’ in today’s technology arena like artificial intelligence (AI), machine learning (ML), and of course, advanced analytics. Every retailer is trying to get ahead—and in many cases just keep up or stay viable. It’s easy to get caught up in today’s retail whirlwind, but so many of our clients are simply not prepared for the fast lane and not taking the critical initial steps to truly assess if they are ready. This means not only having the necessary support systems in place, but also from an organizational perspective—the people and roles necessary for success.

It’s also important to have the foundational business processes in a good place. If the ‘simple’ things—basic retail block and tackle business activities like merchandise financial planning (MFP) or buying—are done differently in each area or the solutions are heavily customized, there is a significant risk in being able to do these well in an omnichannel world where everything is connected and customers expect seamless immediate experiences. It’s great to want analytics, but you need to have a place for the data to be housed and be utilized easily and across your entire business. Further, you need to have effective data governance in place. Otherwise, it’s just data…and potentially expensive data.

Why are retailers slow to adopt to new technologies like machine learning and artificial intelligence?

Many leaders think investment in the science is the silver bullet. However, the more they research or talk to experts in this space, they start to understand it’s a much larger undertaking. Regardless of how AI and ML are touted to be integrated into modern retail systems, there is still very much a human factor that must be taken into consideration. There must also be acute understanding about where the data comes from (and ensuring it’s clean) and how it will be utilized in downstream business processes. This is where the business process, roles, and responsibilities come into play and are extremely critical for success.

What must retailers do to get ready for the ‘shiny objects’ like ML and AI?

To successfully adopt any new technology, your data needs to be accurate, clean, and well-governed. Business processes that are core to your operational strategy should efficiently support your desired capabilities. It takes some work, but you may need to take a step back and really assess where gaps may be. Standardizing and streamlining across the business units where it makes sense will help with user adoption and ease training. I would also mention having a solid implementation plan, skilled implementation partners, and a strong change management (OCM) strategy. Change management is one piece we hear many companies say, “We can cut OCM out and still be ok.” I would 100% disagree—while you may be able to implement an AI or ML solution out of the gate, there is a huge sustainability risk with respect to user engagement and adoption.

Is there an area of retail where you see the heaviest investment activity in newer technologies?

We’ve been working with many clients implementing advanced analytics and AI in forecasting, demand planning, and merchandise planning (MFP). The focus is on accurate forecasts that are universally adopted by the company’s different functional areas, as opposed to multiple forecasts used in a single company. In these areas, it makes sense to give users visibility to some of the science, because it helps with user adoption (meaning, getting them to trust the system and move away from their beloved spreadsheets). The users understand how the system derives a forecast, and they can choose whether or not to accept it, or make select changes to some of the demand drivers.

Starting with MFP is far easier than trying the ‘black box’ approach which is the case with some allocation or replenishment solutions. Those solutions also have other levers that can muddy the waters for a user when you start talking forecast accuracy.

The hardest part is always user adoption. You can have the best data and the most advanced software, but if the user (and leaders for that matter) don’t adopt…it’s dead in the water.

Amanda Astrologo, Partner