After cutting through hype about the role of artificial intelligence in retailing one thing is certain – it will drive a stake through the heart of linear merchandising processes and managing by averages.
Here’s why: Merchandise managers simply cannot keep up with the speed of omnichannel consumers and competitors using traditional methods. For example, in a chain of 200 stores, it will take more than 18 hours just to open each store plan for five minutes and make a few rapid-fire adjustments. How much can a merchandise manager do in five minutes? Not much. But those five minutes per store takes more than two days of work. If the stores need 10 minutes of work then the job will take nearly a week.
And that’s just for 200 stores. In a chain of 500 stores it will take more than 45 hours to work five-minutes per store and in a chain of 1,200 stores it will take more than 100 hours. Again, a manager can’t get much done in five minutes so these hours (although surprisingly high) are actually low.
In the RIS Targeted Research report “The Future of AI-Driven Merchandising” implications of using AI to transform linear merchandise management processes are examined and recommendations made for setting a path for AI implementations.
Law of Averages Enshrines Loss
To get around using excessive labor hours to manage stores individually, retail chains simply punt, i.e. they don’t do it. Instead, they manage stores by groups. In a chain of 500 stores, the retailer will create four to five groups of about 100 each to cut down on time needed for merchandise management functions.
The problem with this approach is obvious – no two stores are alike and treating them alike produces a measurable financial loss even when a category manager hits targeted sales goals.
The reason is that in every group of stores roughly half will have over performed and half under performed. The under performing stores will have left money on the table because adjustments were not made to account for different customer bases, location characteristics, physical sizes and shapes, employees and planograms.
Using an AI engine will address the problem because it will manage each store individually, simultaneously and holistically.
Key Takeaways from the Report
- AI in the merchandise management function is not a reality today for most retailers. The study shows 56% currently have no AI engines in place and only a quarter are currently testing or piloting at least one AI engine.
- However, there is a dawning sense of urgency as a big majority (69%) believe AI will play an important role in hitting financial targets in the future and a quarter of these say it will be a major factor.
- A big majority say the top two strategic goals they want to achieve using AI in merchandise management functions are improving sell-through of inventory (59%) and improving margins (56%).
- Three specific applications are critical for meeting merchandise management goals and therefore will benefit most from using AI engines: demand forecasting (56%), replenishment (44%), and price management (41%).
- Retailers that have AI in place today or are testing/piloting it now will have the technology to themselves for about two years, according to a finding that nearly two in five retailers have no plans to invest in AI-powered merchandise management at all and just 19% plan to invest in the next two years.
AI enables merchants, especially those at larger chains, to weigh the value of economies of scale against personalized assortments, identify meaningful patterns in customer demand, price sensitivity and omnichannel complexity, and enable retailers to hit the most profitable sweet spot for each store while catering to customer needs.
To read the full report, with a complete set of charts and recommendations, click here.