By Ziad Nejmeldeen, Director of Science, Quantum Retail
For retailers, turning customer analytics into actionable business decisions can be a daunting task. This is especially true for the largest retailers, where transaction data streams in for tens of thousands of products in thousands of locations, possibly from across the world.
In order for retailers to make use of analytics in targeting an optimal range of products, they need an acute awareness of product behavior. Since each store has dozens of unique product behaviors, retailers need to look at product attributes like package size, brand, value, price point and store attributes including demographics, location type, size, etc. While understanding this information on a store-by-store basis can be downright tedious, it is an essential part of measuring product performance, customer behavior across stores, how products behave similarly and how important they are to the categories in each of your store clusters or grades.
Where to begin
One of the best places to begin assessing your performance metrics is to look at customer buying patterns. To do this, you must begin at the cluster level and narrow down customer behavior by store size, demographics, region and traffic. Then you can begin to look for patterns within categories, brands and products; the more granular you can get when analyzing these behaviors, the more efficiently you will serve your customers.
Examples of product behaviors to compare:
Seasonality – How do your stores differ in product sales seasonally? Which seasons do each of your stores perform best in? Do these patterns align with store size, demographic or region? Do certain styles or colors perform better in specific regions?
Time of day – Did you stock out? When? What did that mean in missed opportunities for sales? How quickly can you replenish again?
Day of week – Does this location have a weekend traffic boost? Does that product respond to the pattern?
Weather impact – Do products react differently on cold days or wet days? What does that mean to demand? Should that affect how stores are supplied?
Market basket – What products typically sell together? Does this change by region or demographic? Does the supply of these market basket items correlate?
Price points and promotions – When your prices are lower or higher – how does it affect your sales and how do your price points affect your margin? Do certain promotions perform better by region? Are your promotions creating loyal customers? How does your product performance compare during a promotion vs. at full price?
5 ways to turn customer analytics into action
1. Focus on the customer- The customer should be the most important element in your retail strategies. Retailers need to focus on real-time local demand — this means creating a dynamic inventory plan that is highly reactive to local demand fluctuations, allowing the retailer to be flexible and respond to how their customers are behaving now.
2. Set objectives- Each product should have a role with specific objectives that can be measured and executed against. A product may be in your assortment to drive traffic, generate profitability, present an image or opportunistically acquire impulse sales. Each of these roles come with unique objectives that can result in different inventory requirements.
3. Shift focus- While forecast accuracy is important, it is not the only way to improve inventory placement. If you are adjusting forecasts to achieve different inventory results, you’re already reacting to this fact. Shift focus to find the best way to utilize inventory to achieve goals while understanding that forecast accuracy and variability are realities.
4. Get local- There is no substitute for understanding product behavior at the store level. Ideally, you have a host of customer data that lets you not only map customers to purchases, but also links the changing customer buying strategy with your product strategy, and this may be different by location. It will be important in this changing environment that these product/location strategies are continually monitored and updated.
5. Revisit and rationalize– Product behavior constantly evolves with the changing consumer. The item that fulfilled its role last year or last quarter may not be doing so now. You need to be alerted to situations where this change is happening and have a mechanism to understand and react to the way that impacts your offerings to customers. The initial assumption of the product assortment is an important part of the process. This enables retailers to understand which stores will offer the greatest potential for full price sales — and appropriately decide what inventory is best and where.
Creating an efficient and intelligent execution process
Understanding your customers is key to increasing your profit. This means that retailers need to take the generalizations out of their methods and create a way to react to each cluster – and ideally, each store. You can begin to analyze and optimize your inventory execution by comparing store and cluster metrics, but the most efficient and profitable way to turn your data into action is by investing in technology that can react to shopper behavior automatically, giving you real-time visibility of consumer demand.
After receiving his PhD in Econometrics from MIT, Ziad spent 3 years at ProfitLogic working on retail optimization solutions including Planning, Assortment, Allocation, Size, Pack, and Markdown Optimization and another 3 years at Oracle Retail managing the science team responsible for Store Clustering, Size and Pack Optimization, Regular Price and Promotions. Ziad joined Quantum Retail in February 2009 where he leads the Science team.