Customer In-Store Analytics for Data Science Noobs

‘Do you know anything about store & customer journey analytics?’ — asked an AI store analytics startup founder.

Uh… No, not really + I’m a newbie aspiring data scientist — So if you’re like me, then I guess it’s time we find out about this new technology and get ourselves out of the Noob-zone.

Feel free to go along, comment additions, and expand this article!

Did you ever watch that Vox video about how the store layout can incentivize you to buy more? Store analytics means the process of measuring and analyze the performance of every aspect of a store to optimize operations and create sustainable revenue increases.

In the last decade, store analytics had brought an entirely new approach to how stores can have more control over customer purchases and leave less to chance. Now, with the power of technology and AI, stores can have real-time analysis and more complex data analysis at the tip of their hand. Below, I explore the following topics.

Content

(General knowledge)

1. What is store analytics: key definitions
2. What are the tools used for store analytics
3. Typical business owners’ questions about their store analytics

1. What is customer in-store analytics: key definitions

Image from Walkbase. In-store analytics companies aim to be the one-stop-shop for processing all store data and produce analytics and services and assist business growth.

• In-Store (or Customer In-Store) Analytics refers to the process of collecting and finding patterns in brick-and-mortar retail stores that offer insights to optimize its operations.
• In-store Devices: The advent of heat map technology (like this) and the cloud helps many stores track the physical customer experience in their stores and store it in the cloud to optimize data processing and analysis. From customer footfall to their interaction with each product category, store owners can capture all in their video footage.
• AI is a new and exciting component to optimizing store operations. Some examples of AI applications for in-store analytics are pricing adjustment and suggestions for product combination placement. With AI, we can automate the process of collecting the most common combinations of product-price or product-product pairs in the store data and trust that our computers are better than us in suggesting new adjustments — and works faster.
• Weather forecasting: Especially for multinational retail companies, or chain stores that expand over various cities, it is useful for managers and store owners to have an overview of all the weather forecast of all their store locations over a specific time period to optimize business opportunities related to supply chain, retail pricing, marketing opportunities. IBM explains this in detail.
• Point of Sales (POS): is the exact place in the store where selling happens — the cashier, or any place that the customer makes a purchase from a company. A POS system is often used in the restaurant and hospitality services industry. It essentially captures financial information reports on revenue, sales over time, warehouse stocks, and record customers’ contact details. Read more about POS here.
• Marketing optimization for brick-and-mortar stores is different for online shopping websites. Online shopping companies can track website traffic as customer traffic and clickthrough rate to measure the effectiveness of their marketing campaigns. For physical stores, the staff might have to physically count the number of people visiting their stores on a sale period to know if the campaign was successful or not, which not only difficult but not a productive task for the staff. Hence, video footage allows store owners to see how many people visited the store from one campaign period to another with more precision.
• Inventory and Staffing: Store analytics will also let the store owner knows when to stop advertising for a particular product when it runs out from the warehouse or to get more if the product sells well. As for staffing, the store owner will know who performs well and have the highest number of customer engagement and successfully convert that customer to purchase a product.

2. What are the tools used for customer in-store analytics

This is a non-exhaustive list, but I try to capture the main technologies utilized in the space of customer in-store analytics.

Great image from RetailNext.

• Video cameras (device)
• Heat maps (tool)
• Weather/temperature scales (tool)
• Marketing software/SaaS
• Employees software/SaaS
• Data manipulation software: Python, R, etc.
• Data visualization software: Tableau, PowerBI, etc.

3. Typical business customer in-store analytics questions

Image from Ipsos

Product (sales and inventory records)

• Which product sells the most/least (aggregate data)
• Which product sells the fastest in a campaign (aggregate data, time limit)
• Which product sells the slowest over a year (aggregate data, time limit)
• Which product is often bought with another product (i.e., accessory, complementary products)
• Which high-value product sells the fastest/slowest?
• Which product could not sell even with price reduction?

Store layout (using video footage, heat maps)

• Which location correlates with the highest amount of footfall (capture rate)?
• Unique footfalls, travel paths, dwell times (image 2)
• Which store layout encourages customers to stay the longest?
• Which storefront display attracted the highest number of customers (A/B testing storefront)

Sales personnel and conversion rates (

• Which sales staff generated the most/least revenue weekly/monthly/yearly?
• Which staff had the highest conversion rates (convert store visitors to buyers)? (efficiency/productivity)
• Which staff often sold the highest value products? (value)
• Which types of information incentivize customers to buy (conversion rates)

Customer segmentation (customer data)

• What are our easy-to-convert customer segments (by age/gender/location)
• Do our customer segments each has favorite products/product combinations?
• What are the shopping habits (time/products/sales personnel engagement) of the customers?
• How do customers react to physical/digital marketing campaigns?

AI solutions

• Which product combinations are the most appealing to a different segment of customers?
• Where do we store data from multiple cities/countries cheaply and still maintain fast data performance with easy access from different locations?
• How do we integrate all data from heat maps, video footage, and other SaaS products to create monthly reports on metrics we care about, in visually appealing formats?
• What are the common high-value product combinations?
• What is the most optimal pricing for each product category?
• Can we satisfy customer pricing preferences and still make a profit?
• What can potentially predict a dip in purchases/visits/footfall/etc. over a certain period?

Do you feel like less of a noob? I do. But the important thing about noob-dom is that no one will tell you when you stop being a noob — you can only try your best to trust that you are no longer ignorant.

The next step I’d suggest to you and myself is to go get some real-life experience exploring in-store customer data. Follow me for updates about my in-store data analytics project soon!