Market Basket Analysis (MBA), also known as affinity analysis, is a data mining technique. You can use MBA to uncover meaningful correlations across products and customer purchase patterns.
At GoGoGuest, Market Basket Analysis is a module that is part of our Analytics and AI platform. It helps brands predict customer behavior and find purchasing patterns.
But if you’re a restaurant-type business, you may find yourself asking: what is Market Basket Analysis? And even more importantly, how can it help you? Large and multi-location hospitality, quick service retailers, retail brands, and restaurants use Market Basket Analysis to unlock information at the purchase level.
It functions by analyzing which items are most often purchased together, even if they are unusual combinations. For example, Market Basket Analysis might show that your customers buy quinoa bowls and sweet potato fries together. This information helps you understand the profitability of the combination, how frequently the items are purchased together, and the revenue the combination brings to your business vs. other menu items and combinations.
You can use what you learn to drive changes and grow your business. Market Basket Analysis will give you fast, actionable insights that help you increase your profitability. Let’s dive in and explore more!
Market Basket Analysis helps businesses understand product correlations by predicting consumer purchasing behaviors. It is especially useful for identifying unusual trends that would otherwise be impossible to recognize.
You can use these insights you learn from Market Basket Analysis to sell more, identify cross-selling opportunities between your locations or stores, drive profitability and manage inventory waste or demand.
For restaurant-type businesses, Market Basket Analysis will take you from reactive to proactive. You’ll have the tools to:
There’s no limit on how you can use the insights gleaned from Market Basket Analysis. You can draw insights on a location-by-location basis or by ordering channel, so it’s just as useful for digital and takeout businesses as it is for brick-and-mortar retailers.
You can also use Market Basket Analysis to drive product recommendations, exactly like how you see product recommendations each time you browse a product on Amazon.com. Amazon offers you products that they think might interest you, based on what other consumers with similar profiles have bought in addition to the product you are looking at. You’ll be able to do the same.
You can also suggest a menu item to a customer who is browsing your website or app when they’re looking to place an online order or order for delivery. Marketing Basket Insights can help you cross-sell more effectively and make digital ordering channels smarter.
Market Basket Analysis is about understanding large sets of data and applying rule sets. There are two types of MBA:
Predictive Market Basket Analysis forecasts which items are frequently bought together. It is mainly used to leverage cross-selling opportunities. For example, if lots of customers with a certain profile buy specific products together, you can offer them selections or choices that they are likely to find highly appealing, increasing cross-selling.
Differential Market Basket Analysis is broader. It can include what your customers buy, when, if a specific category of customer buys certain products and if there are monthly or seasonal differences in what they purchase. Comparing the results, differential market basket analysis lets you identify exactly which factors are responsible and so on.
Understanding the purchase patterns of your customers ultimately helps you customize your offerings. The more customization you can offer and the more you understand what your customers want, the better experiences they will have and the more they are likely to spend.
Market Basket Analysis works by analyzing the purchase patterns of certain products that are consistently bought together by customers with similar profiles. An algorithm determines connections between two or more different variables (in this case, purchases).
It works a little bit like this:
37% of your customers who purchases a large latte will also buy a strawberry donut, they might also want to buy a chocolate covered donut, increasing the average order per visit (AOV) to $12.50. 🎉
When the chocolate covered donut is not available, 5% of the large latte and strawberry donut consumers do not buy a second donut, adjusting your AOV to $8.75.
Meanwhile, 10% of the large and strawberry donut consumers downgrade to a large latte and a plain donut when the strawberry donut is not available, adjusting your AOV to $7.75.
The maths and formulas behind Market Basket Analysis are pretty complex (thanks for doing the hard work for us, AI!) However, at its most basic, the algorithm will work out if there is a statistically dependent connection between two or more items. The fact that Market Basket Analysis links connections through a formula is important. When you sell multiple products in high volumes, you can get random connections. This can be a result of the repetition of items bought at the same time but without there being a statistical pattern backed by consumer preference.
Some correlations are obvious. For example, your data might show that health-conscious customers that purchase salad might also be inclined to buy a smoothie and an energy bar. Other relationships might not be so apparent: think back to our previous example of how customers might commonly buy quinoa bowls and sweet potato fries together. And finally, you’ll learn the combinations and products that repel each other. For example, you might learn that customers who order a veggie burger are highly unlikely to also purchase ice cream.
Let’s say you run a coffee shop, and you have an item on the menu that isn’t selling as much as it did when you launched it: a triple chocolate chip cookie, for example.
Your data (i.e. triple chocolate chip cookie sales) shows that you’re only making $1,000 a month from this type of cookie, which you’re selling for $2 each. All your other types of cookies, sold at the same price, are making $5,000 a month.
Initially you might think of taking the triple chocolate chip cookie off the menu as it looks much less profitable than your other cookies and it’s selling much less. But using Market Basket Analysis, you see that 60% of your triple chocolate chip cookie customers are also buying a large latte and a soda, bringing your business an additional $7 in revenue each time they buy, totalling an additional $4,550 in revenue every month. On the other hand, the analysis shows that customers buying other cookies aren’t making any additional purchases: they’re just buying a cookie.
Most restaurant-type businesses use Market Basket Analysis to increase profitability and make smarter decisions around the products you offer.
You can turn data into useful insights that drive your business forward. Change won’t be trial and error and you can make decisions confidently, knowing they are based on solid customer data. Market Basket Analysis is primarily used to:
If you’re a restaurant-type retailer, you probably have lots of data, but you possibly don’t know how to use it to boost profitability. You may also have to spend hours manually sorting and analyzing data from multiple sources, only to come up with a few ideas that you’re unsure of. Market Basket Analysis is a quick way to analyze your data to help you use it to meet your business goals.
As well as the primary use cases listed above, Market Basket Analysis has some other unorthodox uses as well. These include:
Fraud detection / criminology. Companies can also use market Basket Analysis to detect potential cases of fraud. The algorithms used in Market Basket Analysis can analyze the connections and relationships between events to pick up similarities or repetition. These are primarily used in finance and insurance.
Web or email filter. The same algorithm used to understand connections and relationships can also understand the difference between spam emails and ‘normal’ email traffic. It can be used to sort emails automatically into various folders.
Customer behavior. Market Basket Analysis helps all types of companies get a better understanding of customer behavior. That can be from understanding unusual items that are bought together to knowing more about what’s driving customer loyalty, why customers buy certain products, and what customers enjoy most about their experience buying from you.
Pharmaceutical bioinformatics. Association rules (the maths behind Market Basket Analysis) are often used in the pharmaceutical and bioinformatics industries. Companies, pharmacists and biologists can use association rules to categorize different groups or products, patients, plants or animals based on their similarities or connections.
By now, you should be starting to see just how Market Basket Analysis can help your business. The insights you gain will be concrete and easy to understand. You’ll be able to make smarter decisions with more confidence.
The major benefits of Market Basket Analysis include:
If you’re like most restaurant-type businesses, you probably have a lot of data on sales, revenue, and inventory. You’re also probably struggling to understand how to use your data and how it translates to actionable insights. Market Basket Analysis will help you take the actions that move the needle on your business. From understanding what’s driving profitability to identifying easy cross-selling wins, Market Basket Analysis will give you all the answers you’re looking for.
When you understand your customers and their purchase patterns better, you can create more appealing and customized promotions. This can be at the customer level, offering a deal or product pairing you know customers will like. Or you can offer more standardized promotions that offer frequently bought products together.
Pinpoint exactly which products and groups of products influence sales. Understand what part these products play in profitability, how to use them to encourage cross-selling or upselling, how to manage inventory and how to use them to your advantage in things like menu design.
Whether you have a digital store or a brick-and-mortar location, you will be able to improve the placement of your products. You can use Market Basket Analysis to place keystone products front and center, provide recommendations, put products that are commonly bought together next to each other, and so on.
Some menu combinations are obvious. Others aren’t. Market Basket Analysis gives you the whole picture. You’ll uncover more pairings or product associations and understand more about what customers want, why they are buying from you, and how these associations drive revenue.
What are keystone products? Keystone products are usually the lead products in your menu or product catalog. Keystone products aren’t simply ‘popular’ products. Instead, they are items that are usually bought together with other items in your menu, or they are products that are purchased for a very specific reason.
That can look something like this: a certain group of Starbucks’ customers are calorie-conscious. These customers usually buy Starbucks’ Nitro Cold Brew. Why? Because a Nitro Cold Brew Grande is only five calories. The unavailability of a Nitro Cold Brew could turn away calorie-conscious customers.
Knowing what your keystone products are can help you increase sales by cross-selling them with your other products. If the item isn’t available, it can also help you offer an alternative that will keep your customer happy. For example, an alternative keystone product is the Starbucks Reserve Cold Brew. Not as smooth as the Nitro, yet it meets the calorie count.
There is a science in merchandising design and display that is the result of Market Basket Analysis. When a restaurant, quick service or retail brand understands which products are frequently bought with a keystone product, it’s a natural extension of this to display a combination of add-on products in the line of sight of a customer.
For online menus or eCommerce sites, upselling and cross-selling is much easier with a recommendations engine that is powered by AI. When you visit Starbucks online, for example, you’ll often see pop-up recommendations of add-on products to go with your coffee drink.
Upselling and cross-selling in-store is often easier since the merchandise is right in front of the customer. Upselling and cross-selling is much easier to execute and drive online, but equally, it is also easier for a customer to ignore recommendations in a digital environment. This is why data-driven merchandising and recommendations are essential for brands.
Maintaining optimal stocks in the inventory becomes super-efficient when you know your customer purchase patterns. You’ll know what you’re likely to sell on any given period or day of the week.
When you understand variables and rules (including a customer’s purchase pattern and engagement), applying rules, specific product combinations and offerings could increase frequency and spend.
Customer engagement driven by deep analytics and market basket analysis can:
Analyzing customer’s response patterns from various ad sources or media can remove the guesswork on which ad channels are driving engagement and conversion.
Starting a restaurant concept or looking for help to grow your current restaurant?
Join our newsletter for actionable tips on how to start and grow your business.
It all starts by asking the right questions to find your most valuable answers and insights.
How To Use A Captive WiFi Portal To Market Your Business If you’re the owner or operator of a restaurant-style business, coffee shop, QSR, small …