How to use Product Recommendations to Boost Revenue and Conversion
By Ivan Borovikov, CEO and co-founder of Mindbox
Product recommendations have been hyped since Amazon first introduced them in 2008 and have led to numerous myths such as being a fully autonomous magic bullet to increase profit immediately after being installed.
However, things are much more complicated. Product recommendations can indeed boost sales, but they may also be hazardous for conversion rates. It’s also challenging to measure the added value as many free analytical tools are misleading and can create losses. Here’s how to do it right:
Start with workflows and campaigns, not tools
First, you need to understand which of your customers’ workflows may benefit from recommendations.
In classic eCommerce, recommendations look like product cards on the website. For example, clicking on a T-shirt on a brand’s website will entail other similar T-shirts being recommended and shown to the customer.
Offline recommendations can be shown at the POS to a store employee on their tablet. Also, a call center operator could suggest products based on the customer’s purchase history.
Product suggestions can be fine-tuned to better fit specific business goals or audience segment needs. Recommendations can display only discounted products (to increase unit-per-transaction), and only highlight a specific brand (e.g. high-margin ones) or products which are in excess stock.
-
Online channels
-
- Homepage — popular products;
- Catalog — popular products from the category;
- Product card — related or similar products;
- Cart—related or most frequently purchased products.
-
Offline channels
-
- POS—related or promotional offers;
- Call center — related or frequently purchased products;
- Sales manager’s tablet — bestsellers or product collections.
Pic. 1 A product recommendation example on United Colors of Benetton’s website
It is crucial -that your recommendations be consistent across touchpoints and communication channels. To achieve that, consider a solution that enables you to centrally orchestrate your product recommendation logic across the touchpoints, e.g., a Customer Data Platform (CDP).
The same logic applies to product recommendation algorithms at different stages of the customer journey. You don’t want to promote cheaper alternatives at checkout.
Evaluating Your Efficiency
Different product recommendation combinations affect metrics in various ways.
Popular on-sale products displayed on the homepage can reduce the average order value but increase the conversion rate. Recommending similar, more expensive products in the cart can reduce the conversion rate, but increase the average order value and overall profit.
Proxy metrics (clicks or product views) are collected faster, however, they don’t always reflect success. For instance, one pharmacy retailer, as part of an experiment we launched, added product recommendations to the cart page to increase the number of products per order. Over two days of testing, the company lost $30,000 compared to the control group. It turned out that customers simply started abandoning the cart to view the recommended products and forgot to complete their purchase. So, page depth increased but revenue decreased.
These metrics are most affected by product recommendations:
To simplify the task of compiling a list of product recommendation campaigns, I suggest cloning this Miro board and brainstorming solutions relevant to your business.
Pic 2. Miro board product recommendation strategy visualization
Personalizing Product Recommendations
After determining the list of product recommendations and setting target metrics, you must think about optimizing the quality of your product recommendations. To suggest the most useful items for customers, product suggestions must be based on the following:
-
- Business goals — revenue growth, profitability, average order value, UPT, sales volumes;
- Customer behavior — browsing history, items added to cart and favorites, online and offline purchase history;
- The behavior of similar customers.
Product suggestions will consider the interests of both the business and the customer. The more data is considered, ideally from multiple sources, the more accurate the suggestions can be. If the customer bought a shirt offline, the mobile app would recommend matching pants, because other customers who bought this shirt loved that pair of pants.
The history of a customer’s interaction with the brand and the product range is often stored in several systems: offline sales in ERP software, and online sales in Shopify, while customer actions are on another platform. In this case, we may not know that an online customer already made an offline purchase an hour ago.
Crocs Eastern Europe used to send out email blasts to their entire audience on a weekly basis. Online purchase history, offline purchases, emails, SMS, and web push notifications were all stored in separate systems. As a result, their marketers did not have access to a single comprehensive data source as a basis for campaigns. Customers could receive emails that recommended Crocs that they’d already purchased. When you feed an algorithm inconsistent or incomplete data, you’ll likely receive unsatisfying results.
To solve this, data should be centralized in a single system. The most cutting-edge way to solve this bottleneck is a class of technologies called Customer Data Platforms, which provide a full scope of ready-to-use solutions.
CDPs allow companies to automatically upload customer behavior data from an unlimited number of sources, cleanse and unify data, and get a complete history of customer interactions with the brand, on the basis of which you can launch tailored marketing campaigns, including product recommendations.
This data can also be used to train machine learning algorithms. Algorithms create a profile of the customer’s interests, find similar users and, based on what they bought, recommend the customer other products they may want to buy.
Another bonus of centralization is cohesive omnichannel marketing when online channels consider the popularity of items offline, and product recommendations on the website and in campaigns are synchronized.
Measurement tools
While performance measurement tools differ depending on the recommendation channels, they share the same principle. Customers are divided into two groups. One group receives recommendations, while the second does not. If the sales are higher in the group that gets recommendations, then the customers found the recommendations useful.
Online Channels
I recommend using Google’s free Optimize tool, which is quick to set up and doesn’t require programming skills.
Using Google Analytics, Optimize can utilize e-commerce data to evaluate experiment efficiency via two website versions: one with no changes and one with product recommendations. For each widget, you need to set up an experiment, making sure that the data doesn’t mix, otherwise, you won’t know which specific recommendation widget helps. Incanto, a lingerie and swimwear store, earned a 5.5% increase in revenue after launching product recommendations.
Pic 3. The results of the test were carried out with a control group: with a 95% probability, the version with recommendations was ±5.5% more effective.
Experiments in email campaigns are similar. Only a portion of recipients receives emails with recommendations. A/B tests are available on nearly every email marketing platform.
Pic 4. Example of an Incanto campaign with similar product recommendations
Offline Channels
An offline A/B test with a control group works based on the same principle. This can be done when testing product recommendations within a call center. The operator’s software sends a request to the CDP, where the audience is already divided into two groups. In half of the cases, it displays product recommendations on the operator’s screen. In other cases, no recommendations are provided. Then, the behavior of both groups is compared to determine where customers made more purchases. The same approach can be applied to POS software.
Key Takeaways
- Create a campaign list. Understand in which situations online recommendations may be useful. Make a “wish list” of everything you think may work, starting from simple solutions, such as popular products on the homepage, to pop-ups with personal recommendations when a customer is about to leave the site. Arrange the resulting set of hypotheses according to the reach. The more people that see the recommendations, the faster you will get a statistically significant result in the tests. To create a list of campaigns, use the Miro Mind map template above.
- Define metrics that help you understand which products you want to recommend and determine the criteria for success. Look at the revenue and page depth for online or revenue and an average order value for offline.
- Show this “wish list” to your product recommendation service. Developers should tell you how long the implementation will take, and the service representatives will tell you how to quickly set up campaigns. In services like Bloomreach, Klaviyo, or Mindbox, the most popular campaigns are provided out of the box. Feedback from colleagues will also allow you to adjust the launch plan — certain stages can be implemented faster.
- Ensure manageability and customizability. Check that you’ll be able to coordinate your recommendations logic across different touchpoints and customize it to better suit specific customer segments (brand lovers, heavy purchases, etc.).
- Upload the history of the customer’s interaction with the brand and the product range to the product recommendation service. Data from online and offline channels as well as mobile apps will allow you to generate better suggestions and guarantee uniform marketing in all touchpoints.
- Set up a testing tool. You can use Google Optimize for online channels, and a control group for offline channels. The distribution of the main and control groups can be 50/50, while efficiency can be assessed by revenue.
- Monitor the progress of the experiment and adjust the product recommendations. It may take up to 2-3 months before your first success, and some widgets may decrease revenue. However, once everything’s up and running, you might see an uptick in revenue, just as Incanto did.
