Want to save this article for later?
The concept of recommendations has been around since Amazon first added “inspired by your shopping trends” or “top picks for you” to its website.
When customers are on the lookout for an item, they are hoping to find the best possible one out there. Their search is comprised of hundreds, if not thousands, of digital touchpoints. Through these interactions, customers expect brands to learn about them. In fact, they’re willing to spend more money if they receive a thoughtful, targeted experience.
While generating recommendations has become fairly standard, there continues to be an increasing level of frustration when they’re not personalized.
One of the key reasons is that growth teams have approached recommendations as a short-term product-centric strategy instead of a long-term brand strategy. So recommendations are seen as a means to drive more conversions on the website or app instead of using them as a sustainable way to increase customer lifetime value and loyalty and eventually optimize the end-to-end customer experience.
So how do we bridge this gap from a short-term campaign strategy to a long-term growth strategy? How do we ensure that we account for rapidly evolving lifecycle stages?
A report on personalization from Segment assessed how it impacts shoppers1. Findings show that 40% of consumers have purchased something more expensive than they had originally set out for. Moreover, nearly half (49%) of the shoppers made an impulse purchase after they received a personalized recommendation and a majority (85%) of them were very happy with their decision.
It would be impossible to keep recommendations relevant to each customer in real-time without using a scalable data science model that works well for your business. Data science helps automate customer recommendations to not only make them relevant to each user but also to consider the dynamic changes that occur along the way.
For example, if you recommend vacuum cleaners to someone who just bought one, that message is bound to be ignored.
Similarly, if you recommend chocolates to a former chocoholic who’s now buying more fresh vegetables from your grocery app, you aren’t keeping up with her updated demands.
A recent report2 on recommendations for ecommerce shows that the average number of items purchased increases by 50% when customers actually engaged with the recommended items – through clicks, impressions or purchases. It further finds that the average order value of a recommendation that catches a customer’s attention increases by nearly 33%.
This means that once you have created a set of data science enabled recommendations, you want to find a way to engage with those recommendations at the best possible time on any number of channels.
But now with Product Recommendations, we’ve taken the next step to assist you with your engagement strategy. With this latest feature, you can automate the ‘what’ of customer interactions with 1:1 personalized recommendations that dynamically adapt to customer purchase behavior, buying patterns, and usage trends.
Using an AI-powered system that allows complete control over merchandising, growth teams can create intuitive targeting rules for millions of catalog items. This means our recommendation engine can generate recommendations unique to each user.
Once the recommendation engine generates specific content for each user, growth teams can send these out via In-App, Push, Webhooks, SMS, and App Inbox. They can build rich marketing communications using images, videos, deep or external links, with custom fields in their catalog definition to personalize the message. Omnichannel campaigns can also be triggered based on specific user behavior such as “add to cart” or “searched” to bring further context.
You can create multiple types of recommendations based on different filter criteria.
Ecommerce: Increase upsell opportunities using multiple custom recommendations in one message
For example, trigger an In-App message in real time when a user adds an item to cart:
Media/OTT: Increase viewing time per user with recommendations optimized to user preferences
For example, once a user completes watching a teaser video of a fantasy TV series:
Travel: Optimize customer experience by avoiding repetitions and ‘not available’ scenarios
For example, for a user browsing for a hotel on your website:
The goal of our recommendation algorithm is to provide diversity and novelty, so you can expect product recommendations that have low count to figure in the top recommendations while maintaining the relevancy of engagement. This industry-leading capability will allow you to create unique recommendations for each user from your product catalog. This will then be available for you to run your engagement strategies through channels like In-App, Push, and more. Each user, based on their generated recommendation, will receive a relevant and highly personalized message.
CleverTap’s Product Recommendation is currently in Beta. You can write to us at [email protected] to get a quick demo or request early access to try it out.