Product recommendations are an effective way to enhance average cart size and maximize revenue per customer. Big players like Amazon attribute almost as much as 35% of their conversions through product recommendations. However, for relatively new players it may not always be easy to recommend products effectively.
Although automated recommendation engines have made it possible to show personalized recommendations to visitors, optimizing its functionality still remains a challenge.
Merchandizers make some common mistakes while recommending products and simple corrective measures such as analyzing performance, A/B testing etc can bring about a significant improvement to your bottom line.
In this post, I’ve highlighted a few of these mistakes and mentioned ways to optimize your recommendations engine.
Mistake #1: Not showing the right recommendations on the right pagesSo you’ve started using recommendation widgets to show more products to your visitors but people are not clicking on your recommended products. What’s going wrong?
Simply put, you may not be showing right products at the right time to your visitors. It’s important to understand that the shopper’s psychology is directly related to the recommendations you are providing.For instance, if you’re showing complementary products like ties, belts or pants on a shirts’ search result page, it becomes irrelevant for the visitor whose original intent was to purchase a shirt. Instead, showing category bestsellers on the search results page will help visitors narrow down to the right products. I recently wrote a post on best practices for placing recommendation widgets on your site, you can take a look at it here.
Dogfunk, an online retailer that sells snowboard gears and apparel, shows different products like sunglasses, jackets etc on the watches search results page. A better recommendation would be to show top-selling watches on this page.
Mistake #2: Offering too many recommendations on a single pageMost retailers provide a lot of choices to visitors, thinking it would lead to more sales. They’re wrong. Too many choices confuse your visitors and could lead to a phenomenon known as the paradox of choice. Let me give you an example. Lets say, a visitor comes on your site to buy a checked shirt and lands on the product page. This page contains ‘More like these’, ‘People who bought this also bought’ and a ‘topsellers’ widgets which are way too many recommendations on a single page.
You must be clear with the strategy behind your recommendations and show only the most relevant products per page in order to avoid confusion.
“A recent survey proved that visitors who were given more choices had a conversion rate of 3%, whereas visitors who were recommended lesser products had a conversion rate of 30%.”‘Less choice means less confusion’. It’s as simple as that! Snapdeal is India’s largest online marketplace which shows too many recommendations on the product page.
Mistake #3: Missing the metrics
A lot of ecommerce merchandizers concentrate on achieving business objectives through product recommendations but pay less attention to performance. This prevents them from understanding what works and what doesn’t work for their site.Reports and analytics give you an insight on how your recommendation widgets are doing and prompt you to optimize your strategy accordingly. For example, our Merchandizing Dashboard gives you in-depth insights on the performance of your recommendation widgets. The reporting section contains multi-dimensional reports which allow you to slice data based on category, brands, locations etc. It involves important metrics such as impressions, clickthroughs, conversions etc, making it extremely simple for you to understand how recommendations are impacting your conversions.
Mistake #4: Lack of A/B testingSome merchandizers often ignore A/B testing as a practice while testing recommendations. As a merchandizer it’s important to optimize and test smaller details like number of recommendations per page, look and feel of widgets, their positioning etc. A/B testing these parameters give you an idea of how your visitors respond to such changes. A simple example is of Carelogger website, that managed to increase its conversion rate by 34% simply by changing the colour of the sign-up button from green to red!
It a similar way, you can A/B test different recommendation parameters and optimize product recommendations on your site.
As Optimizely puts it, “A/B testing takes the guesswork out of website optimization and enables data-backed decisions that shift business conversations from “we think” to “we know.”
Mistake #5: Recommendations engine does not align with Merchandizing
Many recommendation engines use widgets to show personalized recommendations, however merchandizing helps optimize site performance by letting you apply rules to achieve various business objectives.
For instance, if your business objective is to clear high-stock products, you can promote these products while recommending similar products to your visitors. Aligning your recommendations with merchandizing helps in optimizing overall business performance.
Mistake #6: Limiting recommendations to product pagesYou should treat your visitor’s shopping experience as a stepwise process and at each step you can capitalize on their shopping impulse by showing personalized and relevant recommendations. For instance, when a visitor lands on your site, show him personalized products with widgets like ‘Recommended for you’ or ‘You recently viewed’. This will personalize his on site experience and help him go to the desired products instantly. Similarly, you can show specific recommendation widgets on your product recommendations page, search results page, cart recommendations page. ———————————————————————————————————————— Keeping in mind these couple of elements when using product recommendations on your site will help you optimize your recommendations and avoid the common pitfalls most merchandizers get stuck in.
I’d love to hear if you’ve come across any other type of product recommendation flaws, do share it with me in the comments below!