Effective marketing isn’t a case of set and forget. It requires ongoing testing and optimization to make sure you’re delivering content that resonates with your audience, converts shoppers into customers and continues to do so amidst fluctuating digital trends and consumer tastes.
But when it comes to optimizing your website and email marketing, there are so many different elements to optimize that it can leave you scratching your head and reaching for a strong coffee before you’ve even begun.
In this blog post, we take a look at how to optimize your Product Recommendations, sharing basic optimizations to get you started, next level optimizations to step it up a gear, and advanced optimizations to make your Fresh Relevance personalization best in class.
Let’s dive in.
Basic optimizations
Look and feel
First things first, give your product recommendations a chance to shine on your website with an optimized look and feel. Test your overall design and layout and try out different copy, content and image sizes. Your call to actions (CTAs) are also important, so play around with different sizes, colors and messaging, for example testing the difference between a solid color CTA and a white CTA with a colored outline.
Number of products
This optimization is all about balancing product exposure with showing the most relevant products. A carousel layout can work well here but it’s worth testing how effective this is with your audience. Try testing the difference between displaying 3 products vs 9 products vs a carousel layout, for example.
Data source and fallback optimization
Fresh Relevance offers several types of data sources and depending on the touchpoint, there may be several options available for your tactic, such as business as usual (BAU) emails.
These options include:
- AI visitor personalization, where you can intelligently predict what your customer wants to purchase based on what others have done, e.g. ‘People like you buy’
- Crowd sourced, which displays the products that are hot on your site right now, e.g. ‘most purchased’
- Visitor’s historical behavior, which displays products based on what the individual shopper has previously browsed, carted or purchased
This optimization doesn’t just apply to the primary data source but also to the fallback options, particularly if the primary source is highly personalized.
Try testing ‘people like you buy’ vs ‘previously browsed’ vs ‘most purchased over last 7 days’, for example, to see which types of recommendations resonate best with your customers.
Learn more about different types of Product Recommendations.
Ratings and reviews
With Fresh Relevance, you can take your recommendations to the next level by including the product’s ratings and reviews within the recommendations to draw on the wisdom of the crowd and increase confidence in products.
You can also filter by ratings to only display products with specific ratings, for example only showing products that are 4* or above.
Test the difference in performance between recommendations without ratings vs with ratings vs with 4* ratings or above.
Next level optimizations
Advanced data source optimization
Once you’ve established the data sources that work best for you and your business, it’s time to optimize them further as each data source has different options, including:
- AI visitor personalization – ‘people like you buy’ vs ‘people like you browse’
- Crowd sourced – frequently purchased, carted or browsed
- Person history – frequently purchased, carted or browsed.
It’s worth testing which of these versions resonates with your customers as it might not be what you’d expect. One Fresh Relevance client tested frequently purchased vs frequently carted and found the latter to be most effective, much to everyone’s surprise.
Filtering
Data sources are just the beginning. You can also enhance product recommendations with filters. Here are some key filters we recommend testing:
- Best tag value – limiting recommendations to core tags/categories browsed. Test the number of categories to restrict, for example top 1, 3 or 5 tags.
- Price affinity filter – AI recommendations that predict the price level that will appeal to each website visitor. This can be particularly helpful in engaging new shoppers to your website where you don’t have any behavioral data yet.
- Price higher or within range of history (browse, purchase and cart) – recommending products based on their price points in relation to a customer’s previous behavior. This can be an effective way to upsell and increase average order value (AOV).
- Excluding recently recommended products (this increases product exposure when repeatedly using recommendations)
- Set price boundaries – recommending products that are in a certain price range, for example products under £20. Consider restricting this to your most purchased price range for more compelling recommendations, or use at Christmas to recommend stocking fillers.
Top tip: Pay close attention to all metrics for these tests as you may find positive improvements aside from revenue, such as AOV.
Crowd sourced data window
With Fresh Relevance, you can look over different options for a crowd sourced data window, for example best sellers of the day, week or month. The best option for your business can really depend on your web traffic levels and whether you want to label all products or highlight a select few so they stand out. For example, if you experience high traffic and you only want to highlight labels on some products, a shorter timeframe would typically work best. Conversely, if you want to label all products and you experience low traffic, a longer time frame could be the best option.
Try testing 24 hours vs 7 days vs 30 days to get you started.
Social proof messaging
Social proof is an effective way to boost customer confidence in your products and increase urgency. But it’s worth testing which messaging is most effective for your business. Are customers driven by how many people have purchased a particular item or are they also motivated by how many people have browsed the item? It could be the case that browse numbers are best for driving urgency, as the customer wants to make their purchase before a product goes out of stock, and purchase numbers are best for boosting confidence.
Try testing product recommendations without social proof vs with only purchase messaging vs with purchase and browsing messaging.
Advanced optimizations
Advanced templating
With Fresh Relevance, you may be able to use our advanced templates such as direct add to cart and bulk ordering functionality. Try testing the difference between having a link to the product page on your recommendations vs direct add to cart functionality vs direct add to cart with bulk ordering functionality.
Social proof
There are a couple of ways you can take social proof within product recommendations to the next level. First, explore messaging optimization. This is all about finding how best to communicate with your customers. Do they want to see numbers of people that are viewing products or simply a ‘best seller’ tag, for example?
Then look at optimizing your social proof threshold. Popularity messaging is most effective when it stands out; if all of your products appear to be trending and popular the message gets diluted.
With Fresh Relevance, you can change thresholds for each to find the sweet spot, for example testing the difference between only displaying popularity messaging when at least 5 have been purchased vs 10 vs 50.
Affinity optimization
What drives affinity for your customers? Categories tend to be the most commonly used, but that might not be the categorization that produces the most affinity, but it’s worth testing the level of categorization that produces the most affinity, for example top or sub level, or SKU variants such as color, size, style and material. And how many affinities are most effective.
Try testing top level category vs sub level category vs color affinity, for example.
Final thoughts
Optimization is a crucial part of effective product recommendations but with so many elements to test, you need a plan of action before jumping in. Once you have your product recommendations set up, start small with our basic optimizations, such as the number of products to display. Once you’ve gathered results and made changes, move onto our next level optimizations, such as enhancing recommendations with filters, and finally try out the advanced optimizations, such as affinity optimization.
Our Testing and Optimization tool empowers you to optimize your product recommendations in the ways discussed above and our customer support team is on hand to make your optimizations go further. Book a demo to learn more about testing and optimizing your marketing with Fresh Relevance.
