Personalization is a means of meeting the customer’s needs more effectively and efficiently, making interactions faster and easier and consequently, increasing customer satisfaction and the likelihood of repeat visits. But what does this mean in practice?
At Nosto, we approach personalization as the practice of leveraging Artificial Intelligence to expose products relevant to each particular shopper, and in so doing, augmenting the shopping experience. At the core sits our recommendation algorithms and effectively chaining them through something we call “Fallbacks” is an essential tool in any retailer’s toolbox.
The Early Days of Fallback Product Recommendations
Amazon introduced a layer of personalization to their ecommerce store as far back as 20 years ago by rolling out a Browsing History element but also by calculating scores between books based on custom reviews. This use case seems rather trivial today and seeing a row of products declaring “Other customers also liked” is hardly a rare sight when traversing the internet of the 21st century.
But what actually happens between these elements and when leveraging a personalization solution such as Nosto, how can you control these to better tailor towards individual business goals and needs?

Initiating Interest-based Product Exposure After an Initial Visit
The front page of Flight Club (above and below), an exclusive sneaker marketplace and industry leader since 2005, greets new customers with their top sellers which, at the time of this writing, is 4 quite gorgeous pairs of Adidas shoes. The underlying idea with exposing best sellers on the front page is to try to get the shopper to commit to one more click and delve deeper into the site for a higher chance at converting. Statistically, if most people buy these items, it should work on a large portion of the new customers as well – and this is true. However, this element is not actionable for a repeat, loyal customer shopping at the store, and something a bit more specialized should be exposed if the retailer knows the shopper’s individual preferences.
Let’s assume these are the kind of sneakers I am interested in as a shopper:

If I view a few of these and return to the homepage, what sort of treatment would I expect as a returning user who has exhibited a clear interest towards a certain item type and brand?

This is now the experience the shopper is exposed to when coming back to the site. The hypothesis here is that if I show a clear interest in Nike, then trying to sell me the best selling Adidas sneakers is probably not the way to go.
The strategy works well across most retailers that carry different competing brands, especially in the fashion and sports verticals. This is mainly due to high loyalty shoppers have for certain global brands that have invested heavily into marketing and branding to the point that wearing their products is considered more of a lifestyle choice than just a necessary purchase.
Leveraging Fallback Product Recommendations via Nosto
By leveraging Fallbacks and chaining different recommendation types to each other. Allow me to elaborate:

What is happening in this example is that we are exposing the Primary Recommendation type: Personalized Recommendations. However, this recommendation type expects the shopper to have viewed a few items so a shopper who has not seen any would not end up seeing this element. This is why we leverage the type Best Sellers as our fallback, leading to a separated experience for new customers, and for returning engaged users, leading to an increased shopping experience.
Keep in mind that retailers don’t need to restrict themselves to just one fallback. One of the most popular fallback stacks we have seen looks something like this:
- Order related recommendations (Requires 1+ orders)
- Personalized recommendations (Requires 1+ viewed products)
- Browsing history related (Requires at least 1 viewed product)
- Best Sellers (No requirement)
Tailoring product recommendations is a great way to optimize for different goals, while also keeping messaging consistent with the shopper’s current position in the shopping lifecycle.
Driving Increased Profitability Through Fallbacks
One of the more coveted functionalities that retailers can achieve by using Fallbacks is to up-sell with strict price or margin filters, allowing for fallback mechanisms to kick in if there are no relevant results to expose.

Route One is a skate / streetwear store founded back in 1989. There are multiple price points across many brands, which means shoppers stick to a certain price range and usually continue converting on similarly priced items that they have a affinity towards.
You can see this in the “You may also like” – cross-seller element leveraged to upsell from the viewed hoodie. The price range can roughly be bucketed into a “similar or higher” price. This enables the shoppers to easily discover similarly priced items.

So, How Can Fallbacks be Leveraged Through Nosto?
Easily – by chaining fallbacks and leveraging the relational Price filter. In this example, we are primarily upselling by keeping to a minimum 100% relational price and then leveraging a similar fallback without a price filter to allow for situations in which the shopper is browsing the most expensive items in the entire store.
This does not have to be tied purely to price but can also be achieved by leveraging Margin filters to drive profitability instead of pure revenue metrics.