20th November 2018 -

AI-Driven Recommenders – The Power of Real-Time Ecommerce Personalization

Online recommender engines are becoming increasing popular across many sectors. What was once introduced by Amazon as a concept of ‘you may also like’ has now spread not only across retail but into TV and media, professional networks, music and social media networks. Online recommendations are the perfect tool for e-commerce professionals to increase their average order value and online conversion rates.

Early Adopters

There are not many retailers now that don’t offer some form of additional product recommendation on their website. Any retailer who hasn’t yet adopted this is now certainly in the minority. Recommendations usually sit on the retailer’s product pages and offers an introduction to standardized items such as best-sellers or showing what other people have looked at who have viewed the same product.

These standardized recommender features are accessible via ecommerce platform marketplaces and do allow a degree of customization. They allow retailers to choose where the recommender sits on the page, how many products are shown and what the header of the recommender is.

Restrictions on Standard Recommendations

The issue with the way that the majority recommenders are used is that they are not dynamic but are based on fixed rules as to which products they show. A commonly seen recommender is ‘People who bought this also bought’. They do not adapt to be unique and relevant for each visitor nor pull on real-time data. There are many times when visitors will see the same products recommended throughout the website and when they are not shown in real-time quickly become obsolete and irrelevant.

What isn’t always apparent to the visitor is that these recommendations are quite static and based on the sales results of the website overall and not on their own individual profile and preferences.

Real-Time and AI-Driven

What retailers need to see is that there is so much more benefit in using AI-driven real time recommenders.  A good recommendation platform should be able to continuously learn and adapt itself flexibly to new user behavior.

What retailers need to see is that there is so much more benefit in using AI-driven real-time recommenders.  These new types of recommenders are so much more powerful in the art of conversion and increasing average order value as they are based on the true needs of the individual and not the performance of the website as a whole.

The recommenders now available through ecommerce personalization specialists like PureClarity are driven by powerful AI algorithms and machine learning which evolves the recommendations made with every click.

What Determines the Recommendation?

Based on the behavior of a visitor from the minute they land on a retailer’s website, they look at the behavior of not only the individual from current and previous visits but also the segment that the visitor belongs to and the crowd, meaning the trends and patterns of the website overall.

This three-tiered analysis forms a deep dive into the visitor’s profile and can accurately predict requirements and trends of what that visitor will be interested in, present relevant products and in turn increase the websites propensity to convert.

Effective and real-time AI collects data 24/7 from every visitor interaction onsite but also collects data from offsite behavior such as demographics, location, source of traffic, key words, campaigns and device. So, from the second a new visitor arrives on a website the AI has already built enough of a visitor profile to start providing a truly personalized experience from the second a visitor land on a site. The recommendation engines should deliver the best recommendation strategy for each individual, presenting the most relevant products, categories and brands.

Where to Place Recommenders

On average only 65% of traffic to a retailer’s site will arrive at the homepage. It is important to provide recommenders throughout the entire buying cycle. Personalized AI recommenders should be able to be placed anywhere on your website. More importantly recommenders should be included throughout the entire buying cycle from your homepage through to your search results, product page, basket, checkout and order confirmation page, through to being included within your email marketing campaigns.

What Sets PureClarity Apart?

Recommenders should be extremely targeted and relevant and as unique and as your customer.

The beauty of PureClarity’s AI Recommenders is that they get to work from day one regardless of the amount of data you have. You don’t have to lift a finger for the AI to get to work straight away.  The second great feature of PureClarity’s AI Recommenders is that it intelligently strategizes as to which recommenders to present on which page.

It evaluates what products are already being shown and uses the AI to present the most relevant recommendation based on the three-tiered data dive into behavior of not only the individual, but the customer segment and the crowd.  PureClarity’s AI ensures that there is no duplication of automated recommenders throughout a customer’s single journey.

PureClarity’s options for recommenders can range from standard suggestions such as ‘People who bought this also liked’ through to complex recommenders such as – ‘People who viewed this, added the following to cart, value under $40’. The options are endless.

The Race To Be Relevant

Recommenders are now the expected norm and provide online retailers the opportunity to provide an extremely personalized experience for the online customer. Only with real-time AI machine learning and intelligent algorithms will recommenders work effectively.

The data used to drive the recommenders needs to evaluate both onsite and offsite behavior of not only the individual, but each customer segment and the website as a whole. Then and only then will a recommendation be successful in the race to be relevant, convert into a sale and create online brand loyalty.

Leave a Reply

Your email address will not be published. Required fields are marked *