The fastest way to win new customers.
LEAP FORWARD WITH AMAZON PERSONALIZE.
Amazon has pioneered personalization for over 20-years. Constantly improving the performance of algorithms on Amazon.com.
We’ve partnered to accelerate eCommerce. Every merchant – small to big – needs access to top-tier machine learning technology.
“Obviyo allows brands of any size to benefit from personalization based on the same technology that is powering product recommendations on Amazon.com.”
Head of Business Development for AI, Amazon – AWS
STOP MAKING GROWTH COMPROMISES.
10% of Visitors
90% of Visitors
THE BIG DISCOVERY
Live buying-signals are better predictors of visitors’ needs and preferences than customers’ historical data alone.
“This! This is the app I had been waiting for! Obviyo gives me access to Amazon’s recommendation engine without having to be part of the Amazon ecosystem. It recommends MY products to MY customers, and WE make the revenues. No commission to the 1,000 pound gorilla, Amazon. Over the last 30 days we’ve made 16.7% of our revenues from having Obviyo running on our site. If you’re thinking of adding this app, DO IT. It’s worth it. Also, the development team is highly responsive, and the business managers are friendly, easily accessible, and helpful. This app is top notch.”
Founder and Publisher, Genius Books (Shopify)
Maximizing Amazon Personalize with layers of eCommerce intelligence.
Every interaction by your web visitor could be a buying signal if your personalization solution can detect and act on it.
Using historical data to train personalization engine.
Low-relevance recommendations using visitor’s history.
Using diverse real-time audience data to fuel personalization engine.
Highly-relevant recommendations driven by visitor’s live buying session.
Did you know?
Visitors’ needs are always changing. What someone purchased in the past may not predict their future preferences.
The Impact of Hyper-Personalization During Peak Holiday-Shopping Season.
Hypothesis: If hyper-personalized recommendations are adapting to visitors’ actions and context in real-time, they should produce improved results during the busy holiday season when everything abruptly changes.
HOLIDAY SEASON RESULTS
Oct 1, 2020 – Nov 15, 2020
RPV (non-personalized): $2.41
RPV (hyper-personalized): $6.85
Nov 15, 2020 – Dec 25, 2020
RPV (non-personalized): $4.10
RPV (hyper-personalized): $14.70
New Revenue Attribution
Before Holidays: +5%
During Holiday: +10%
Full Case Study 🔗
Exclusive to the Amazon AWS blog.
Before holiday season, your visitors are usually buying something for their own needs. During busy holiday periods they are mostly buying presents for other people.
If your conventional personalization system is trained on pre-holiday data, recommendations will be less relevant. Your visitors are in search of something different that they want to gift to others.
With hyper-personalized product recommendations your site will always automatically sync with current visitor needs and preferences.
Product descriptions and meta data hide powerful indicators of customer preferences to increase relevancy of recommendations.
Did you know?
Conventional personalization solutions ignore under-sold products because of sparse data. Natural language keyword indexing enriches data and expands the range of products to recommend to visitors.
It’s essential to align your business objectives, like top line growth, margins, and inventory management with your own customer needs.
Did you know?
What’s good for your visitors may not always align with your profitability goals or inventory status. Using business goals as a key personalization criteria is essential for long term success of your business.
Unlocking revenue potential from assumed non-buyers needs broad recommendation strategies. Designed for different buyer personas, actions, and context.
Showing visitors a small number of product recommendations.
Very few visitors engage with recommended products influencing a low percentage of overall sales.
Eclectic recommendations at every step of buying journey.
Significant percentage of visitors engage with recommended products, influencing large percentage of overall sales.
Did you know?
When visitors engage with hyper-personalized recommendations, revenue per visit (RPV) goes up exponentially compared to visitors who don’t engage. A wide-net strategy will increase the percentage of visitors who are engaging with recommendations and grow revenue.
What Running Dozens of Hyper-Personalized Recommendations Taught Us About Growing New Revenue.
% of engaged visitors: 1.2%
% of revenue influenced: 2.4%
% of engaged visitors: 9.6% (▲ 700%)
% of revenue influenced: 37.9% (▲ 1,479%)
Timeframe: 4 months
Most stores have a few product recommendations on a handful of pages, like product or cart pages, while the majority of their site is just static content.
Amazon.com is blowing ‘conventional wisdom’ out the water and treating almost the entire site as a dynamic, product recommendation driven experience.
This case study proves the value of Amazon’s approach. As the merchant increased the number of recommendations the percentage of visitors engaging with product recommendations increased in a linear way, while new revenue influenced by recommendations increased exponentially.
Placing a wide variety of product recommendations – aimed at different buying personas – is known to unlock new revenue from 90% of non-buyers. Visitors who never meaningfully engage with your brand today.
Not type of device but type of shopper. Mobile visitors have different product needs and they shop differently.
Did you know?
Mobile eCommerce is not about smaller devices but about different kinds of shopper. Whose needs are different, with shorter buying sessions and more impulsive decision making.
Buying logic changes at every step of the buying journey. Personalized recommendations must adapt to those changes.
Did you know?
What live visitors have done at one step is a pre-text to interests in a next. Creating a seamless personalized experience across in-session buying, post-purchase marketing, and return visits is essential for your success.
WHAT MAKES HYPER-PERSONALIZATION DIFFERENT?
‘Amazon Grade’ Machine Learning
Hyper-personalization is only made possible with a new generation of machine learning algorithms. Their development requires expert knowledge, deep pockets, and extensive empirical evaluation.
We selected Amazon Personalize as the top-tier technology. Perfected through 20-years of innovation and experimentation to net billions in sales on Amazon.com.
PRIMARY RESEARCH PAPER
Personalizing Session-based Recommendations with Hierarchical Recurrent Neural Networks 🔗
Massimo Quadrana, Alexandros Karatzoglou, Balázs Hidasi, Paolo Cremonesi
For hyper-personalization to work it’s essential to have a hyper-performing loop for capturing and submitting data to algorithms, plus getting and presenting recommendations – all of which must happen in a blink of an eye.
We provide highly integrated front and back end systems with CDN based computing end points.
Hyper-personalization is driven by a mix of highly-accessible activity, visitor, and product data that is normally stored in different systems, like web analytics, customer data lake, or product data store.
We integrate and structure highly-fragmented eCommerce data and make it accessible with high-performance systems.
“Might be too advanced for some brands because it is powered by Amazon’s own product recommendation technology. Once you go live, the reporting dashboard makes is easy to measure the overall impact of your campaigns. You’ll quickly see why they evangelize adding as many product recommendations as you can to your site!”
Founder, Synergee Fitness