Growing an e-commerce business with recommendations
“A lot of times, people don't know what they want until you show it to them.”
This is a quote from the late Steve Jobs, co-founder of Apple. Although this quote relates to Steve’s attitude towards market research (he famously often preferred to launch products based on intuition and feeling), it is definitely also applicable to two of the current major e-commerce trends: personalisation and recommendations.
Shopping online is about more than simply buying a product. Like browsing in a brick and mortar store, shopping online is an emotional and aspirational experience. One of the key challenges for online retailers today is building deep and meaningful relationships with customers by providing these experiences.
One way of achieving this is by serving personalised recommendations to give customers the sense that they are being understood, valued and properly served.
What is a recommendation engine?
An effective recommendation engine enables you to target the right products (and content) to the right people across different pages on your website as well as across a variety of marketing channels.
Recommendation engines are everywhere. For an example, take a look at Amazon, one of the most valuable companies on the planet. Arguably, much of Amazon’s growth comes from its success with personalisation and integrating recommendations into every part of its marketing. It even has a dedicated page for “Your Recommendations”.
Put simply, a personalised online experience sells more products and engages website visitors (and existing customers). A Forrester study reported that over 15% of website visitors admitted to buying recommended products and over 70% preferred personalised shopping experiences.
So, how exactly can a recommendation engine help you to grow an e-commerce business?
Deliver live recommendations – it’s possible to analyse a website visitor’s current site usage (and pair it with any previous history) to deliver live recommendations as they browse the website further.
This helps to engage shoppers – website visitors will be able to explore product offerings more deeply and efficiently without having to use website navigation or search.
On top of this, engage customers across multiple channels – for example, it’s possible to serve recommendations via retargeting and email marketing.
Improve conversion rate – use personalisation to show that customers are valued as an individual as well as making your site easier to navigate.
Increase the number of items purchased – when a customer is recommended a product that meets their interests, they are more likely to add it to their order.
Inform strategy and decision making – by analysing reports from your recommendation engine (or getting support from a partner like Station10), you’ll be able to uncover further customer insights and further refine the system.
One of the top aims of every e-commerce marketer must surely be to achieve Amazon-like success with personalised recommendations. Almost every online store has basic recommendations i.e. “I bought/viewed this product so I may be interested in another product” but achieving true personalised recommendations (serving recommendations based on a number of individual attributes as well as any other information held on a customer) is a different ball game.
The first challenge to overcome is the availability of data. Serving recommendations to visitors requires data from across an organisation: website data, transactional data and CRM data to name a few. A key step in implementing a recommendation engine is unifying this data.
Next, if all of the data you need is available, do you have the technology to handle this data? There is a range of off-the-shelf solutions but it’s not always a case of one size fits all. Not every business’s product, visitor and behavioural data is the same. In Amazon’s case, scale was the issue. The online retailer built its own engine because existing technology couldn’t handle their hundreds of millions of customers.
The final hurdle to overcome when implementing a recommendation engine is the ethical intelligence of the system. A report from Malmo University uses the phrase “recommendation ethics” and explains that whenever the functions of a recommendation engine are exploited for reasons beyond serving user needs, an ethical problem arises. Taking the betting industry as a specific example, serving recommendations that may influence customers to bet more brings up a number of questions on ethics. Can the customer afford to bet more? Do they have a gambling problem? Other potential recommendation ethics issues include privacy, censorship and discrimination issues.
In an era where online shopping is growing massively, personalised recommendation engines are becoming increasingly popular and arguably necessary. It is clear that recommendation systems are providing huge value for not only retailers but also media companies who aim to keep readers on their websites for as long as possible. Available technology is already sophisticated but there are still potential problems to overcome including organisational issues and ethical dilemmas.