How has buying and selling evolved online?

Product Recommendations are quietly shaping e-commerce behaviours

For as long as humans have existed, we have been bartering, selling and consuming goods. Back in 2013, archaeologists discovered a 2,700 year old shopping mall from Ancient Greece, reinforcing the concept that as much as things change, they still remain the same.

In more recent times, ‘Mom and Pop’ shops led the way through the 1700s and 1800s, with department stores powered by the invention of the cash register ushering in a new consumer era. This led to modern shopping malls emerging slowly in the 1950s in some regions across the globe.

Throughout this time, the one consistent element has been the shop keeper or shop assistant, there to guide the customer and suggest some items they might want to purchase during their visit. Making recommendations.

New era

But the internet has now spawned a new era of shopping. With businesses moving online at record pace, a 56% increase in global e-commerce shows that competition is sudden and fierce. Finding ways to add the personal touch and service from the brick and mortar experience to an online world has always been a challenge, but also brings with it the advantages of data — which can be viewed in many ways as the new ‘world currency’.

In the local shop, a good assistant might know you and your interests to help you make decisions, while in e-commerce machine learning serves that purpose. Suggestions can be made based on the consumer’s activity on site, along with additional user behaviour with similar interests and many other factors. This takes the shop assistant role to a new level, with a truly personalised approach beyond what could be captured in conversation.

Product Recommendations

Amazon led the innovation in product recommendations, taking this concept and building an empire with recommendations as a core feature for growth, generating as much as 35% of its revenue from this machine learning technology.

Online shoppers are now well accustomed to seeing website sections with titles such as ‘You May Also Like’ serving as the modern-day shop assistant, recommending products when browsing a site. The feature is now one of the 7 must-haves for product pages.

But the recommender system is but one part of the puzzle and comes with great responsibility.

Machine learning and AI play their part but are only as good as the data you feed it. Rich and detailed product titles and descriptions are critical therefore to the technology, so the algorithms can compare products in detail.

Although not universally true, a simple way to think about it is that anything on your product pages that enhances SEO is generally good for product recommender systems as well.

People don’t buy products anyway, they buy ‘what they do for them’ or how they make them feel, or look, in the case of clothing.

How you tell the story of your product through the title and the description sets the tone for the entire buying process. How you provide product information says a lot about your company; everything you write tells someone how to feel and speaks to the emotive side of the buying experience for consumers.

Combining well-written copy and the right type of technology is not always an easy task.

Smaller retailers don’t have the resources to build their own tech solutions and must rely on third-party applications for proper machine learning engines that drive quality product recommendations.

Not all technology is created equal

Some e-commerce website pre-built themes or templates come with a basic product recommendations engine built in, but the technology is not very powerful because they are ‘one size fits all’. Many recommendation or personalisation engines claim to enable you with business rules, but are limited in scope and effectiveness, or require constant monitoring and tweaking that only large enterprise companies can afford to staff on a regular basis.

In reality, most of them use basic algorithms, and while some algorithms work well for specific situations, not all algorithms work for every scenario.

The most cutting edge product recommenders use artificial intelligence to select which algorithm is suitable for every user click, in real-time. This removes the need for configuration and business rules for many scenarios and makes it more feasible for small businesses to derive value by providing recommendations to their customers without constantly configuring the technology.

The challenge for the business owner is that the tech all looks the same on the outside, and it is challenging to decipher what technology is underneath the widgets. Finding a trusted partner is key.

Automated machine Learning evolves and improves personalised consumer experiences on websites over time.

A more holistic approach

Overall, the end result of these ‘digital shop assistants’ doing all the work, is a better user experience for the consumer and increased conversion rates in the short term for retailers, with improved customer lifetime value in the long run when managed properly and ethically.

But since customer lifetime value is the ideal metric for all e-commerce businesses to optimise but notoriously difficult to measure, a true long term strategy needs to be developed beyond just selling more products (which inherently encourages consumption optimised for profit margins alone, not focused enough on the consumer or environment).

Product recommendations ought to be paired with a strategy around reviews, loyalty campaigns and more detailed information for consumers. This will deliver more value and increasingly, demonstrate the impact your company intends to have on society.

Industry-specific

Small businesses are disadvantaged when using large website platforms and combining that with various apps or plugins from multiple third parties for conversions. They cannot build a truly circular data strategy that informs each part of the business in the way that massive companies like Amazon and Ebay can.

This is primarily because e-commerce tools have been built mostly in a ‘vertical agnostic’ manner, to enable any industry to use the various conversion apps and solutions from fashion to electronics to travel.

But e-commerce is maturing and the trend over the next few years will be more bespoke solutions for specific industries, enabling more efficient and tailored offerings for consumers, benefitting brands, retailers and suppliers.

By developing industry-specific solutions, more cohesive and affordable data strategies can be developed for smaller businesses, which will enable more efficiency in supply chains informed by consumer behaviour. This will allow small and medium enterprises to access the same level of data intelligence that, until recently, has been only available to the largest companies.

As supply chains are increasingly moving from globalisation to localisation, retailers and suppliers will need to be willing to share more information with each other to compete with the large monopolies and marketplaces.

Increasing transparency

The next fundamental shift and evolution in e-commerce involves increasing transparency and Environmental, Social Governance (ESG) data around products and supply chains, incorporating information highlighting the quality, longevity and reusability of products.

This is where small businesses will enjoy a huge advantage, by innovating faster and being more transparent with consumers.

Enterprises with shareholders will take longer to make the shift to full transparency, providing a massive window of opportunity for the ‘buy local’ movement to reclaim market share from the monopolies.

Society expanded the supply chain globally to support a growing population, but we are now more aware of the resultant impacts of the linear economy. We are seeing a ‘full circle’ movement evolving and we are returning to our local roots (even in e-commerce) just like the Greeks so long ago, with local, sustainable and innovative circular models on the rise.

Not just Generation Z

The forthcoming generation of consumers have proven to be very aware of the impact of their purchasing power, and earning their trust will require more information. Many brands are already adding certifications and badges to product pages and even product recommendations.

But age is not the indicator in these consumer demands, with the pandemic accelerating the trends universally as a “wake up call” to protect the environment for all ages. In an Accenture survey, 60% of people globally reported making more environmentally friendly, sustainable, or ethical purchases since the start of the pandemic. While 73% of UK consumers want to be more sustainable, 42% of over 55-year-olds said the reason they were finding it easier to go green was that they now ‘felt more knowledgeable’. It’s not just the next generation.

Technology & Humanity

Overall, technology is only as effective as the human strategy behind it. Keeping the consumer in mind at all times will help to build the trust of all those voting for their version of the future of e-commerce with their pocketbook.

The retailers, brands and manufacturers that work together to provide more information about how their products are produced and include that data in more of their processes like product recommendations, will truly deliver a user experience focused on customer lifetime value and expectation.

Ultimately every tool should still be designed to boost the most impactful recommendation of all. Word of mouth.