Personalisation and recommendations - in practise

Personalised content is said to improve your conversion rate dramatically. But what should you adapt - and how hard is it to implement? Ratings are a feature of most commercial websites - but which do customers trust?

Profiling or personalisation?

Customer journey mapping and profiling are a good place to start when aiming for an e-com experience that feels personal and relevant. Two or three distinct customer profiles can be enough when designing menus and filters. Landing pages for the most common Google search phrases with an choice of products and helpful information are perceived as relevant and personal, since they connect to what I’m looking for just now. It is also important to think about tonality and imaging in order to feel “right” for your typical customer profiles. But what we usually refer to as personalisation is a more specific adaptation to the preferences and buying patterns of the individual customer. And that requires data.

First-hand or second-hand data?

First-party data which we collect by tracking and encoding visitor behaviour - maybe even across channels and touch-points - is valuable when building a personalised experience. Data can be collected from your e-com platform, ERP, newsletter platform or CRM - as long as it can be connected to the same individual. Even more valuable is what is referred to as “zero-party data”, which is consciously shared information such as preferences, goals and problems, usually collected from a test or guide on your site. Data purchased from a third party and shared via cookies has become increasingly taboo, and even blocked by Apple and (soon) Google. Retargeting in the form of ads that follow me cross devices are usually seen as “creepy” and should be avoided. Personalisation is based on identification, ie that a user can be tracked between sessions and that certain basic data such as geographical location, frequency etc can be collected. There are even CDP’s (Customer Data Platforms) that log massive amounts of data such as on-site search phrases, visited pages, scroll data etc and use AI to add assumed age, gender and “propensity to purchase”.

A more personal experience

OK, so now we have the data. What can we do to give a great customer experience and increase our CR? We can suggest products that complement previous purchases. Fashion and beauty e-tailers can suggest how to complete a particular “look”. A bookseller can suggest the next part of a trilogy. With the help of AI we can arrange products on the start page to suit the individual visitor’s age, gender and favourite colours - but this is complicated and requires massive amounts of data to be successful. A good idea in order to increase your conversion rate can be to adapt the (internal) search experience by personalising the way results are presented.

Reviews FTW?

According to the latest “e-barometer” report from PostNord/HUI, reviews were seen as an important source of information by 69% of “frequent e-com customers”. Recommendations/ratings and user-generated content such as Instagram posts are more or less standard, especially in fashion/beauty. There have been some scandals involving fakes (Tripadvisor comes to mind), but for the most part reviews can be trusted. There is ample room for experimentation here - try balancing positive and negative reviews to increase credibility - or using catchy turns of phrase from users’ reviews in your product descriptions. A great idea is to feature posts about how sizes and colours match your customers’ expectations.

Is it worth it?

ROI for personalisation is hard to calculate and advanced solutions can be both complex and costly. However, there are strong indications that simpler personalisation, such as product suggestions based on previous purchases work really well. A recent study from Adobe showed that 34% of frequent shoppers are likely to buy more when recommendations are based on previous purchases. In sum, it’s possible to create a “personal” and relevant e-commerce experience with fairly simple methods. As always, it is imperative to do the research in order to understand your customers - and to make sure that the data you collect can be used together to form a more complete view.

/David Aler, Strategist

Do you want more tips and need our help to develope your e-commerce? Reach out to Johan Cassel, Sales Manager, 070-494 24 60 or johan.cassel@cloudnine.se.