Personalisation, also known as customisation, involves tailoring products or services to accommodate specific individuals or segments. It is typically offered as a B2B SaaS — Business to Business Software as a Service. Now that the definition is out of the way, we can get down to the cool stuff.
Prescriptive versus adaptive personalisation
Personalisation falls into two main categories, prescriptive personalisation and adaptive personalisation. Prescriptive personalisation is also known as segmenting. It is based on sets of rules and is triggered by interactions with a user. These rules filter out possible user offerings, such as offering a Bluetooth Speaker if a customer has already purchased one. Makes sense, right? This is the old school way of personalisation, so we’ll put it aside for now.
Adaptive, or predictive, personalisation uses collaborative filtering, data analysis and user profiling tools to adapt content based on visitor characteristics, interactions, intent or any other parameter one might desire to use. For example, a fashion webshop can detect if visitors are from a cold climate, and subsequently choose to offer warm coats to them. Building adaptive personalisation systems involves creating powerful machine-learning algorithms, many big data tests and a lot of coffee. The figure below visualizes adaptive personalisation — the old way and the Taglayer way.
This new way of personalisation is extremely difficult for a website to integrate because you have to connect hundreds of sources (which are constantly changing), which requires armies of developers and data experts. Not exactly a piece of cake.
Improvements thanks to personalisation
Because it takes a proactive approach in shaping customer experience, personalisation has been proven to improve customer satisfaction, web metrics, digital sales conversion, marketing results and branding.
However, many personalisation softwares are lacking:
They can’t personalize real-time, meaning that the system first needs to gather information, then change content and repeat — losing many potential customers and negatively affecting web metrics.
They can’t personalize automatically. This means that marketers have to manually create targeted content, after finding out their visitor profiles by tagging their characteristics.
They segment (Oh no!). Not as bad as it sounds, but most personalisation softwares personalized based on segments of visitors (age, geolocation, gender) rather than on each individual visitor. In this case, there are still a lot of individual-level characteristics that can be personalized to (interests, personalities and personas).
They only use intra-site browsing data. Websites like Amazon.com offer you recommendations based on what you have purchased or browsed before. This is a form of personalisation, but it only uses data gathered from your time on amazon.com — limiting personalisation possibilities. When cross-site browsing data is accessed, personalisation can become much more accurate and expansive. Privacy is also an issue.
They require access to their client’s user database. For example, a big e-commerce website can only be personalized if the personalisation software has access to its user database. This means that personalisation can only be done for users who are logged in on this website. To optimize revenue, you want to be able to offer relevant products to visitors regardless of if they’ve been on your site before. This links to the previous point because the client’s user database is considered as intra-site data.
Okay, so there are clearly many things missing in contemporary personalisation softwares. Fear not, Taglayer has tackled all of these problems. Stay tuned for the web revolution.
Written on April 24, 2016