NEWS: BlueVenn to join Upland Software. Read more

Predictive personalization for better conversions

Predictive personalization

Personalized marketing used to refer to the ability to merge names and key information into a marketing communication. Today it means something very different, with the ability to customize the entire content of an email, including the offer and the product recommendations based on what you already know about them.

Having a personalization strategy be it rules-based or predictive is at the forefront for many organizations, to improve response rates, conversion rates and revenue, along with enhancing the customer experience.

However, personalization done badly can have very negative effects on the customer experience and your budgets. How often has your web browser served you an advert for a product after you’ve just bought it? Marketers need to be smarter about how they deliver tailored content and ‘predictive personalization’ is the route many are taking.

Predictive personalization involves observing how a consumer behaves online and using those observations to determine how likely future behavior (such as making a purchase, completing a form or searching a specific web page) might be. This data can then be used to serve customers content and/or products most likely to connect with their interests.

Delivering customer predictions – say, showing products that are not only relevant to their interests but within their most frequent spending range – are well received by shoppers and have a higher chance of converting.

With rules-based personalization, marketers rely on explicit data. This includes clear and specific information relating to demographics, personal data and data that customers have willingly provided. This will tailor the content you show them when certain rules and conditions have been met.

Predictive personalization uses implicit data, too. This assumed information is an indication of a visitor’s intentions or needs, such as data from browsing history to derive implications about their interests.

Essentially, predictive personalization makes a decision about what content to show them based on their actions, location and characteristics.  

How can predictive personalization be used?


User personas: It’s nothing new creating personas in order to group your customers and creating a profile with attributes relevant to different types of personas is a very useful way to serve up personalized content. A car manufacturer, for example, will have personas who are interested in different makes and models: a mother with children may require practicality and safety as top priorities, while a single, successful saleswoman may prioritize image. These two very different visitors are profiled in real-time and have can have content (i.e. based around car types) predictively matched to their interests.

Browsing data: New and returning visitors to a website can be segmented based on the types of pages they have visited, and a website can be personalized based on the type of content they view the most, or most recently, recent purchases, or whether they are a customer who has been to the site on several previous occasions. For example, your first visit to a new retailer will present you with a promotional message encouraging you to open an account, while returning visitors will be displayed a homepage that highlights the categories and products they showed an interest in during a previous visit.

Demographic data: This type of personal data (age, gender, income, employment, etc.) is often captured through registrations and forms. For example, this could be a travel agent delivering holiday suggestions that are popular with retired couple, or a retailer showing the latest range of men’s watches to a male visitor.

Geographic data: This type of data (location, region, time, weather, etc.) can be collected from a visitor’s IP, Wi-Fi or GPS data and helps brands show personalized content that is relevant to their location. This can be used to ensure that customers go to the right page for a multi-lingual website, serve local offers (such as discounts for a nearby restaurant), or even display product recommendations based on the visitor’s current weather.

Whatever the form of personalization your organization opts for, be warned that it isn’t without its risks – many will tell you that getting personalization wrong is worse than no personalization at all. Show them the wrong recommendations or make and incorrect assumption about their interests and this can damage how a customer perceives your brand. It goes without saying that any strategies you plan to implement should be split tested and observed to ensure they are achieving the goals you set for them.

Cross-Journey Optimization On-Demand Webinar

Cross Journey OptimizationA deeper understanding of the customer journey is changing the way organizations engage with their customers.

In this webinar, we will discuss how marketers need to change the way they think about old metrics and campaigns that have little or no relevance in the new age of the customer, and how marketers can use ‘cross-journey communication’ to better manage the customer experience.

Watch now

Topics: Article