Skip to main content
data-successful-optimization-ai-personalization
In-depth articles

What data do you need for successful AI personalization?

Author
anneclaire
Anne-Claire Bellec

1 The benefits of online personalization

Brands are starting to understand the benefits that AI personalization can bring to their digital marketing. Essentially AI personalization enables companies to deliver a tailored experience to each and every visitor, based on their preferences and behavior, rather than providing a generic, ‘one size fits all’ experience. It is the online equivalent of the personal service a skilled retail salesperson offers in-store.

While the market is still maturing companies that embrace AI personalization are seeing:

  • Greater engagement – with online visitors staying longer on their sites and accessing more relevant content
  • Greater conversions – browsers become buyers as they respond to the right products, offers and messages, targeted at them personally
  • Greater satisfaction – customers are happier as they see content and offers that are relevant to their interests and needs.
All of these benefits lead to greater revenues and profitability –in terms of higher sales, stronger loyalty and increased efficiency for the brand. Applying this personalization approach digitally, at scale, requires brands to analyze the right type of data. How can you collect this and use it effectively, especially in an era where privacy and consumer data protection are key?

2 Collecting the right data - hot behavioral data

There are two types of data that feed into personalization platforms such as Kameleoon’s. The first is hot, behavioral data, generated by visitors when they are on your site. Essentially, it is everything that the visitor does while browsing. This therefore covers a wide range of characteristics and conditions - Kameleoon looks at 40+ factors, which broadly split into three groups:
  • Visitor behavior on your website. For example, what have people clicked on, what is the frequency of their clicks, where have they come from, how much time have they spent on specific pages and how has their journey progressed through your site? You can also look at their history with your site and the number (and timing) of the visits they have made.
  • Information on themselves. Hot data can also include their location and the type of device they are using (i.e. mobile or desktop), as well as their browser.
  • Wider information. Thanks to knowing where the visitor is, you can establish what the current weather is in their location, as well as being able to view the day of the week/month and the exact time where they are.
It is vital to understand that hot data, especially behavioral information, is always changing. This means it needs to be collected and acted on in real-time across the entire journey. There’s no point delivering a personalized experience that is solely based on their last (or first) interaction on your website. You’re building a continually evolving, deeper picture of your visitor that you can use to trigger actions and personalizations.

3 Combining hot and cold data for additional depth

As well as browsing information, brands already have a great deal of data that their customers have shared with them, including:
  • Transactional: What they have previously bought (preferences), what channels they normally use, how much they have spent, when they tend to shop, both in terms of time of day and when during the week or month,
  • Data from their wider technology ecosystem (such as from CRM, DMP, CDP, Tag Management and Analytics solutions) : gender, age, whether they are a VIP or not, location.
Given the range of solutions that hold customer data, you therefore need to deploy an AI personalization platform that can integrate with them all, with open APIs that make it simple to connect and to share data in real-time.

4 Bringing hot and cold data together with AI personalization

It is important to stress that AI personalization doesn’t require huge amounts of information to operate successfully. For example, our algorithms are able to be trained to reliably predict after 2,500 conversions. Based on an average rate of 2.5% this usually corresponds to 100,000 visits. The speed of learning is therefore more about the time needed to reach the 2,500 conversion threshold rather than the actual amount of traffic.

Combining hot and cold data gives a complete picture of the customer and enables deep insight. However, in the real world it isn’t normally possible to identify those people you have data on when they are on your website. 98% of your visitors are either not customers or have not signed in, so you only have hot data to work with.

Using Kameleoon this is not a problem as the range of hot data it collects in real-time gives sufficient insight for personalization, based on their actual behavior. After analyzing the results of AI personalization for three years, our data science team has concluded that hot data provides the information that really counts.

Kameleoon works by scoring visitors to see if they fit into specific, defined targets, which then triggers particular personalizations that match their needs.

Here’s how it works:

  • The visitor arrives on your site and Kameleoon checks to see if they are part of a particular target segment
  • They navigate around the site, with their score evolving and changing in real-time
  • As they do this Kameleoon’s AI continually checks to see if the change in score means they are now part of one of your target segments
  • If and when they cross the scoring threshold, personalization actions are triggered. This could be specific content, offers or messages, created by the brand

5 Ensuring GDPR compliance and optimization

As every digital marketer knows consumers are now much more conscious of the amount of data that they are sharing with brands, and demand that it is protected and only used in their interests. This is backed up by regulations, with the GDPR laying down strict guidelines on how the data of EU consumers is stored, processed and used. Similar legislation is being enacted across the world, with the California Consumer Privacy Act (CCPA) coming into force in January 2020.

Marketers therefore need to be sure that their AI personalization solutions are compliant and secure. They don’t want to face financial and reputational damage due to breaches or poor practices. In the case of Kameleoon, they can rest easy. Our solution only uses anonymized data - this makes it GDPR compliant by design. There is no storage of personal, identifiable information within our platform. All of our software is hosted within specific, secure datacenters located in the European Union. It isn’t just stored somewhere within the cloud. This all provides peace of mind to brands when it comes to compliance and security.

6 Delivering the store experience online with AI personalization

The rise of AI personalization delivers the ability for brands to successfully achieve digital personalization at scale, replicating the tailored experience that a shopper receives in-store. Data is at the heart of this. The combination of hot and cold data gives digital marketers unparalleled insight into customer behavior, enabling them to trigger individualized actions that will improve the experience, drive conversions and improve the bottom line.

To learn more about our AI personalization technology download our whitepaper which explains how Kameleoon benefits online businesses and increases revenues.

anneclaire
Anne-Claire Bellec