Personalization Demystified

A woman shops online on a MacBook with a notebook on the desk nearby

In the early days of the web, people thought about content in terms of what every visitor to their site would see. For several years we were stuck with static content that was manually curated and displayed to all visitors. Then content management systems (CMS) made dynamic content possible. You could show related or popular content (remember when those “most viewed stories” widgets started appearing on websites?). In the last two years, technology has evolved to enable not only basic demographic data collection on site visitors, but also to then tailor the experience of a site – from the general content to the encouraged behaviors or calls to action – to each visitor or group of visitors.

What sounds like some sort of miracle of complex technology is actually data-driven and fairly straight-forward. It begins with geolocation and basic profile information. Then we can look at behavior on a site. How are visitors navigating through it? What actions are they taking along the way? When do they share content? What did they see or do just before leaving the site? Finally, we can tap into third-party data stores that track users’ actions across multiple sites and apps, so we can further fine-tune the experience to meet their needs. Personalization has become a powerful tool not only for enhancing user experience, but also for more targeted, effective marketing.

But it has the potential to freak people out and bring your brand bad publicity as well. The key to doing personalization right is understanding how it works and how others are doing it and ensuring as much transparency as possible when rolling out a personalization strategy.

Personalization is here and it’s not going anywhere

Whether people like it or not, personalization is happening now, and it has been for quite a while. Google has been doing predictive search and making search suggestions for years. The company also feeds ads based on email keywords in Gmail and on user behavior across the Internet. Google relies on over 200 data points collected from each user to personalize search results. Facebook similarly suggests content and friends based on your activity elsewhere on the Internet and other indicators, such as your Facebook likes, your age, gender, location, and your ‘ideological fragmentation’ (e.g. your friends’ political leanings, pages they like, etc.). Netflix recommends movies or shows to users based on what they’ve watched in the past, what they’ve rated, and what’s popular in their city. Twitter recommends people to follow based on accounts you already follow and on your interests, as determined by your activity on Twitter.

Consumer sites do it, too

It’s not just the more obvious search engines and social networks, either. Most smartphones are now equipped with technology that enables the tracking of users offline as well, and more and more retailers are deploying sophisticated strategies that enable them to track customers from their online site to their offline store and back again. Consumer websites are increasingly leveraging web technologies to produce multi-channel marketing campaigns, presenting content to users based on a variety of behaviors, including:

  • Purchase history
  • Search history (not just on the consumer site, but on search engines as well)
  • Page visits
  • Shopping cart activity (what users have added to carts and whether they completed a purchase)
  • Comments
  • Ratings and reviews
  • Inbound links (if user followed a link in an ad, search result, or email)
  • Device (whether the customer is accessing a site from a desktop or mobile device)
  • User profile and account information

FedEx Small Business Center is a great example. The company has created a central knowledge center catered to small- and medium-sized businesses. Not only does the site make it easy for users to find articles and useful, relevant information, it also provides recommendations based on interests and browsing history, shows relevant and related articles based on what articles the user is viewing, and shows popular content.

How to succeed in personalization without being creepy

Although personalization has never been more prevalent, users have also never been so concerned about privacy. And while the majority of users are still willing to sacrifice some amount of privacy for convenience or an improved experience, the line between providing great service and invading someone’s privacy is thin and depends on the user.

Here are a few guidelines to engage users and avoid scaring them away:

  • Don’t be overly personal. If you think it’s odd that you know something about someone that they didn’t share with you, don’t use it.
  • Don’t stereotype. There’s a big difference between personalizing user experience and limiting it. Recommend content but don’t rely too heavily on complex behavioral rules that drive a user’s browsing experience. Gender-based marketing, for example, can backfire. Urban Outfitters discovered that female customers not only regularly bought men’s clothing items, but also took offense to being obviously marketed to as females.
  • Don’t overload users with content. In their zeal to tailor content, some companies go too far and overload their users with articles, tips, links and suggestions. Users don’t have the time or the attention span to go through it all. Learn and analyze their actions on the site and other sites so you can effectively tailor both the type and amount of content to each user.
  • Do segment. Identify the personas of your users so you know how to best segment your audience for creating effective marketing campaigns.
  • Do be relevant. Provide customized and relevant information to the user that is easy to find.
  • Do look for overlaps between your business interests and your customers’ needs. The best personalization campaigns exploit the synergies between user needs and business requirements. Offers are great incentives for users, for example, but companies need to ensure that they’re offering users something they want. Finding geographic overlaps is another great way to do this — if a user lives in Los Angeles, you’ll want to promote events or conferences in the area.
  • Do listen to customers and fine-tune their experience. You can’t roll out a personalization strategy and call it a day. Companies must be prepared to continuously test features and content, gather user feedback, and tweak their sites accordingly.

Don’t sacrifice everything on the alter of personalization

In the quest for more and better user data, it’s easy to forget that basic site stats are equally important. Collecting and analyzing data on what works across various touchpoints and campaigns in terms of which pages are most popular, which calls to action are most effective, which emails are opened, the conversion rates of various email and advertising campaigns, which sites are most often referring users, and other metrics are just as important as collecting information about your users and who they are.

Testing both content and campaign strategy in combination with a personalization strategy is the best way to figure out which variations of both content and campaign are working in terms of conversion and bounce rates. Then you can optimize your site accordingly.

––

Learn more about how personalization strategy is easier than you think, best practices for implementing a personalization marketing strategy, and which brands are using personalization marketing successfully.