Building a Personalization Roadmap: 6 Considerations
The term “Personalization” is a massive phrase that spans industries, tools, and teams. Therefore, it’s no surprise that personalization is on almost every organization’s roadmaps.
One of the most frequent things we hear after starting a conversation around personalization sounds something like “we know we need personalization, but how do we do it without seeming intrusive?” The good news is that there are strategies you can use to during the development of your personalization roadmap to help find the balance between underpowered initiatives and creepy data utilization. The key? A deep understanding of your industry, users, their needs and your data.
Know Your Industry
Understanding your industry, its players, and how they relate to personalization is the highest-level consideration that will need to be made. Personalization strategies in one industry may not apply to another. The products you sell, the maturity of personalization in the industry, and the demographics you serve are all variables you will want to consider when evaluating your personalization strategy.
Some industries can withstand (and some demand) more granular and direct personalization strategies that call directly on users’ personal preferences. Other industries thrive on campaigns that are more suggestive in nature, avoiding an over-reliance on personal preferences and details.
This can be seen, for example, when comparing the cosmetics and medical industries. Where cosmetics and medical related products are both personal in nature, in most cases makeup is less emotionally sensitive as compared to medical related products. Cosmetics shoppers want their preferences to be known in order to streamline their shopping journey. In contrast, medical shoppers are actually dissuaded from purchasing from organizations that make it obvious that their medical/health-related data is being put to work.
Where both of these industries could benefit from helping users understand which product most closely fits their needs, the demand for assistance that personalization brings greatly differs. Being conscious of these sensitivities is the first stepping stone to finding the balance required for successful personalization.
Understand Data Collection Methods
Collecting data and understanding how that data was collected is important. Most companies have multiple data sources and multiple methods of collecting that data for each source.
Understanding how the data was collected and how it relates to the larger customer profile is paramount to creating a seamless personalization campaign. This will allow you to be cognizant of how the campaign could be perceived from the user’s perspective if you plan on using multiple variables from multiple data sources. This understanding reduces the “how did they know this about me” reaction that can occur when users feel that the brand knows more information about them when they feel they have provided to the brand.
Users provide data in two ways, either overtly or covertly. Overt data is when a user knowingly provides the brand with a piece of information, like when they sign up for email, create an account, take a quiz or place an order. Covert data is normally seen as behavioral, mined or third party data where we are assuming something about someone based on something they did, data we mined, or what a look-a-like segment did.
Using data that was provided in an overt way tends to be seen as less intrusive and provides a more accurate personalization experience. On the other hand, using behavioral and mined data allows you to personalize to a larger quantity of your users, but when the personalization campaigns are too accurate and users don’t understand how you know that information about them, you run the risk of the campaign feeling intrusive.
Finding a balance between using both types of data and understanding how the data you have collected is important to understanding how your personalization campaigns are perceived by your audience.
Consider Context
When deploying a personalization campaign, it is important to consider the context through which it will be experienced. This entails understanding how different segments may encounter the message alongside other personalization campaigns (segment overlap), the delivery method, the timing of delivery and the degree of personalization that is occurring.
Personalization efforts that fall outside of the appropriate context can more easily be seen as intrusive, not useful, or even just flat out wrong. Once you identify an attribute that you want to use for a personalization campaign, over-using it, under-using it or delivering the campaign during the wrong interaction can cause confusion.
The key is to match expectations to reality. If your user expects to see or experience something, but you personalize incorrectly for that experience, your personalization campaign can fail. This does not mean that the entire campaign needs to be tossed, but rather you should start by taking a closer look at the context. Perhaps the same personalization campaign would work better on another page, for another segment, or in another delivery method. When optimizing a personalization campaign, make sure to consider the entire context of the experience to see how you can better fulfill your user’s expectations.
Test Delivery Methods
Automated personalization tools may help you determine which piece of content to show to specific audiences at specific times. But a human still needs to select the pieces of content to be uploaded into the tool.
The same goes for the layout of a page (though some tools can personalize that too) and layout of individual page widgets (think product carousels, hero images, and banners). Though personalization tools can help serve the right message, image, or CTA to each user segment, the layout of the widget containing the content should be tested to ensure that both the content and delivery method are both optimized.
Here are some examples:
If you are using a “Recommended Product” carousel on your site, personalization tools can help populate the carousel with products that have the highest probability of being purchased by each user segment but most likely all are being populated within a carousel widget that has a consistent look and feel. Testing the widget layout, how you show sale prices, offering an add to cart CTA or number of items shown are all variances that can be tested, regardless of personalization algorithm used.
The same goes for hero images and banners. Though the personalization tool can help determine the correct content or message to show to whom, test how that content is being shown. With hero images, you can the layout of the image with the text, call to action, and the size of the hero image, just to name a few.
When it comes to content, the personalization tool can help determine which piece of content from the lot that you provided it is best for each user but it’s up to a human to determine what 10 pieces of content to upload. Continually testing the tone, theme and overall aesthetic can help reassure you that you are providing the most optimized content to the personalization tool.
Just Because You Have the Data Doesn’t Mean You Should Use It
Do you have a ton of clean data? Congratulations, you’re ahead of the game. It may be tempting to start personalizing off of a bunch of that data because, hey, that’s why you collect it, right? Well, yes, but use it cautiously.
Using too many data points, gathered from multiple sources, in a different context, can quickly result in a personalization campaign that is too specific. Highly specific personalization campaigns require users to have a high tolerance and for you to have delivered the personalized content very accurately.
Having a user base that has a high tolerance for personalization usually comes from one of two things. 1) Your user base is a bunch of personalization strategist (not likely) or 2. You have built trust with your users by first personalizing with more generic or simple campaigns and moving into more specific or personal campaigns.
When it comes to the type of data to use that will be dependent on your company, industry and what is available to you but it’s important to remember to always be cognizant of laws around data use and respect user’s choice to opt-out of being tracked.
Just because you have the data doesn’t mean you should use it all at once. Use it to build a relationship with your users and when your users become more engaged with your brand you can become more engaged via personalization, regardless of when it became possible for you to personalize against a certain data point.
When in Doubt, Let the User Know Why they are Receiving Personalized Content
If a personalization campaign is touching on highly specific things or is being delivered via a new medium, simply let the user know why they are receiving the content. For example, you can say “we noticed you looked at our blog about breathable fabrics for summer and thought you would be interested in these 100% cotton products”.
Personalization can be extremely powerful but striking the right balance between helpful and overt can be tricky when a good strategy is not in place. Start by understanding the problems that you are trying to solve, your user base and data to make sure you are serving an experience that is helpful to your users, helpful to your brand and scalable. Finally, become hyper-focused on customer intent and understanding how to satisfy your users’ expectations during that experience.
