Personalization Best Practices 2019

Website Personalization: Best Practices in 2019, from the Visitor’s Perspective

Website Personalization has proven more and more significant in recent years, as life has become increasingly digital. Much more than simply a trend in marketing, it’s a natural and inevitable evolution in our relationship to technology, where digital experiences increasingly adapt to us as individuals.

This also means that we, as digital marketers, have to constantly innovate and test in order to stay ahead of the curve, in a highly competitive market. From the website visitor’s perspective, experiences on any given site should simply be easy, intuitive, and relevant. This is a bar being continually raised by big names and major innovators like Amazon, Netflix, or Spotify. Anyone who wants to stay relevant into 2019 and beyond must equip themselves with the tools and strategies necessary to keep up.

Luckily, SaaS personalization solutions have been making it more and more cost-effective to do so, putting the tools in just about anyone’s hands to drastically improve website performance and revenue, using personalization technology, for a relatively minor investment.

So, what do best practices for website personalization look like, from the visitor’s perspective, in 2019? Here are some quick guidelines.

Product and Content Recommendations

Today, the applications of recommendation algorithms in various types of online businesses is only expanding, and whether you’re in e-commerce, publishing, or even just a company with a blog, there are uses for a powerful recommendation engine to increase engagement on your site.

Bottom Line: using algorithms and/or machine learning to show tailored content based on known data about the visitor always makes for a more engaging user experience than a static site.

Product Recommendations

Requirements

1. Product recommendation engine

2. Product feed with relevant data such as interest category for each item

3. Product interactions tracking, either by the recommendation solution if available, or internally on your site and sent to the engine somehow (JS pixels, for instance)

4. Enough traffic and transactions to generate relevant recommendations, ideally 100k per month or more

By this point, it should go without saying that e-commerce customers have been trained by giants like Amazon and eBay to expect to be shown relevant product recommendations throughout their journey with a brand, from homepage to remarketing emails.

That means it’s your job to deliver insightful product carousels powered by a data-driven product recommendation engine that will keep them oo-ing and ah-ing all the way to checkout, and beyond.

These will not be universally the same for all markets and e-stores, but you can consider these ideas to be a combination of the tried-and-true as well as the state-of-the-art, in 2019.

Homepage

As a return visitor, a best practices example for ecommerce recommendations might look something like this:

On the home page, you are greeted with a combination of view it again recommendations, reminding you of products you’re interested in but haven’t purchased, and the somewhat mysterious “Recommended for You.” The key to the latter is a machine learning algorithm, also known as predictive personalization, which uses a combination of crowd data and your own individual interests and demographics to show you products you didn’t even realize you wanted.

Scrolling on, you also see another “Items from Your Previous Carts” carousel, which reminds you of items you’ve added to your cart in the past, but for whatever reason didn’t complete. Overall, the feeling these types of recommendations will inspire is enjoyment and engagement with the site, and which make it clear that you are using information from their past visits and behavior. In data-paranoid 2019, that can be a critical point.

Product Page

As you find your way onto product pages, your personalized journey continues, as you view individual items. On each product page, you’re shown items that were frequently bought together with the item you’re currently viewing, in an equation-style display reminiscent of what you’ve seen while shopping on Amazon.

On items with a value over $100, you also notice a new recommendation carousel, “Similar Items for Less.” This carousel shows you items from the same category that are below $100 but is not present on cheaper items.

However, on items that are priced below $100, a different carousel takes its place: “Most Popular from This Category.” On every product page, you have two carousels of AI-recommended products that learn from every interaction, and keep you clicking and engaged.

Cart Page

When you’re finally ready to check out, you’ll not just simply see the items you’ve placed into your cart. Additionally, you notice a product recommendation carousel below the items in your cart, which suggests “Popular Accessories for These Items”. On the side, you might also see a vertical side-bar titled “Care for an Upgrade?” In it, you see tempting, more premium items similar to those in your cart.

Should you have doubts or indecision and attempt to leave the page, you’re presented with an exit popup. This isn’t any generic exit popup, however, but recommends products similar to those in your cart, and also presents you with the option to either buy now with a 10% discount or to have your cart sent to your email, for later checkout.

Content Recommendations

Requirements

1. Content recommendation engine (may be same as product engine)

2. Content feed with relevant data such as interest category for each article, author, etc.

3. Content interactions tracking, either by the recommendation solution if available, or internally on your site and sent to the engine somehow (JS pixels, for instance)

Content recommendations are important not just on publishing sites, but also for any business that uses content marketing, such as any website with a blog. These can also coincide with product recommendations, and can even cross-over. For instance, on a product page you can recommend content relevant to that product, or vice versa. Just like product recommendations, it’s important to show the visitor what is most relevant to them at each point in their journey.

Although content recommendations can be used in a variety of contexts, in the example of a content or publishing site, the idea is always to increase engagement, as measured by metrics like session length, time on page, and scroll depth. This means that the use of recommendations will be slightly different, while still utilizing a combination of individual and crowd data to optimize the visitor experience.

Homepage

What this might look like from the reader’s perspective is that the homepage will contain recommendations such as “Finish Reading,” which reminds you of posts which you viewed in the past, but didn’t scroll through completely. Another example would be to show the latest posts from your favorite (most read) author, and/or the most popular articles from your interests.

Alternatively, if you are a new visitor, you might see a more general recommendation which utilizes machine learning to guess what you might be interested in, based on whatever data is available.

Article Page

On individual article or post pages, you’ll see recommendations that are related to what you’re reading. This could be something like “Most Popular from This Author,” for instance, or “Related Content” showing popular articles from the same interest. You might even see separate recommendations showing posts related to different interest tags associated with the piece, if it pertains to multiple interests.

For instance, on a news article about the president’s meeting with a diplomat from Turkey, you might see one carousel showing articles related to the president, and another about US foreign relations in general.

Category Page

This is another place where the reader’s interest is relatively straightforward. If you are viewing a content category page, then you’ll probably be seeing content recommendations such as “Most Popular from This Interest,” as well as “Most Popular Authors for This Interest,” in addition to the standard Most Recent queue.

Targeted Website Content

Aside from the more automated and personal AI recommendations, there’s also a great deal you can do towards a more engaging experience on your site by simply creating different experiences per segment. These segments can be based on a variety of factors, such as:

• Geo-targeting: visitors from certain locations, regions, or countries

• Behavior: those with certain interests derived from behavior in past visits

• Site Visits: those who have returned to the site however often, in a certain time-frame

• Referral: who arrived from a certain referral source, like a Google ad

• CRM Targeting: something like account type or industry, derived from your CRM data

Whatever you use to identify a segment for whom the content on the site will be slightly different, you’ll want to maximize their exposure to the offers, products, content, or messaging that is most ideal for them. To do this, you’ll need to both utilize your existing customer knowledge, as well as state-of-the-art personalization tools. Generally, targeting can be divided into two categories: adding new content, or changing/removing existing content; in both cases, the changes will be seen only by your selected segment.

Adding New Content

Targeted Banners

Requirements

1. Targeted banner/popup/forms content solution, or personalization solution with these tools included

2. Creative for many different banners/popups/forms per segment

3. Traffic insights and strategy to know who to show which content

4. Ways to identify segments, such as ad referral, on-site behavior, geo-targeting, etc.

In 2019, people have been overexposed to banner advertisements on the web, and are also accustomed to them being personalized, thanks to Google Adwords. This means they don’t want to see too many banners, and when they do see them, they’d better be relevant to their interests!

Luckily, with modern SaaS personalization technology, this is relatively easy to achieve.

Targeted Banners in Ecommerce

From the visitor’s perspective, on an e-commerce site, this will look like offers that are specifically relevant to you, as a customer.

This feels like a unique offer that is just for you, or for a limited group you are part of. For example, “FREE SHIPPING only for visitors from Ohio!” Now, the reality may be that you are only getting the offer because you are close to a distribution center, or it may not be exclusive to Ohio at all; you’ll never really know, but either way, you’ll be happy not to pay for shipping.

On the other hand, it could be related to your specific product category interest, such as an ad for the most popular line of products in camping gear, if you’ve shown an interest in the outdoors.

Targeted Banners On Other Types of Websites

On other types of sites beyond e-commerce, banners may have a variety of different purposes, but they can still be personalized.

For instance, in a SaaS website personalization example, your banner may have a personalized call-to-action, which is different depending on the industry or role in company of the visitor, as derived from CRM data. If you’re someone from management visiting, you might see a banner CTA emphasizing the statistical improvement to ROI by big-name companies who used the software, while if you’re from IT, you might see one that emphasizes specs and features.

Targeted Popups

Popups are the quintessential “potentially annoying” type of advertising, so this is another area where you must be very careful and strategic. Again, personalization technology can help immensely with that, because you can show them only when they are most likely to be interested, and show messaging and offers that are relevant to the individual.

There are several ways to make popups more tailored via personalization, for different types of websites. In e-commerce, one of the most powerful applications of this technology is the personalized exit popup. This is partly because, at this point, when the customer is attempting to bounce, it's more of an all-or-nothing moment, to attempt to keep them on site or increase the likelihood they’ll come back. But how best to do that?

Annoyance-Avoiding Strategies

First of all, the exit popup can be limited to only show to a limited segment, such as those with items in their cart. For instance, if you as a customer try to bounce with items in your cart over $100, you can be offered free shipping via exit popup, while if you don’t have cart items, you might be offered a discount on the actual price, such as 10% off if you continue to checkout.

On the other hand, you may want to incorporate recommendations into your popups, whether they’re based on exit intent, or just set to display on certain pages, or after X amount of idle time.

This would look like a recommendation carousel that appears either on the sides or bottom of the page or in the case of an exit popup, would probably appear in the center with a glass panel behind. It might show you the most popular items from the category you’ve been viewing, or even the cheapest, with messaging like, “Before you leave, here are the best deals from this category:”

Other Strategies for Personalized Popups

Of course, there are many related strategies for popups. One of my favorites is the “Send Me My Cart” option, as it’s a great excuse to collect customer emails for further remarketing, and can also be added to any of the previous examples, as a secondary option button.

Another cool strategy is to show the visitor a map exit popup, which works like an automatic store locator, displaying their nearest locations. Of course, this only works if your company also has physical, brick-and-mortar locations, but there are plenty of appealing ways to present it based on your industry.

For instance, if you run a shoe site with physical locations, you can say “Want to see how comfy it fits? Come in and try it on!”

Forms

Forms can also have many purposes, but often they are prompting the user to sign up for a newsletter, email list, or similar email content marketing. This is a case where personalization can make the offer seem quite a bit more appealing, such as by emphasizing that the content of the newsletter will be specifically related to their interest.

For instance, if you are visiting an online scientific publication’s website, and are presented with a popup form prompting you to sign up for updates about conferences, what you see might be different if you’ve somehow identified yourself as being interested in a particular area of scientific interest, based on what you’ve read, your original referral, or other sources.

Personalization Opt-In

Targeted forms can also be used to prompt visitors to reveal information about themselves, making personalization more explicit. In 2019’s data-paranoid times, this can actually be a good idea, to reassure the visitor that it’s their choice to have a personalized experience.

It’s not hard to imagine how if you, as a visitor, can tell a site what you’re interested in with just a few clicks, and ultimately consent to have a better more personalized experience, there’s a good chance you’ll want to do that. If not, you’ll likely still respect the fact that they gave you the choice.

Changing/Removing Existing Content

What you may need is not new content added to your website, but to actually change content that is already there. Here are some examples of the kinds of personalization tools and methods at your fingertips in 2019 to create optimal site experiences for your visitors, by strategically changing content.

Requirements

1. A/B Testing or segmentation solution with editing tools, or personalization solution which includes these tools

2. Editing tools such as WYSIWYG editor, and others, in the chosen tool

3. Insights from your customer/audience knowledge to hypothesize which edits would improve site experience for which segments

4. Segmentation/Targeting built into the tool

WYSIWYG Editor

Many personalization solutions include a WYSIWYG (What You See Is What You Get) visual editor, to make changes to your page which will only be seen by a certain targeted segment. This editor allows you to change images, text and other page elements throughout the site.

Examples:

• Remove/Change Menu Item(s) - For when you want a certain segment not to see certain items on your main site menu, or to see a different menu option, for whatever reason.

• Adjust Homepage Creative - Change the text and/or images on your home page in part or in whole, to be more appealing to various segments.

• A/B Test Design Tweaks - Curious if your site might perform better if the logo was smaller if the mid-page CTA were moved up higher, or similar small changes? WYSIWYG editor plus A/B testing tools will allow you to find out.

Replace a Key Word, Phrase, or Image

Change a keyword, phrase, or image as it appears throughout your entire site. This type of tool allows you to tweak your messaging or creative throughout your whole website in one fell swoop.

Examples:

• Replace the word “best” with the phrase “most exclusive” everywhere it appears on the site, for visitors known to have high spending habits.

• Replace American-spelled words with their British spelling, for visitors from the UK, or vice versa.

• Replace an image with “rural” appeal that is used more than once in the site with an image that has a more “urban” appeal, for visitors in major cities.

Test Perpetually

Of course, testing, analytics, and evidence-based marketing strategies must be built into your digital marketing procedure on the whole, but especially when it comes to such a rapidly changing field as personalization. When Facebook’s data privacy fiasco hit the newsfeeds, for instance, data collection and privacy concerns changed almost overnight.

This means A/B and multivariate testing should be a core component of your digital marketing and personalization strategy. Luckily, there are a variety of A/B testing tools out there, and many personalization solutions will include this as part of their functionality, allowing you to A/B test any given personalized content you create, and view detailed statistics internally, as well as integrate seamlessly with your third-party analytics solution.

When it comes to testing strategies, personalization technology allows you to test not only different versions of your personalized content but also different targeting rules or recommendation algorithms. This opens up many new possibilities for optimization.

In Conclusion

Too many digital marketers, the task of personalizing a website can seem daunting and full of uncertainties. How much is really necessary? What is the right approach?

The good news is that modern SaaS personalization solutions make it relatively simple to implement the tools and strategies described here. In 2019, one of the keys seems to be just enough personalization to make the visitor’s experience more engaging, without being creepy. Another key to this is being relatively straightforward about the fact that you are using their data to provide a better experience.

Not Optional in 2019

Failing to implement personalization always means leaving money on the table. Every exit popup you deploy, every recommendation widget you install, will be generating some percentage of your revenue. Total lift from personalization can range from 5-30% on average, depending on a variety of factors, especially how well and to what extent they are executed.

To find the right personalization tools for your company’s unique needs, check out this list of personalization solutions.