In 1999, David Weinberger, a technologist and co-author of The Cluetrain Manifesto, wrote, “Personalization: the automatic tailoring of sites and messages to the individuals viewing them so that we can feel that somewhere there’s a piece of software that loves us for who we are.”
Two decades later, ironically, personalization is being used by companies attempting to make the online experience more human.
Personalization has grown rapidly since David’s statement. So much so that personalized experiences have become the norm, not an option. [Tweet it!]
But is personalization really the key to a better user experience? Does it actually impact conversion rates?
Table of contents
- What is web personalization?
- Static content vs. dynamic content: which converts better?
- Where personalization goes wrong
- Can you “go too far?”
- Getting started with personalization (in any industry)
What is web personalization?
“Personalization” is rather self-explanatory, but more complex than it’s given credit for. For example, you could be dealing with different channels (on-site, in-app, mobile web, and so on).
Essentially, it’s the real-time individualization of a site to suit each visitor’s unique needs and guide them through a custom conversion funnel.
Here’s an example of on-site personalization from Bunting. Visitor 1 is a woman from the USA, who is browsing in June or July…
Visitor 2 is a man from the UK who has browsed Nixon watches before, but never purchased. He is visiting the same site (not a separate landing page)…
Notice how the two pages are different based on either the visitor’s demographic or previous behavior? That’s on-site personalization in action.
Why is this relevant? People are drowning in information and options. Personalization reduces the amount of information and the number of options to help guide visitors through a funnel that’s designed just for them and their individual needs.
So, how do marketers feel when it comes to personalization? According to a survey by Evergage, an overwhelming 98% of marketers believe that personalization advances customer relationships.
When it comes to using personalization, this is the current channel breakdown:
Personalization keeps rising. Compared to data from previous surveys by Evergage, positive results from personalization keep growing year-over-year, and even more organizations plan to maintain or increase their personalization budgets.
Increased conversion rates, increased visitor engagement, and improved customer experience are seen as the biggest benefits to personalization…
Conversion rate is the main measure of personalization success…
The hypothesis is that the more personalized the site / user experience, the more likely visitors are to convert.
So, how exactly are marketers personalizing their sites? Here’s how the popularity of certain formats holds up…
Popups, information bars, call-outs—nothing you haven’t seen before. In-line content and edits are at least holding their own, but there hasn’t been much innovation in the space when it comes to “types of personalized web experiences.”
Personalization is powered by big data, which means you have a lot of options when choosing how to segment your audience for personalization purposes.
Evergage’s data from a few years ago revealed that many marketers rely on elements such as device, geolocation, and content viewed as personalization criteria. Fresh data revealed and added to that list campaign source, email opens and clickthroughs, purchased products, and more:
Customization vs. personalization: what’s the difference?
This is a common question: What’s the difference between customization and personalization?
Often, you’ll see the two terms being used interchangeably. That’s unfortunate because there is a distinction between the two. Definitions may vary, but to summarize…
- Customization: The visitor deliberately chooses between options designed to make the user experience more personal.
- Personalization: The visitor is automatically shown personalized pages based on anticipated needs or wants.
Customization is the visitor’s conscious decision to alert the user experience. Personalization is making predictions and automatically altering the user experience based on big data.
Good(ish) example: Netflix
Amazon is an obvious example of personalization. Their product recommendation engine is often applauded. Instead, let’s look at another big company that’s personalizing the user experience well: Netflix.
When I log into my Netflix account, here’s what I see…
As you might’ve guessed, I watch a lot of dramas and sitcoms. If I suddenly start watching a lot of comedies, you’ll see that category move closer to the top of the page.
Netflix is predicting, based on my viewing history, that I’d like to watch a drama or sitcom. It also shows me “Top Picks” for me based on what I’ve watched.
One thing Netflix could do to improve the user experience (vs. focusing only on showing me personalized content) is de-prioritize shows and movies I’ve recently watched from their recommendations. For example, the only movie under “Dramas” that I haven’t recently watched is Lila & Eve.
Yes, I’m likely to enjoy those movies, but if I’ve already watched them, why show them to me again? The experience isn’t as personalized as it was before. The system didn’t keep learning.
Next, scroll down the page a bit and you’ll see this…
Recommendations based on what I’ve recently watched. These recommendations could be based on actors, directors, genre, rating from people like me, and so on. In this case, a lot of data is being combined to personalize my experience.
However, there’s another problem. I rated Inside Out 1/5 stars (unpopular opinion, I know). Why show me suggestions based on a movie I didn’t enjoy?
A lot more goes into personalization strategy than most people think. It’s not just popups, call-outs and “product recommendations like Amazon.”
What tools can you use?
When you’re dealing with so much data, you’re going to need a tool to help you out. Here are just a few of the top web personalization tools available today…
Static content vs. dynamic content: which converts better?
The honest answer is a frustrating one: it depends. Perhaps all of your visitors expect / need a similar experience, maybe you personalize the experience with the wrong messaging, maybe you try to personalize at the wrong time, etc. There’s always the need to test (and optimize) it for yourself on your own audience.
However, there is a lot of logic behind the hypothesis that a more personalized, guided experience converts better. There are also quite a few case studies to support it.
Product recommendation case studies from Bunting
We’ve already talked about how Amazon and Netflix do product recommendations, so this should feel familiar. Bunting worked with At Home In The Country to add personalized product recommendations to their site, which resulted in a…
- 13% increase in revenue;
- 3% increase in average visit duration;
- 12.5% increase in conversions;
- 4.5% increase in product views.
Here’s what they did…
You’ll notice they placed alternative product options below the product currently being viewed. Some products were given accolades like “Most Popular” or “Best Seller.”
What’s interesting is the use of the word “Buy” in “What other items do customers buy after viewing this item?” Normally, you’d see something like “Customers also viewed…” Buy is much more active, giving the visitor more confidence.
Bunting also helped Bras N Things implement product recommendations, which resulted in a…
- 9.6% increase in average order value;
- 3.5% increase in average visit duration;
- 3.3% increase in product views;
- 2.4% increase in site visits.
Here’s what they did…
Now, Bras N Things had already built their own product recommendation engine. Working with a tool that could handle more data and make the personalization process more efficient, Bras N Things saw improvements across the board.
Note that many of the recommendations include sales and promotions in red font to capture more attention.
In-app messaging and onboarding from Evergage
Personalization for SaaS sites is less commonly discussed, but it exists and can be just as effective.
Evergage worked with Citrix to serve their visitors with the most relevant content based on their interests and stage in the funnel.
Citrix now offers in-the-moment, personalized experiences and content to guide the visitor toward the right Citrix solution. This is powered by Citrix’s existing data sources and account-based marketing features.
They no longer make broad assumptions about what could work (and waste significant time making it happen).
With Mindflash, Evergage deployed a personalized, relevant in-app experience by tracking visitors from anonymous to named users, which resulted in…
- A 30% increase in leads converted from organic search;
- A 15% increase in user adoption of key onboarding actions;
- A 40% increase in accounts updating payment information.
Here’s what they did…
This approach allowed Mindflash to focus their efforts on the needs of users based on which campaigns and sources they arrived from. Using Evergage, they tested each campaign on a smaller scale before rolling it out to all the pages.
To convert more free trials fo paid users, Mindflash needed to encourage more users to complete essential steps, which only some of them did. With in-app, in-the-moment messaging, they increased the completion of these steps by 15%.
Furthermore, they segmented certain user groups based on their usage and presented them with relevant hints and tips to get the most out of the product.
Where personalization goes wrong
While the four case studies above demonstrate that personalization does work, it’s important to remember that it can go wrong. And when it does, it goes very wrong.
There are two major mistakes marketers make when it comes to personalization: not truly personalizing the experience, and segmenting audiences ineffectively.
1. Content vs. experience personalization
Most marketers have content personalization, not user experience personalization. What that means is that personalization has become synonymous with “the right message at the right time to the right visitor.” While that is certainly part of it, that is not the be-all-end-all of user experience.
Karl Wirth, CEO of Evergage, suggests that there are four core principles of user experience (UX): remember, understand, help, and surprise / delight. Personalization should touch on each.
Imagine you receive an email from your bank about refinancing and home equity loans. You’re interested, so you decide to click through. Unfortunately, the site’s UX is lackluster and you end up taking the action, but choosing another bank…
Karl Wirth, Evergage:
“First, instead of sending me to a landing page, they could have made every part of their site a ‘landing page’ that continued to speak to me about the topic that got me interested in the first place. Secondly, the bank could have paid attention to what I was doing, which was perusing their site and others’ for home equity line options, risks and rewards. I even put my information into their calculator, but they didn’t take that data and turn it into an opportunity to talk to me in a personalized manner. Furthermore, they could have tied these two things together to close the sale by telling me why I should do business with them instead of going somewhere else. And lastly, since I’m often also on my mobile phone, they could have tried to reach me there as well, via advertising or an app.” (via Evergage)
Context is important here.
- Who is this person?
- Have they been here before?
- Have they entered the conversion funnel at all?
- Have they made a purchase before?
- What pages have they read?
- How thoroughly are they reading those pages?
Once context is accounted for, you can focus on intent, motivation and anxiety for that individual visitor.
Instead of stopping at a popup advertising a home equity eBook, go the extra mile to cover the entire experience. Remove distracting sections of the site that are unrelated, simplify the navigation, change images to match their demographic, and so on. There are tons of possibilities.
Let’s say you’re a distributed startup team frustrated with your web meeting software. You start looking for alternatives and find a viable one, but the UX is a disaster…
Karl Wirth, Evergage:
“They could engage me based on my high-level attributes. For instance, my IP address would show that I’m from a high-tech company in Boston, so they could offer me relevant case studies from other respected high-tech firms in Boston. Or since they were actually in Boston exhibiting at a trade show the very day I was on their site, they could have invited me to come meet with them and engage in person. Or they could have paid more attention to my affinities—such as the fact that I was reading articles about tech-focused businesses, and looking at web meeting software versus payment software or any of the other types of software they offer—and provided me more focused content based on those interests.” (via Evergage)
In this case, Karl is a new visitor. They don’t have the same information his existing web meeting software company would. What they do have is his behavioral data, demographic information, and more.
As Karl browsed the site, he undoubtedly displayed certain traits that would signal that he is comparison shopping, likely to switch software. There should be conditions in place to detect that type of behavior and then an entire user experience for people likely to switch from another provider.
If you combine personalization and customization, you can get as granular as switching from Product A, Product B, or Product C to your product.
What’s important here is that you understand:
- Who the visitor is;
- Why the visitor is there;
- Where the visitor is in the buying process.
When you understand that, you can create entire user experiences for very specific segments.
Karl bought a microwave and a dishwasher online. When he went back online to buy a new stove, the retailer dropped the ball. Even though Karl had a track record of buying expensive items, no personalized effort was made to help him in his discovery and make his next purchase…
Karl Wirth, Evergage:
“They could have helped me right away by keeping track of my browsing and buying history. They could have paid attention to the fact that I recently bought a microwave and a dishwasher, factored that in, and recommended the best stoves for me. And then when I’d shown a lot of interest in a few stoves, but left to look at other sites and came back, it would have been nice if they’d directed me right to my recently viewed items instead of making me search for those products all over again.” (via Evergage)
Help your visitors help themselves. It starts with qualitative research because you need to understand the different buying processes you’re dealing with.
Karl gives a good example when he mentions leaving to look at other sites and then returning. Someone who spent a good deal of time on your site browsing products, but then left only to return 10-15 minutes later should have a much different UX than someone who is visiting for the first time.
It’s not a matter of a call-out, it’s a matter of fundamentally changing the UX. Focus them on the product categories they were searching in previously. Show the products they recently viewed. Show reviews of those products.
If you truly understand, you should be able to help subtly and intuitively.
Using the same stove example, Karl explains how delight comes into play…
Karl Wirth, Evergage:
“I want to buy a decent stove, but I’m not an expert and I really don’t care to be one. I need to learn from other people or sources of content. If the first retailer would have proactively put me in touch with other stove-buyers who were like me, I could have learned from them. Instead I was poring through reviews that weren’t very relevant. It was challenging and frustrating; the store was putting the onus on me to become an expert on stoves. I simply want to let the store know what I need and have others help me. That would have gone a long way, but there’s more. The store missed a real opportunity to reward my loyalty.” (via Evergage)
It’s about going that extra mile. Not only do you need to anticipate the visitor’s next want / need in personalization, you need to anticipate the nice-to-haves and the I-hadn’t-thought-of-thats as well.
Delight is about modifying the anticipated UX in a way that unexpectedly eliminates work for the visitor.
For example, what Karl described above about “poring through reviews that weren’t relevant” to “become an expert on stoves” is fairly standard. Ecommerce shoppers expect to conduct their own research, shop around for the best deal. You can eliminate that work for the visitor by automatically showing reviews from people like them who have purchased the stove they’re looking at or comparing the price of that stove to the price of the stove from three other retailers if they attempt to exit.
Jordan Julien of Hostile Sheep Digital Experience Lab has a few UX principles of his own…
Jordan Julien, Hostile Sheep Digital Experience Lab:
- “Be contextual: It’s often easy to think of a user journey like a storybook. If you open most books to any given page and select a word, you’ll be met with an abundance of context on the page. You’ll usually see the title of the book, the chapter, the page number, and the word will appear contextually within a sentence, paragraph, and page. Ensure that users are contextually aware of where they are within their journey.
- Be human: Be approachable, trustworthy, and transparent. Provide human interactions over machine-like interactions.
- Be findable: Establish a strong information scent. Provide wayfinding signs.
- Be easy: Reduce the user’s cognitive workload whenever possible. Be consistent and clear, and establish a strong visual hierarchy.
- Be simple: Establish a strong signal-to-noise ratio. Avoid distractions, jargon, and long loading times.” (via UXmatters)
When innovating with personalization and attempting to think outside the box, start with these five principles. Ask yourself what you can personalize to…
- Provide more context to visitors;
- Provide an experience that’s less machine-like;
- Help visitors find what they’re looking for faster;
- Make it easier for the visitor to do what you want them to do;
- Reduce the amount of irrelevant or secondary information and options presented.
It’s a change in mindset. It’s no longer about “What message can I deliver when, and where?”; it’s about changing the UX itself to feel more personal, more individualized.
2. Segmenting ineffectively
When you’re segmenting for personalization, you have options. Here are just a few examples of the type of information you can use to segment your audience…
- Behavior on any channel (web, email, mobile).
- Demographic information like age, gender, education, ethnicity, martial status, etc.
- Geographic information like city, state / province, country, time of year / season, time of day / day of week.
- First-party data (i.e. data you have collected yourself).
- Third-party data (i.e. data from your CRM or POS).
You aren’t limited whatsoever. This is a big data industry, so take advantage of it instead of sticking to site behavior, gender, and age.
For example, Nielsen Norman Group found that personalizing intranet portals based on the person’s role at the company vs. the person’s previous behavior is more effective.
You can find a lot of segmentation inspiration in your first-party data. The best way to do that is to go to your Google Analytics account and begin segmenting your data as you normally would. What segments are you using?
Some examples include…
- Browser / Device type. How would you change the UX for an Android visitor? Or someone who uses IE 9?
- Page load time. Should you treat visitors who waited longer for your site to load the same way?
- Day of week / Time of day. Does someone browsing your site at 3 a.m. on a Friday have different motivations, intentions, and anxieties than someone browsing your site at 1 p.m. on a Monday?
- Conversion funnels. Where do people drop out of your funnel most often? Combine the quantitative data with heuristic analysis. Why are people dropping out? What is so difficult about this step in particular? How can personalization fix it?
- Landing page type. Are you dealing with someone who is on your main site or a dedicated landing page?
- A/B test variation. If you’re running A/B tests, visitors who land on variation A will have a slightly different experience than visitors who land on variation B, right?
- New vs. returning visitors. Will someone who has visited your site three times in the last 24 hours have different motivations, intentions, and anxieties than someone who has just landed on your site for the first time?
Use your existing data to inform your personalization efforts. It’ll show you problem pages and the parts of your conversion funnel that need work, which is where you should begin.
Can you “go too far?”
There’s also the risk that marketers will “go too far” with personalization. Like any new(ish) UX theory, personalization has its limitations, its threats, and its drawbacks.
While personalization and machine-learning have come along way since Jakob Nielsen wrote this in 1995, the same limitations still exist…
Jakob Nielsen, Nielsen Norman Group:
“Having the computer personalize the website to the user assumes that the computer can guess the user’s needs. This is difficult to do and even more difficult when you consider that the same person may have different desires at different times. It is annoying to have the computer try to be smarter than it really is and second-guess your needs, only to have to spend extra time to correct it when it gets it wrong.
Personalization does work in a few, limited cases that are characterized by being
- very simple to describe in machine-understandable ways, and
- relatively unchanging.” (via NN/g)
Predicting behaviors, wants, and needs will always be difficult. It will never be an exact science. That’s a problem that personalization faces. What if you predict incorrectly? What if you make an assumption that you shouldn’t have?
You could end up misreading someone, eliminating information and options that the visitor desperately wants. The price of bad personalization is a bad UX.
However, you can mitigate the risk of this happening by investing heavily in qualitative and quantitative research. You have to truly understand your audience and your identified segments to create a system that thinks and acts like them.
Since personalization can go wrong, it’s important to provide self-serve alternatives…
Jakob Nielsen, Nielsen Norman Group:
“A good example is the weather forecast. 95% of the time people want the weather forecast for the area where they live (so it’s unchanging). But 5% of the time they will want the weather for some other area, so personalization does not free the designer from providing a way of choosing another location. Also, it is easy to describe the desired area: simply type in city name or ZIP code and you have uniquely defined the area you want covered. (In case of non-unique city names, the system can easily come back with a question, ‘do you mean this city or that city?,’ after which it can store a unique city descriptor in its internal database).” (via NN/g)
Finally, personalization “goes too far” when visitors begin to think “this is creepy” instead of “this is helpful.” Personalization should be incredibly intuitive, which means “what one feels to be true even without conscious reasoning; instinctive.” In other words, it should feel like a natural, useful part of the process vs. Big Brother.
Use personalization when it’s necessary and when it’s helpful. Otherwise, don’t.
Here are the three elements of effective personalization to help guide you in the right direction…
- Automatic and always learning: Always be gathering new data, always be automatically analyzing and deploying that data. Remember, intention often matters more than behavior.
- Timely and contextual: Base personalization on individual context and deliver the adjusted experience in real-time.
- Subtle and intuitive: Aim for helpful and useful, not overwhelming and creepy.
Getting started with personalization (in any industry)
Personalization can be used in any industry, so forget the myth that it only applies to ecommerce. What matters is that you set a goal, define KPIs, experiment with new and innovative personalization techniques, and continuously optimize your personalization efforts for the best results.
Here are some common personalization goals…
- Engage indecisive shoppers;
- Increase overseas sales;
- Increase first-time visitor sales;
- Get product feedback / reviews;
- Reduce cart abandonment;
- Encourage repeat purchases;
- Increase the free trial to paid customer conversion rate;
- Onboard new customers more effectively;
- Increase the conversion rate of lead generation efforts (e.g., webinar registrations, blog subscriptions).
Here are some personalization ideas to get you started…
- Recommendations. What’s trending, what goes well together, similar buyers also purchased, what you might like, etc.
- Low in stock alerts.
- Search results. What products or services should be prioritized given the visitor data and the search query?
- Exit, on-load, on-scroll popups.
- Favorite brands, styles, price points, etc. given preference and prime on-site real estate.
- Preserve your cart call-outs, count-downs, etc.
- Automatically apply coupons that are relevant to the visitor (this is delight).
- Personalize forms. Why make them enter information you already know?
- Hovering calls to action. If a visitor hovers over a price plan, indicating they’re reading, present a call to action or automatically trigger a popup that compares the features / price to your top three competitors.
Most marketers have personalized content, not a personalized UX. [Tweet it!]
It’s the shift in thinking that you need if you want to take full advantage of personalization. Marketers need to be more creative than deploying a popup when exit intention is shown or when a cart is abandoned.
Here’s a simple process you can use to excel at web personalization…
- Set your goal and choose your KPIs.
- Conduct qualitative research. How do your “power users” use your site? How do new users? Personalization can bridge the gap there.
- Conduct quantitative research. What part of your funnel are dropping out most often? Start there.
- Choose your segments and conduct additional research to understand their motivations, intentions, anxieties, habits, etc.
- Design entire UXs for each segment. Eliminate distractions, simplify navigations, update creative and copy, etc.
- Choose a personalization tool and deploy your strategy.
- Think outside the box to come up with creative hypotheses to test. Rarely will your original UXs be the best available options. Always be tweaking and optimizing for the best results.
Join the conversation
Add your comment
Great article and tons of very detailed info Shanelle. At NiftyImages we do personalized images by putting subscriber data in each image, but this helps remember that there is so much more to continue doing. A lot of what we do is based on segmenting for what feature that clients have used, getting them to try other features and going over results. We thought we were on top of it, but it looks like we need to do some touching up.
Well, I’m glad you got some value from it.
NiftyImages sounds interesting. Any case studies to share?
We don’t have any case studies posted but here are some quick examples of what can be done.
Please reach out to me at Kevin (at) NiftyImages.com and I will share everything we have.
Very cool. Thanks for the examples, Kevin!
Thanks for the detailed and resourceful article!
Happy to help, Eric. Are you doing any sort of personalization currently?
Great article Shanelle. One remark; when testing with personalisation, you def need enough conversions to test certain segments.
Thanks, Maurice. So glad you liked it.
Absolutely… that can be difficult for low traffic sites.
Stellar as always … esp. the distinctions between customization versus personalization as well as content versus experience.
I’m wondering …
Does CXL use any particular personalization tools?
I’ve personally used Evergage, but that was quite a while ago. I’m sure the tool is even more comprehensive now.
Do you have a favorite tool? I know some specialize in eCommerce, but I’ve yet to find other specializations.
This information was very excellent and easily understand all given information,thanks for sharing that valuable information.
Glad you found it helpful. Thanks for reading!
Excellent article Peep! I just have one amendment to make:
You mentioned the fact that other people are more prone to seeing “What Other Items Do Customers Buy After Viewing This Item”, because buying is much more active, giving the visitor more confidence.
We found a recent article which shows that people aim to hold a balance between conforming to other people and maintaining their uniqueness. When social proof is framed in terms of concrete actions, the motivation scale tips to uniqueness, making people more likely to buy alternative but complimentary items.
Thus framing the same statement in terms of liking instead of buying, people are more likely to buy. You can read the article here:
That’s an interesting concept, Joost. Do you have a link to the original study, too?
I think most of the examples cited here are for ecommerce websites with multiple products – in which case personalization is much more obvious. I wonder if there are cases where this has been successfully done in a stand-alone product.
When an eCommerce site is using personalization, it’s quite obvious. For lead gen or SaaS sites, it’s more subtle. You could be experiencing it without realizing it. Actually, that’s sort of the point. If you stumble across any examples, please let me know! I’d love to feature how they’re optimizing UX, too.
Hi – fab article, thank you.
Have you ever seen / had any experience of registration / booking forms employing different micro copy / layouts to increase sign ups depending on visitor behaviour / personas?
Thanks for reading, Jonathon.
I haven’t, actually. I’m not sold on the concept of rigid personas, so I’m not sure how effective that would be. However, visitor behavior would be interesting. Anyway, have you? Would love to see an example.
Comments are closed.