< Back to main blog page

Personalization trends and challenges: Are you doing it right?

Words by: James Flory Last updated: August 26th, 2020 16 min. read

This is the year of personalization! Just like last year, and… well, it may seem like every year is “the year of personalization”. You know, for the last five years in a row.

Personalization is still relatively new in the grand scheme of things and companies are clamoring to implement it, without necessarily understanding why they’re doing it. The result?

A whole lot of, “Hi $NAME”.

In theory, personalization is useful and impactful. In practice, it’s still falling flat for many companies.

So, how do we bridge the expectation-reality gap for everyone involved and finally make this the year of personalization for your company?

First, let’s assess where we’re at in 2019.

The Current State of Personalization

According to the 2017 Gartner Customer Experience in Marketing Survey, more than 66% of marketers responsible for customer experience believe their companies compete almost exclusively on the basis of customer experience. That number was expected to jump to 81% by 2019.

While the expectations and demands of customer experience are rising, budgets are not. 52% of marketers responsible for customer experience expected their budgets to remain the same or decrease year over year.

In Gartner’s 2018 CMO Annual Spend Survey, however, it was revealed that 14% of the average marketing budget went to personalization, which is certainly no small chunk of change.

Either way, consumers just aren’t feeling the love. According to Acquia’s State of Personalization Today, 87% of marketers feel confident that they are delivering a superior customer experience. Yet 79% of respondents feel like a generic customer rather than a unique individual when they engage with a business online.

Gartner and Acquia aren’t the only sources reporting on this apparent disconnect between expectation and reality. Segment’s 2017 State of Personalization Report found similar data:

  • 71% of consumers show some level of frustration when their buying experience is impersonal.
  • 54% of consumers expect to receive a personalized discount in less than 24 hours.
  • 40% of U.S. consumers say they have purchased something more expensive because their experience was personalized.

Yikes! So… what gives?

Let’s recap. We know that marketers think they’re providing winning customer experiences that are personalized to individual consumers. We also know that consumers don’t necessarily agree with that performance evaluation. We know that when personalization goes wrong, it goes really wrong:

Customer frustration due to personalization gone wrong

But we also know that when personalization goes right, it goes really right:


We know that personalization has a tangible impact on important metrics, like total revenue and average order value:


So, how can marketers design customer experience realities that actually meet consumer expectations? There are a few common types of personalization to choose from:

  • Behavioral: Based on measurable actions.
  • Demographic: Based on physical or situational factors.
  • Psychographic: Based on shared attitudes, needs and goals.
  • 1st Party: Based on proprietary data.
  • 3rd Party: Based on data from an entity that does not have a direct relationship with the visitor, lead or customer.
  • Temporal: Based on time and frequency.

The list is limited only by your imagination, but they all face a similar set of challenges.

The Challenges of Personalization

Entire articles could be written on each of these challenges, but here’s what you really need to know before diving into the world of personalization.

  1. Page speed is a huge factor. We can all agree that page speed is a conversion factor, right? Personalization tools, whether internal or external, often have a negative impact on site performance.
  2. Tech is questionable. While there’s no shortage of personalization tools to choose from, 74% of marketers feel that tech has made it harder to deliver personalized experiences, not easier.
  3. Data is siloed. Often, data is stored in different warehouses, managed by different teams. Bringing all of that data together in a manageable way is no small task.

Data silos and structures are struggling to catch up to personalization trends.

Right now, it’s a lot of shiny things: personalizing based on on-site behaviors and random third-party data inputs. Much of the true value of personalization lies in what companies already know (or should know) about customers, but have scattered across different systems that can’t talk to each other (either for legal or technical reasons).

Rethinking these structures can be a huge pain for companies and, as a result, they get bogged down by network lock-in and end up with bandaid solutions, like surface-level personalization strategies that really fall flat. As a result, their personalization programs fail to show value and personalization gets cut as a strategy.

It’s a dangerous cycle that can really only be broken by doing the hard strategic work before just diving in and getting started for the sake of getting started.


  1. The death of the landing page. With 1:1 personalization, the idea of a landing page becomes obsolete, really. If the landing page is different, even in small, subtle ways, for every visitor, how can marketers optimize that landing page? Of course, it’s still possible, but the proverbial death of the landing page does complicate experimentation.
  2. Real-time. Take behavioral personalization, for example. Every behavior should, in theory, have an effect. It’s a new insight, which can be used either in isolation or with a group of similar insights, to provide a personalized experience. Delivering personalization in true real-time is very difficult, especially when you might receive 10+ behavioral insights in the span of just a few minutes.
  3. Privacy laws. First, GDPR rolled around. Now, CCPA is just around the corner. As more and more ambiguous privacy laws are written into existence, the way marketers can and cannot collect and use data becomes more ambiguous.
  4. Extending beyond landing pages and page elements to entire customer journeys. Personalizing the heading on a landing page based on the Facebook ad campaign the visitor clicked is one thing. It’s an entirely different thing to personalize entire customer journeys from start to finish. Creating small moments of delight can be simple. Crafting entire customer experiences that are personalized for each individual? Not so much.
  5. Cross-device and cross-browser usage. Let’s say you’re browsing anonymously on a mobile device. You’re looking at gym clothes: sweatpants, headbands, socks, lifting shoes. You start receiving personalized product recommendations based on this behavior. But now you’re running late for a meeting, so you make a note to come back later. The problem? When you fire up the same website on your desktop at home, all of those personalized elements are⁠—poof⁠—gone. Unless you have sophisticated technology and a robust data structure, personalizing cross-device and cross-browser is very difficult for sites that don’t rely on account-based selling.
  6. Balancing helpful and creepy. This is an extremely fine line. Do I secretly want to be reminded of those lifting shoes I’m coveting? Yes. Do I want to know you know how much is in my 401K and my mother’s maiden name? Not so much. Ask yourself how the personalization makes the customer experience better. The goal of personalization is not to simply show that you can, the goal is to remove friction.
  7. Segmentation is not a straightforward concept. Let’s say you decide to segment based on business type. What if that doesn’t have a causal connection to conversion? What if the same person fits into multiple segments? How do you assign them? How do you know which of the segments is most impactful for them to be in? Truthfully, segmentation is not an exact science. Sometimes it’s a success story, sometimes it hurts your marketing.

Behavioral science is perhaps the best way to segment. Normally with segmentation, you only know the what. You know that Sarah is a 27-year-old product manager from Seattle who recently purchased a spaghetti strainer and chocolate chip cookies from her 2017 iPad. But you don’t know why. With behavioral science, you can understand emotional states and contexts to uncover that why.

Through a combination of color, emotionally resonant copy, multimedia storytelling and usability, you can create an emotional first impression through customer experience design. As you conduct conversion research and run experiments, you will begin to collect quantitative and qualitative data that will inform your understanding of relevant emotional states and contexts.

Knowing who your customers are is one thing. Knowing why they do what they do is another. Segmenting based on the latter instead of the former is a game changer.

Finally, segmentation can complicate experimentation. Let’s say you’ve run a test and the results are in. But you want to know how the test performed for various segments. Maybe you’ll uncover a hidden insight or two! Maybe a test that lost across all segments wins big in a specific segment or two, for example. This is known as post-test segmentation. While it’s very common and not inherently a bad idea, it is complex and can be dangerous for those without a strong understanding of statistics.

You calculated your sample size upfront, right? Great. But once you begin to segment, the sample size becomes smaller. If you stopped the test when you reached the originally calculated sample size, your segmented sample sizes are likely too small to analyze with any degree of confidence.

Biases also sneak into post-test segmentation. For example, the more segments you compare and analyze, the higher the chance you’ll uncover a false positive. With even a 95% confidence level, it’s likely that 1 in every 20 post-test segments will be a false positive.

Segmentation is the backbone of personalization, but it’s simply not as black and white as most marketers believe. There are nuances, complexities that are rarely acknowledged, let alone accounted for.

Personalization in Ecommerce, SaaS and More

So, the question remains: how do you get personalization right?

First, there are two questions you absolutely must answer if you want to be effective with personalization:

  1. How does personalizing this page element reduce friction and improve the experience for the visitor, lead, or customer?
  2. Which business goal does personalizing this page element serve? Which metric will this change directly impact?

If you feel like you’re in a good place after that self-reflection, you want to make sure you’re actually ready for personalization. Much like testing is not the best use of every company’s time, depending on their state and maturity, not every company should invest heavily in personalization right now.

Ask yourself:

  • How much data do I have on my visitors, leads and customers? How confident am I in the integrity of that data?
  • Do I have a personalization tool that has a limited impact on web performance and works reliably with my data structure?
  • Do I have someone who can manage the overall personalization strategy and day-to-day implementation?
  • Do I have someone who can manage the experimentation side of my personalization program and optimize it consistently? Do they understand how personalization will impact the experimentation program?
  • Do I have the resources to maintain segment integrity as we get more and more granular with our personalization?
  • Do I have confidence in our ability to validate our personalization hypotheses and implementations?
  • Is our data structure siloed? Is it sophisticated enough to collect and communicate cross-device and cross-browser?

So, what does personalization look like in various industries? Impactful personalization is rarely addressing someone by name in an email. Instead, it looks like:

  • Product recommendations based on what you’ve viewed or purchased.
  • A dynamic call to action in the right-rail of a blog based on the keyword you searched for.
  • Promotion emails timed based on your personal purchase frequency.
  • In-store availability flags based on your shoe size.
  • On-site ads based on your actual personal interests.
  • Not seeing stock that can’t be shipped to your geographic location.
  • Seeing the people you interact with most at the top of your list when you go to share a meme on Instagram.
  • A recommended plan based on the number of employees who work at your company or its annual revenue.
  • Shipping prices automatically calculated based on your zip code.

You get the idea. Personalization is a creative exercise, but be sure it’s grounded in tangible business goals.

Of course, we’re all familiar with some of the ways Amazon does personalization. You’ll see recommendations based on the types of products you buy most often. For me, that’s books:

Maybe you’ll see what’s hot in the categories you purchase from often:

But it’s not just the Amazons of the world doing personalization well in ecommerce. Kylie Cosmetics has complementary product recommendations right on product pages:

Chubbies estimates your shipping based on your geographic location right on the product page:

In SaaS, sometimes personalization is as simple as a chat bot, like Drift, acknowledging that you’re a repeat visitor:

Other times, it’s recommendations based on interests and previous behavior (thanks, Spotify):

You’re also familiar with personalization on popular listing sites. For example, Airbnb will recommend cities based on your interests and your average historical price per night:

Kayak knows where you’re based geographically and will recommend popular destinations based on where your “neighbors” are traveling:

If you have an account, you’ll even get a personalized booking recommendation based on your preferences and upcoming travel plans:

Condos.ca will not only show you similar listings, but listings that are in the same neighborhood as the listing you’re currently viewing as well:

Personalization is possible in every industry, but these are a few you might be especially familiar with. Notice what they all have in common: they don’t personalize for the sake of personalization. Instead, they personalize to improve the customer experience in a way that fuels their business goals. It’s not “Happy Monday, Sarah!”, it’s empowering the visitor, lead or customer to take action.


We know expectation, we know reality, but what’s next for personalization?

The near future of personalization will look a whole lot like the current state of personalization while companies continue to lay the foundation and develop meaningful strategies. It will take at least a few more years before most companies have progressed beyond implementing what’s “new” and “shiny” in personalization.

Subscribe to get experimentation insights straight to your inbox.