If there’s one thing that’s accepted as truth in the customer experience world, it’s that CX personalization is a must-do. From emails that address customers by name to relevant product recommendations based on behavioral predictions, personalization can bring huge value to both customers and companies alike when used wisely.
We’ve seen the power
Thankfully, CX personalization has been tested enough by now that the benefits are crystal clear.
For instance, customers seem to want it. Surveys show that 80 percent of consumers are more likely to make a purchase when brands offer personalized experiences, and 72 percent of consumers will only engage with personalized messaging (e.g. personalized ads and emails).
Brands, too, have seen real business outcomes from personalization efforts. For example, marketers that personalized email subject lines experienced 27 percent higher unique click rates and more than double the transaction rate compared with non-personalized emails.
This trend won’t go away. Personalization has made shopping more pleasant for customers on a number of levels. They may feel they get more personal service. Or it helps them avoid option paralysis by narrowing down the often bewildering array of products available.
Whatever the reason, personalization makes sense: retail giants like Amazon and entertainment companies like Netflix have even built recommendation algorithms into their very value proposition.
But companies don’t have carte blanche for personalization
Even though customers are keen on personalization, they haven’t given companies the right to do as they please. As concerns about data privacy increase, and shoppers become better informed, personalization requires a more careful approach than ever before.
For example, we may wonder, is there such a thing as too much personalization?
As it turns out, yes there is.
What is too much personalization, Michael?
Did we just say Michael? Yes — but don’t worry, we haven’t gone mad. We’re just making a point; if your name were (or actually is) Michael, then this heading might give you pause. “Wait, is this blog personalized? How do they know my name?”
We can draw a parallel with ‘why’ CX personalization doesn’t always work. If it’s random, sudden, or invasive, then it’s not effective. It might even cost you customers.
For instance, more than 50 percent of customers will unsubscribe and 38 percent will stop doing business with a company if they find its personalization efforts creepy.
So, to help you personalize — while also remaining on your customers’ good side — we looked into statistics, anecdotes, and research to find out what’s considered ‘creepy’ and what’s considered ‘cool’.
The Jeepers Creepers: Examples of bad CX personalization
Who’d have thought personalization would provide such rich pickings for Halloween-themed storytelling? And yet…
The stalker 👿
We’re talking about knowing where your customers are and taking advantage of that to offer recommendations or targeted ads. Imagine, for example, using geofencing to trigger ads when people cross a virtual boundary.
That’s a type of intrusive personalization that many people find unacceptable: 40 percent of customers have stated that receiving notifications whenever they pass near a specific location (e.g a physical store) is the creepiest personalization technique around.
On the other hand, 87 percent of marketers that personalized their ads with location data reported more positive customer experiences as a benefit. It seems like a contradiction, but it probably only means there are ways to do it right. Just make sure you’re not keeping customers in the dark about how their data is used.
The omnipresent 👀
Consumers are mostly ok with you knowing their purchase history and some other basic information. But, medical history? Intimate secrets? Information about friends or family? There’s a horror film right there.
According to research, only about 2 in 5 people are comfortable with artificial intelligence accessing their employment or family information, and only one third are comfortable with the technology knowing their health or financial information.
That goes for social media targeting, too — you know, when you’re thinking about buying a robot vacuum cleaner and then suddenly start getting bombarded by ads about it on Facebook or Instagram. Consumers seem to be on the fence about this one: 50 percent of US consumers think it’s more creepy than helpful when companies use social data to make suggestions (interestingly enough, 32 percent say it’s both creepy and helpful, in equal measure).
When apps seem to “listen in” on conversations as well, everything gets a whole lot creepier (though this might not be literally ‘listening’ through microphones, but rather tracking and analyzing online behavior and connections).
One thing is for certain: monitoring a customer’s every move in the hopes you’ll deliver helpful messaging is both inefficient and inappropriate. You need to make sure you’re providing real value to justify using this data for CX personalization.
The fail 👎
Bad personalization is worse than no personalization. When was the last time you respected a brand that called you “Dear [First name]”? These kinds of mistakes only serve to remind consumers that your algorithms don’t see them as human, but just data points being handled by a machine.
Algorithms make other mistakes, too. You’re looking for a graduation gift for your brother, and suddenly you get ads for colleges and universities. At times, these kinds of oversights can stir up real negative feelings. Consider the case when women received congratulations for non-existent weddings. Or a card printing company congratulating women about their babies — when many of them were actually dealing with fertility issues.
Mistakes do happen, of course, and how you deal with them matters, but it will take substantial effort to turn the tables after a personalization disaster of that magnitude.
The persistent 📣
It happens all the time: you buy an item online and then you keep getting promotions or ads for that same item. This tactic may work for consumables or products like shoes and clothes, but it’s completely ineffective for things like laptops, sofas, or toilet seats.
This probably indicates a problem with how data is analyzed. Not all products can be treated the same, and there needs to be a way to differentiate messaging. For example, if a person buys a laptop, an effective recommendation would be laptop cases. In other words, relevance by design.
The Cool Kids: Examples of good CX personalization
Despite the horror stories, personal customer service often works. And that’s why so many people want it as part of their buyer journeys. Here are some examples to get inspired from.
The curated 🎁
B2C technology companies knock it out the park with this type of personalization. Spotify’s yearly reports on what music each user listened to, Netflix’s “Because you watched X”, and so on.
Even retail or ecommerce companies can use this type of curated content. All you need is access to purchase or browsing history to be able to send relevant blogs, videos, and other useful content.
The holistic 🌐
What separates the persistent personalization from the holistic? When you want to do holistic personalization, you send recommendations for products that go hand-in-hand with what the customer has just bought.
Think Amazon’s “frequently bought together” recommendations. Or emails recommending laptop cases for laptop buyers, new furniture for newlyweds, and so on. This requires more complete and accurate customer profiles, instead of just inferring preferences from incomplete data.
The individualized 👤
Most personalization tactics happen automatically and serve large numbers of customers. But, although it’s difficult to scale, there’s no better way to make service more personable than by talking to people individually.
For example, 45 percent of consumers indicated that receiving an apology email after a poor in-store experience is a cool engagement tactic. This may be because this type of communication is unique and individualized, and shows the company actually cares about the customer’s experiences.
This also includes personalized service over various channels — for instance, a customer support agent knows a buyer has had a problem with their order before they reached out again, so they can tailor their response properly. Social listening and responding is also a great opportunity for individualized engagement.
The customized ✂
An excellent example of this is Coca Cola’s Share a Coke campaign that went live a few years ago. The company didn’t need any complex data or advanced algorithms. They just added names and song lyrics on their products, enabling consumers to choose their own personal experience.
This is a lesson for every consumer industry — you don’t always have to target specific people, but you can find ways to help them see themselves in your offerings.
The crux of the matter is trust
Consumers are much more likely to share their data with companies they trust. To get there, you need to focus on nurturing long-term customer relationships. Here are some things to pay attention to when it comes to establishing trust for personalization:
- Be transparent. It doesn’t take much to tell customers that you’re collecting data and how you’re using them. A website banner, a privacy notice, a short explanation, and other options can make it easier for customers to accept your use of their data. Give them full control over their data, too — if they want you to stop using it, you need to comply (and you must if you’re covered by privacy laws like GDPR).
- Be careful with sensitive information. People generally don’t object to you having a basic profile about who they are and their preferences. But, unless it’s extremely relevant to your business to know sensitive data such as religion, sexual orientation, or medical history (and let’s face it, chances are it’s not relevant at all), avoid using it for personalization.
- Just don’t be sales-y. Trying hard to sell is the most sure-fire way of crossing the line into creepiness. Aim to be helpful instead — see personalization as a way to provide customers with the information they need to make the best decisions.
- Make sure your algorithms work. The more your software or AI learns, the better the recommendations will get. Ensure your team takes care of the basics (never call someone “First Name”) and try to refine algorithms. If you get it wrong, apologize and do better next time.
What kinds of customer experience personalization do you use? What do you find creepy? Or cool?
Let us know in the comments.