Why Personalization Should Be Invisible To Your Users

by Richard Bayston

Last updated on July 25th, 2017

Personalization seems like the logical endpoint of data-driven marketing. If we can present our users with an experience that’s tailored to their interests, that should be better for them and better for us. But personalization can come with unexpected costs that eat into its efficacy, and it can even turn customers away and tarnish your brand image.

There’s a tendency to view increasingly accurately targeted messaging as a good: personal good, broadcast bad. We need to look beyond that, and we should start with the state of personalization now, what marketers think of it, and most crucially of all, what consumers think of it.

Right now, personalization is at a tipping point. A majority of marketers would like to be succeeding at it, but few are. It’s already being implemented by large, tech-savvy companies. In a CEB survey earlier this year, the average company said it was using five personalization technologies. Companies with over $1bn in revenue were usually using six. The most common tools in use were web analytics software (71%), CRM (64%) and CMS (56%).

current personalization technologies


But even among companies with deep pockets and strong attachment to marketing tech, personalization isn’t a mature, fully-functioning system.

  • 78% of CEB’s respondents say they need more tools to carry out personalization effectively
  • 14% said they felt their personalization efforts were mature and integrated
  • 59% said they were unsatisfied with their current personalization technologies
  • Only a third of respondents had a technology roadmap in place

In other words, the same process is playing out that we’ve seen in the adoption of technologies like CRM, or strategies like content marketing. If a business is large and well organized, there’s a solid chance it will be getting some value from personalization.

Many other companies are working towards it but it’s not really paying off for them. And a large minority of businesses haven’t invested any serious effort into personalization – yet.

most common personalization techniques


More than half of CEB’s respondents said they were spending 1% to 10% of their marketing budgets on marketing personalization tools, but 74% said returns so far were marginal.

What’s causing this disconnect?

The Data Disconnect

We’re collecting data from users on websites, on ads, on email responses, and in a whole bunch of other places. When you ask marketers ‘do you collect demographic data on your customers?’ 76% say yes.

personalization data collected


Ask the same question about whether they’re actually using that data to offer improved personalization anywhere across the brand experience and the responses are different.

personalization data use


Demographic data is the most-collected type of data – yet it’s used in marketing only 47% of the time – and in real-time just 8% of the time.

One reason many businesses aren’t getting much return on their personalization efforts is that they’re not actually doing much with the data they’re collecting.

Consumers see this as a big deal. Research by the Guardian Media Group shows that 49% of consumers think brands aren’t clear about what data they’re collecting from them. They’re certainly not clear about what brands use it for – in many cases, we hardly know ourselves.

The Consumer-Marketer Disconnect

What do customers think of brands’ personalization efforts?

They’re unimpressed. While 71% wanted personalized offers, if nothing else, 20% felt that personalized offers weren’t well done – and 20% felt that brands’ personalization efforts were too intrusive.

Brands deliver adequate personalization of loyalty programs only about 22% of the time – and loyalty programs are one of the easiest things to personalize. Other brand experiences are even less satisfactory for consumers.

Yet, consumers want personalization. 57% agreed with the statement: ‘I expect companies I do business with to understand my needs and preferences.’

And the younger they are, the more they want it. 70+% of ‘digital native’ consumers demand personalization of websites and other real-time experiences. Among older consumers that number falls: 38% of those aged 50-64 and 62% of those aged 65+ say they don’t think ads should be personalized in any way.

Being served content that they’re uninterested in is one of the biggest irritants reported by consumers.

yucky personalization


Consumers are eager for convenience and relevance: relentless targeting isn’t interesting to them. That’s a brand priority, not a consumer priority.

personalization and shopping


Consumers don’t want to be chased, tagged and targeted: but they do want convenience and relevance, especially in real-time and email.

The Personalization-Privacy Disconnect: Where Uncanny Meets Ineffective

Customers want personalization and brands can’t deliver it, but they don’t want to give up their data – and brands can’t stop collecting it. On the one hand, consumers want to choke off the flow of information they’re giving brands. On the other hand, brands aren’t even using the information they already get from users, yet they’re desperate to offer personalization so they can reap the benefits.

This is what Naveen Awad and MS Krishnan call ‘The personalization privacy paradox’:

Consumers want personalization but not at the expense of surrendering their data.

That’s partly because of rationally-rooted cybersecurity concerns and partly out of an emotional reaction against a level of data gathering that they just find creepy. In a study in 2013, the Guardian Media Group compared the effect to roboticist Masahiro Mori’s famous ‘uncanny valley,’ the creepy territory between definitely-a-robot and definitely-a-human that makes people uncomfortable around over-realistic imitations.

The effect is weakest where personalization has immediate consumer benefits, or when opt-in is clear and data collection and usage is expected. These effects can come together, as with agile emails with highly personalized content, which deliver bumps in open and conversion rates.

It’s strongest off-site, where 40% of adults questioned by Pew don’t even think their search engine or ISP should keep information about their online activities – let alone brands using that data for retargeting. And 28% of US adults in the same survey had sought to conceal or falsify their information online to avoid advertisers.

who users try to avoid


Asked how they felt about ‘online ads that use details about what I have done’ and encouraged to use their own words, the 400 consumers questioned by CEB delivered solid evidence of this ‘uncanny’ effect: 73% expressed negative sentiments, including words like ‘hunted,’ ‘naked,’ ‘unsafe,’ and ‘worried.’

creep factor


The most common response marketers expected was ‘valued’ (62%). The most common response consumers actually had was ‘creeped out’ (49%).

So when Colin Shaw of UX blog BeyondPhilosophy says: ‘Personalization is critical to Customer Experience because of how it makes your customers feel,’ it looks like he’s right – but maybe not in the way he means.

Excessive, intrusive personalization doesn’t make customers feel warm and fuzzy (and inclined to buy) – it makes them feel like Sarah Connor, stalked by a relentless, terrifyingly creepy machine.

If that’s you, cool.

But what if you want your customers to form lasting attachments to your brand, eagerly buy your products and evangelize for your brand experience, and not, say, melt you in a big vat of steel and flee across the border?

In principle, the solution is simple.

Highway Rules: Pick a Lane

Too often, brands that don’t have a defined strategy for personalization slip into that uncanny valley between two strategies, each with clear benefits. There’s personalization that’s barely detectable and feels like it’s just convenience, and personalization that’s being offered as a benefit right out to the consumer. But many brands lack strategic overview of their personalization efforts, so they don’t select an approach.

The result is the situation we’ve just looked at: intrusive data gathering, and clumsily obvious tracking and targeting that freaks consumers out without actually delivering on ROI or revenue.

Walk down the middle of the road, get hit by traffic both ways. Plump for invisible or flashing light personalization.

Invisible Personalization

Invisible personalization should be like good UX: undetectable because it works. From the consumer’s point of view, this kind of personalization becomes visible only when it fails.

Emails announcing the start of deer hunting season to 12-year-olds whose previous interests include Bambi, or bombarding male subscribers with advice on how to ‘rock the wedge’ this summer’ are jarring instances of invisible personalization becoming visible through failure.

When that happens, as we’ve seen, consumers hate it. The rest of the time, they expect personalization to just happen – so that’s what brands should do.

Rather than having high hopes of using a huge wealth of data to deliver ‘1:1 personalization,’ creating categories of just a few or even on customer, it makes more sense to approach invisible personalization as an aspect of UX and focus on customer preferences, relevance and convenience.

As Revenue River Marketing’s Mark Herschberger told Hubspot:

‘It’s all about creating a better user experience for your site’s visitors and providing relevant information to the right people at the right time, not showing off your ability to track every move someone has taken on your site.’

Data collection for invisible personalization

Data collection for invisible personalization should be as unobtrusive and normative as possible.

You should seek to collect the data everyone collects: consumers tend not to mind basic data being collected but do object to retargeted ads following them from site to site, for instance. So start with norms, then see where your customers’ tolerance lies.

The three biggest sources for data acquisition for this type of personalization are your website, your point of sale and/or CRM, and your email data.

sources used to find first party data


Implementing invisible personalization

Invisible personalization can be implemented without hitting that creepiness factor by utilizing modeling. Models are essentially a data-based equivalent to personas. Rather than being invented to personalize customers about whom we have too little data, though, models are there to allow marketers to segment according to data without getting so personal that it’s you know, creepy.

long term high value buyers


Models can be arrived at and marketed to as though they’re real people – without ever making people feel like you’re talking to them, and them alone, without being invited to.

Who’s Acing Invisible Personalization?

Check out the way Zappos does it. They’re an Amazon subsidiary (bought in 2009) that took revenue from $1.6m in 2000 to over $1bn now, and they’re doing it with textbook invisible personalization.

zappo personalization


Select an item from their range, like I did above, and you’ll see ‘customers who viewed this also viewed…’ down there in the bottom right hand corner.

Scroll down the page…

zappo personalization 2


And half the real estate is taken up with ‘you might also like’ and ‘customers also viewed.’ A decent part of the rest is given over to information gathering and feedback.

Do customers like this? Does it creep them out? Do they even notice?

UX Booth’s Maya Nix is probably more prone to spotting this kind of thing than most people but her review of her Zappos experience is positive. ‘My order doubled thanks to successful personalization,’ she says of a recent visit. Yet that personalization didn’t rely on Zappos knowing her date of birth, where she went to college or any other personal information – the kind of thing people find creepy. They didn’t have to follow her from site to site either. All they had to do was connect the dots.

Planning to buy a present for her six year old niece, says Nix, ‘I’d been window shopping on Zappos for months.’ On the day she made her purchase, ‘I was greeted by kids-focused shopping recommendations!’ The result? ‘Not only did I buy the Converse sneakers I originally sought out, but also the denim shorts that Zappos recommended.’

That’s invisible personalization being aced: in a totally unintrusive, passive way that let the customer lead, Zappos made a recommendation that doubled Nix’s order value.

Netflix is another example of solid invisible personalization. The basic Netflix experience all personalization, all the time: the company ‘[uses] personalization extensively and treat every situation as an opportunity to present the right content to each of our over 57 million members.’

In other words:

netflix personalization


Netflix uses a mix of open data gathering, usage reports and complex algorithms to arrive at home pages that are highly personalized for families or subscriber groups.

stages of invisible personalization


As Xavier Amatriain phrased it in a blog post in 2012, ‘Another important element in Netflix’ personalization is awareness. We want members to be aware of how we are adapting to their tastes. This not only promotes trust in the system, but encourages members to give feedback that will result in better recommendations. A different way of promoting trust with the personalization component is to provide explanations as to why we decide to recommend a given movie or show.’

That doesn’t detract from Netflix’ status as monsters of invisible personalization, though, because that’s a background awareness: everyone knows Netflix uses your data to show you other movies and shows you might want. No-one really thinks about it too much. Which is exactly what you want.

Efficacy and KPIs: Measuring Success

It’s vital to know what success actually looks like, in personalization as in everything else. Pageviews and the million other things Google Analytics will measure and crunch for you aren’t proof that your personalization campaign is actually working.

Marketers who aren’t employing personalization yet, but plan to, say they have three main goals

Increased engagement was a main goal for 78% of respondents, and improved customer experience tied with 78% of respondents naming it a main goal. Increased lead generation was a main goal for 60% of respondents.

1. Increased engagement (78%)

If this is what you’re shooting for, you’d want to be measuring KPIs like bounce rate, time on site, and average pages per visit. To determine whether your personalization efforts have a strong effect on long-term engagement, track number of returning visitors.

If you’re one of the 22% of marketers using personalization who have deployed it on your mobile app, you’ll want to measure engagement in terms of:

  • Session length
  • Usage frequency
  • Repeat usage
  • Screen exit rate (Are personalized screens seeing higher or lower exit numbers?)

2. Improved customer experiences (78%)

Tracking customer experience in numbers? Not so easy. But doable. The standard engagement metrics should be showing you something about customer experience (CX), but to get a real bead on things you need more specialized equipment. Google Analytics provides it, in the form of the User Flow function, which lets you see relative volumes of users moving around your site, according to the dimensions you specify. This should let you track user behavior at specific points in your flows and identify where CX is strong or weak.

You should also be using funnels to test CX. Funnels let you map specific actions, called Goals, and see user behavior on your site in those terms. They’ll let you see where your process is shedding customers, and where the conversion bottlenecks are.

3. Increased lead generation (60%)

If this is your main goal from personalization, as it is for many marketers, you’ll need to measure it by plugging the data you get from Google Analytics about website behavior into your CRM. While many CRMs offer a solution that will do some of this for you, with varying degrees of comprehensiveness and success, the best method is to tie Google Analytics right into your CRM data. You can tie the Google Analytics tracking cookie to your lead and pass on that data automatically by reading the Google Analytics cookie and passing that information on as hidden fields in your lead capture form.

To determine the effects of your personalization efforts on customers’ attitudes, brief single-question popups can generate higher response rates and give you a wider pool of data on what customers like and don’t like, without having to wait for industry-wide reports or extrapolate everything from Google Analytics.

Flashing Light Personalization

Flashing light personalization doesn’t seek to be invisible: it seeks to be an attraction.

This is the exception, to be sure, but some businesses will have success by advertising a more personalized experience as a special service and soliciting extra personal information in order to provide it.

The key is to be totally up front: ask for specific types of information and explain to consumers exactly what you plan to do with it. Consumers are often willing to give up additional personal information in exchange for a better user experience, including a more relevant and personalized one, as long as the brand concerned is totally upfront about data collection and use.

In a study spanning five nations and including 8,000 respondents, Columbia Business School’s Matthew Quint and David Rogers found that people were around 80% more likely to share data with a company they trusted.

brand trust strongly influences


They found another lever that makes consumers more willing to hand over specific pieces of data in return for benefits: offers. These broke down across two axes: experiential and financial.

strategic map of offer influence


We should be seeking to combine these two effects by preserving and enhancing brand trust at every opportunity, and by offering highly effective benefits in return for data that lets us personalize experiences to a high degree, in a highly visible way.

What does this look like in practice?

Target is a good example. Based on the realization that still haunts JC Penney – that ‘our core customer… enjoyed coupons more than [CEO Ron Johnson] understood’ – Target’s mobile app is based on couponing.

cartwheel mobile app personalization


That’s smart because it’s offering customers something tangible up front in exchange for their data. Other services that do this can also get away with high levels of data gathering – think Google, or even Pokemon Go, whose constant experiential feedback justifies data acquisition in the eyes of very large user bases in both cases.

More importantly, Target is taking a service that’s separate from the basic shopping or website experience but is still real-time, and personalizing it for individuals for their benefit. A new update to the app takes users directly to the personalized For You tab on startup. It’s a coupon news feed.

It’s paying off for Target: ‘Our efforts to make Cartwheel more personalized have contributed to the great success of the app, which now has more than 25 million users and has generated more than $3 billion in promotional sales for Target,’ says spokesperson Eddie Baeb.

Taking the idea of flashing light personalization to its logical extreme are boutique brands like Trunk Club.

Trunk Club Boutique Personalization


Trunk Club offers personalization not as a service, but as its service: the personalization is the product. It offers personalized stylists who take the place of algorithms and data, building its customers their ideal wardrobes based on an understanding of their sense of style.

trunk club personalization



Personalization is clearly the future of the brand-consumer relationship. (In fact it’s already here; in William Gibson’s words, it’s just not evenly distributed.) But it’s not going to be a future brands build and consumers use. It’s going to have to be negotiated. For brands that want to offer super-tight real-time personalization, they’re going to have to ask their customers to do that with them as a partnership. If not, they risk driving those customers away. Similarly, though, brands that offer weak basic personalization, so that their sidebar offers, pop-ups and website experience feels like it’s optimized for someone else, will shed customers too.

The solution has to be to signpost data acquisition and usage that departs from the norm and own it: put flashing lights on it and tell consumers why it’s such a good thing. For background personalization, we shouldn’t have to tell consumers anything because it’s what they expect: it should remain totally invisible.

*Featured Image Source

No Comments


Get updates on new articles, webinars and other opportunities:

Richard Bayston

Richard Bayston creates copy and content at RBCopywriting.com, mostly about tech, digital marketing and content strategy. When he’s not doing that, he’s arguing amicably with his wife and Googling the answers.


Comment Policy

Please join the conversation! We like long and thoughtful communication.
Abrupt comments and gibberish will not be approved. Please, only use your real name, not your business name or keywords. We rarely allow links in your comment.
Finally, please use your favorite personal social media profile for the website field.


Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Show Me My Heatmap

literally addicted to the stuff I'm seeing in @CrazyEgg. Such a fantastic piece of software! #Marketing #Content #Digital

Chris Vella-Bone


What makes people leave your website?