Examples of dark pattern notifications
Many of the sites we use every day now exploit our psychological weakness for notifications – our fear of missing out. They use pseudo-notifications to deliver marketing messages, or simply to draw us back to using their product when there is actually nothing of substance to notify us about.
Go to the LinkedIn homepage, and you’ll see a navigation bar that looks something like this:
“Hurrah,” I think, “I have 7 notifications!” (Although… Reactivate Premium? I never had Premium – another dark pattern, playing on my fear of loss).
But when I click to see these notifications, they turn out to be nothing of the sort. Instead, I’m served with pseudo-notifications encouraging me to 1) engage with changes in other people’s profiles, 2) sign up to their Premium service by promising to show me who’s been looking at my profile, and 3) browse job ads.
As a LinkedIn user, when we don’t have any new messages or contact requests, there will nevertheless be subtle forms of advertising waiting in our notification feed. By engaging with this, we spend more time on their site, click through to more pages, and complete a wider range of actions.
Facebook was one of the original engineers of the kind of notification feed we’re used to seeing today. The company has in the past couple of years also turned to using pseudo-notifications to draw people into interacting with their service more consistently, and in new ways. For example, when I arrived in Paris recently, I got a notification inviting me to see where my friends had been in the city. First off, Facebook, that’s kind of creepy. And second, it’s not something I want to be pinged about in my notifications feed.
Facebook’s notifications panel: none of these are really notifications
Similarly, Facebook will create notifications in response to how you are using their service. If you are short of “real” notifications (comments, likes, etc.) at the moment you refresh, it will use that lull to propose other forms of engagement – for example, by encouraging you to look back at Facebook “memories”, by giving you information about how many consecutive days you’ve shared content, or by telling you how many views your Pages have got.
Twitter effectively shows you other people’s notifications when you’re short of your own
Twitter uses a similar strategy, populating your notifications feed to ensure there is always something new for you to engage with.
When the service doesn’t have any direct interactions to tell you about, it will start telling you about other users’ interactions instead. In the screenshot above, it’s telling me about what people I’m following are doing on the site – a kind of meta-notification.
Twitter pushes this kind of notification into your phone’s notification bar as well, inviting renewed interaction with their app.
Engagement at any cost?
I’ve picked LinkedIn, Facebook and Twitter here as three big and prominent examples of this trend in notification design, but of course the practice is increasingly common across a whole range of sites and sectors. (Troublingly, my wine club’s website notifies me that I have 117 bottles of wine to review.)
The questions to be leveled at companies that use notifications in this way must be these: Are you acting in your users’ interests, or in your own? And if you’re acting in your own interests, are you fairly balancing these with the interests of the user? Many sites rely on clicks for ad revenue, and have discovered in notifications a quietly coercive way to generate both.
In fairness, companies are faced with the problem that these methods are very effective. Even when we, as users, understand that we are being manipulated, we still carry on clicking and tapping. And in a commercial environment, it can spell disaster to refuse on principle to deploy an effective marketing technique, when all of your competitors are doing it and reaping the rewards.
However, whether this approach to notification design remains effective will depend precisely on the evolution of our attitudes as users. We have probably all experienced notification fatigue at some point. Speaking personally, the intrusive notification practices of platforms like Twitter make me want to use those services less (but I remain problematically addicted to Facebook).
As users learn to identify and avoid pseudo-notifications, just as we learned to identify and avoid ads, notifications are likely to become less persuasive, and therefore less effective as a way of leveraging user action. And perhaps more significantly, if people’s attitudes to notifications harden, their deployment as a dark pattern will begin systematically detracting from people’s perception of a brand. Conversely, in that scenario, companies and services that adopt a simpler, more honest, and more transparent notification regime may benefit from having exactly that selling point.
Notifications and the Tech/Life Balance
If yesterday’s problem was work/life balance, then this story is about tech/life balance. The use of notifications as a dark pattern is important because it raises questions about how we regulate and control our personal use of smart technology in a world not only where it is pervasive, but also where those running key services have no qualms about bombarding us with information.
Technology has the potential to enhance our social and personal lives by keeping us connected. But notifications show that technology also has the power to impair our lives by supplanting real connections with ones that have been commercially brokered, mediated, and processed.
This touches on the question of what it means to be human in the developed world in the 21st century. We are learning together how to enjoy the benefits of our newfound connectedness, without losing touch with why we actually value those connections in the first place, and what we want technology to help us accomplish in and through our social relationships.
The future of notifications
Tech companies are not ignorant of the question of tech/life balance. As it becomes more pressing in the years to come, we can expect services like Facebook to get smarter with when, how, and how often they deploy pseudo-notifications.
I’m sure that before long, Facebook will automatically learn what kind of notifications I tend to engage with, and which I tend to ignore or dismiss. This data will enable the service automatically to send me a personalized mix of notifications based on what it has learned about my preferences. For example, if its algorithms note that I always dismiss or ignore notifications about being in a new city, a future iteration of Facebook might learn to simply stop showing me that information, or show me something else instead.
But with this comes an ethical hazard. At the moment, the pseudo-notification dark pattern is quite crude, which at least makes it easy to identify. But as services refine how their messages are chosen and delivered, it may become less and less obvious when information is being tailored on the basis of machine learning about our individual online behavior and latent preferences.
This begins to open the door to manipulation not only by fellow humans, but also by faceless algorithms that eventually learn to present each of us only with information that we already want to see — perhaps even only the news stories that are consistent with our present worldview.
As users, we should be vigilant to ensure that we retain the capacity to challenge and be challenged. As designers and developers, we must seek to develop experiences that respect users as people with agency, boundaries, and human dignity, and don’t just treat them as a vessel for clicks.