Please Kill These Features
An open letter to Microsoft and anyone else thinking about the next cool feature in their email app
Dear Microsoft,
I know you have the best intentions. Really, I do. I know you want to shape the way people in organizations work. A significant part of that is changing the way we communicate. As a company that develops and sells products, the obvious way to do that seems to be by adding new features and capabilities to the tools many of us use. That’s how most companies interpret innovation: add more capabilities. The alternative, it seems, is to stand in place, and when you don’t move forward, you soon find yourself behind the rest.
But we, your users, just want to communicate. We need to communicate. We rely on communication to do our work, and adding more features to what should have been a simple communication tool is not always the best idea. When it comes to innovation, more is not always better. Sometimes, innovation means fewer features and options. Innovation always means considering the consequences of what you create beyond whether the feature is adopted. When it comes to something as essential and delicate as communication, adding new features is rarely the solution; often, it is the problem.
Once a feature is out there, deployed to millions of users, what you had in mind when you designed it doesn’t really matter anymore; it gets a life of its own. You might have thought you are developing a killer feature, while in reality, your feature is killing your users’ productivity. Sometimes, the results are unexpected, but often, they are easy to predict because people are strange creatures but not entirely unpredictable.
Most people, for example, will use defaults and rarely consider an alternative. That is why we have defaults in the first place. That’s why many applications are preinstalled on our devices, and any application has numerous defaults embedded in its design. Whatever option we decide to set as default will likely be used the most, and most people will use it without giving it a second thought. You need to be extra careful when designing your defaults. One default, in particular, turns our workplace communication into a full-time job and not a productive one.
People also like it when the mess they create is cleaned up after them. Why bother perfecting our work if, in the end, someone or, even better, something will clean up after us? Why bother refining? Why bother even thinking if something can take whatever we dump on the page and turn it into neat and polished text? Of course, when this is the case, we are not really communicating.
People are creatures of habit; when we are used to doing something, we tend to keep doing it. It takes quite an effort to break our habits, and we need a good reason to do so. So, when a new feature takes advantage of our bad communication habits and brings them into the office environment, you can be sure most people will use it, even if it is evident that it makes our communication shallower.
People are strange. Predictably strange. And three features in Outlook make our communication predictably worse as a result.
Reply Who?
I am not a User Interface (UI) expert, but I know this much: defaults matter. Most users will not change the defaults pre-set in their apps (assuming this is even possible); most users will not bother looking for a better option before clicking the first button they see in front of them. At some point, we all become blind to the possibility that there are better options.
So, when you decide to set the Reply All button as the first (and the only visible) option when replying to an email, you can be sure most of your users will not think twice and use it without giving it any thought. And why shouldn’t they? Reply All seems like a good idea. If email is used for small-scale discussions with three or four people, it is not common that you will want to remove some of them from the thread at some point. It might even be considered rude. But many work-related emails are sent to distribution lists of dozens of people. Whether this is effective or not in the first place is a good question, but as long as this is the case, the Reply All option becomes a productivity killer instead of a productivity hack.
Let’s do the math. It starts with an innocent email sent to, let’s say, 20 people (which is in no way an exaggerated or rare number). If only five of them decide to reply with a statement, a follow-up question, or just a “Well done!”, they are likely to do so using the default reply option: Reply All. The result is 120 distractions across the organization caused just by a single email thread. The common estimate for the cost of such a distraction in terms of the time it takes to gain back focus and effectively continue your work is 20 minutes. This single email thread with one message and a modest number of five replies just caused your organization to lose 40 work hours. That’s a whole work week, and before you contemplate this number, a new thread with a similar or larger distribution is already showing up in your inbox.
We are so used to Reply All that we actually believe it is the only option. In reality, many of the replies we absentmindedly share with everyone could have been sent to one person or a limited number of people. Reply All could be a viable option, but when it is the default option, it dictates our default behavior. One way to overcome this is to move the entire distribution of the original email to BBCinstead of using the To field. Alas, this option is not the default in Outlook; it is actually hidden when you create a new email, and you should actively look for and enable it. And this is yet another poor UI choice if you intend to help people communicate more effectively.
A simple design choice would make all the difference and save hours of work every week for every user of your product. Just hide the damn button. How about asking the user the most infamous question in the history of UI: “Are you sure?” before the reply is actually sent to dozens of people? And if you want to be innovative, how about showing us how much our reply will cost in terms of lost focus across the organization before we send it? All these alternatives add friction, which sounds like a bad idea when it comes to usability. But friction is a great idea when introducing an option that should not be used much. Being intentional about replying to dozens of people is much better than doing so just because it is the first option we see.
The Garbage Disposal
When your inbox is populated with numerous threads, each with a conversation between too many people (in many cases because of the Reply All default), a feature that helps you summarize the thread and make sense of it seems like another killer feature. Email threads are impossible to follow. The minute they cross the threshold of a simple question and reply and the more people are involved in the “conversation,” the thread turns into a cacophony of colors, indentations, quotes, and back-references that are impossible to decipher in reasonable effort.
Enters Microsoft’s Copilot, the newest addition to the Office Suite. The premise is simple and powerful: You and your colleagues can write whatever you want in a back-and-forth email thread, and Copilot will summarize it. It seems helpful for any long thread; it's priceless when everyone uses Reply All by default. Wouldn’t that solve our workplace communication problems? Well, not really.
Effective communication must be meaningful and intentional. You must think before writing and write your ideas well to communicate effectively. Whatever you write should have value, and it should be concise. When you reach the point where you need an AI assistant to be able to understand what the conversation is about, you are no longer in the realm of effective communication; you are now in a war zone, trying to minimize damages. We should write for humans. It is not easy. We need to relearn effective writing, but the latest additions to our communication tools let us get away with just about any poor, too-long, low-value text we can come up with. Someone will clean up after us.
Using Copilot is easier than writing well or moving the conversation to a better platform. We don’t need to think whether our reply adds value to the conversation; we don’t have to consider how to phrase it so it will not be a burden to read; we don’t even have to decide what it is we are trying to say. Instead, we can write our stream of consciousness and let the bot do the work for us. It seems more effective but degrades our ability to communicate and connect. It is a killer feature — killing our interactions.
There is no reason to avoid using AI; you can undoubtedly embed it better in Outlook. If AI can presumably understand what we wish to say, it can give us feedback on our writing. If your email is too long, incoherent, or missing a bottom line, an AI assistant can help you improve it. It shouldn’t send a better version of your email on your behalf; it could help you improve and become a better communicator. AI can help you refine and express your ideas until you no longer need to use it someday. And that would be a killer feature. It will save us time in the long run and help us improve our communication skills from day one. Only we know what we wish to say. It would be great to have an assistant that teaches us to say it better.
Like?
“What if we could make email cooler? Something that people will want to use more.” That was probably the thought behind one of the latest additions to Outlook. “Everyone is on social media all day. Let’s add some “social vibe” to our enterprise email application!” And so, one day, we were suddenly able to “React” to the emails we received. React? That’s right. Just send a single emoji colorfully expressing that we like, love, are surprised, or sad by the email. What can go wrong?
Once again, the idea seems great on paper. “We are helping users interact without wasting a lot of time.” Instead of hitting the reply button and thinking about a response (or remembering how to summon the emoji keyboard), we can now access the most popular emojis right from the email. We are used to communicating with emojis, so why not use them in our workplace emails as well? I can think of at least two reasons why we shouldn’t.
First, if all you have to say in response to an email is “Like!” you really don’t have to respond. Remember, any email sent has a cost: it creates a distraction. Yes, even those reactions end up somewhere in the inbox of the original sender. When we communicate, we have to be sure we are adding value to the conversation, and sending a bunch of shallow reactions rarely adds value (and I’m being generous here).
Many people indeed expect some response, and an emoji reaction seems better than nothing, but this is part of our communication contract. It is easy to change that norm and decide that we don’t need to show that we agree explicitly if we don’t have anything valuable to add. Reactions, like their social media ancestors, are used to fulfill some need to be part of the conversation, even without contributing to it meaningfully. The problem is that we are already flooded with so many interactions that adding new ones that have little or no value makes our workplace communication more frustrating, not more fun.
The second reason why adding Reactions is not a good idea is even more profound. We are creatures of habit, and the emoji-based reactions are familiar; we associate them with social media. The risk is that the more we use Reactions on email, the more we treat email as social media. As you can probably guess, social media and effective communication have nothing in common.
At work, we communicate to create something together. That is why we need every interaction to be meaningful. Social media, on the other hand, is typically used for entertainment, and an interaction on your feed rarely ends up helping you create something together. The result is that when we interact with social posts, we typically skim through them. We continually scroll until something catches our eye and end the interaction after a few seconds with a reaction or a brief comment. If workplace communication goes in that direction, nothing will ever get done. It is a shallow form of communication. In a sense, it is communication for the sake of communication, not for achieving a joint goal. Instead of solving some of the many challenges of workplace communication, this kind of interaction amplifies them. Instead of co-creating, we get a scorecard with how many Likes we received in the past week.
Dear Microsoft, if you really think emojis are cool, you can add them as personal markers, like categories, so people can privately tag the emails they receive. This might be useful for recapping your week or finding older emails without creating a distraction for anyone else. It might introduce some social vibe, but it will be confined to my inbox and not spill over to turn the conversation into a cryptic stream of happy and sad emoji faces.
When you design a tool for millions of users, you have great responsibility. Whether you plan to or not, you shape how they work and communicate. You can help them communicate meaningfully and achieve greater things together, or you can create distractions, frustration, and shallow engagements.
The problem with Reply All, Copilot, and Reactions is that you might think they are a great success. We use defaults, we like that our mess is being automatically cleaned up, and we are creatures of habit. If you monitor some usage KPIs, you will see that all these features are being vastly used. People might even say they like them, especially when they just roll out. But this doesn’t mean they have a positive impact on the way we communicate. How many people use these features is not the right measure. In fact, the more people will use Reply All, Copilot, and Reactions, the less meaningful our communication becomes. Either they create an overhead, make the conversation shallow, or both.
Nothing would improve Outlook more than killing these features or radically changing them so they will help us and not just be fun or easy to use.
We don’t need to communicate more and with less friction; we need to communicate better.