From responsive to anticipatory. How connected devices are changing the requirements for user experience
Mobile is one of the forces changing the user experience from being responsive to being anticipatory. I’ve got my first mobile phone in the late 90ies. My parents gave it to me as I moved out, so that they still can have some control ;). And basically for the next 8 years, or in my case for 9 years (as I bought my first phone at the beginning of 2008) nothing really changed in the mobile experience. We were using mobile phones for mobile phone calls and texting, that’s all. Ok, there were PDAs somewhere in between. I was a lucky user of a Blackberry, 2007, which allowed me to send and receive e-mails on the go and, in case of emergency, to load a website — very slowly though. So, I was rather skeptical as the first iPhone came out. I was like
“yeah, it will be the same as Blackberry but with a touch screen, and touch screen is way less comfortable for writing than a Blackberry-keyboard”.
But — I bought one anyway.
And yes, the first iPhone generation wasn’t much more then that. But it was just the beginning. Even though the first iOS didn’t support apps, the Internet connection was slow and the camera bad, already the first iPhone generation was equipped with some sensors. And sensors are one of the main drivers for anticipatory experiences, alongside mobile, location, data and social. They are changing our interactions with technology from being purely responsive to being proactive or anticipatory.
From responsive to anticipatory
But, what do I mean with responsive? Responsive web design is nothing new to anyone today. Already in the prognosis for 2012 it was regarded as one of the top trends and 2013 was designated by Mashable as “the year of Responsive Web Design”.
And indeed, to the satisfaction of the users, this trend has established as standard by now. And, don’t take me wrong — I also think it’s great. But let me take it one step further and talk about the semantic of the word “responsive”. What does “responsive” really mean? When something or someone is “responsive”, it means that it responds to something, it’s receptive to something. So it’s passive, reactive rather than proactive.
The majority of systems and interfaces we use and develop today have very limited proactive or anticipatory capabilities. They react to actions and changes that take place in the present — that’s why they are responsive. Multiple aspects of the past and the present, as well as the expectation of user’s future actions, however, influence the behavior of anticipatory interfaces.
But what exactly does it mean? Let me tell you a short story
My typical morning 10 years ago
My sweet dreams are being brutally interrupted by the annoying sound of my alarm clock. I open my eyes and become instantly blinded by the intense sunlight. My boyfriend is still snoring on the other side of the bed. It would be so great if he once stood up earlier than me and prepared some fresh coffee… I think I have an appointment later today, so I start my computer to check with my calendar. Ok, I still have some time, but what should I wear? I spend something like 30 minutes deciding. I go to the kitchen to make some coffee and grab some breakfast. Ok, no breakfast, the fridge is empty again. Black coffee and some dry bread is the best I can get. I have to remember to do some grocery shopping later today. I make a shopping list.
On the way back to the living room — clumsy as I am — I bump my foot against a closet. It hurts — is it broken? Whatever, I’ll take care of it later.
I check the public transportation plan — I have to leave now and hope that the traffic is not to bad. Otherwise I might be late. And, oh no, my sisters birthday! I have no idea what to buy and when I should find the time to do it…
And in the afternoon, on the way home, I stop by a fashion store and overspend a bit… I don’t dare to look at my bank account, as I know what I can expect there.
My typical morning in 10 years
But how could the same morning look like few years from now. My sleep will not be interrupted by the irritating sound of the alarm clock. Instead, I’ll be gently awakened by the smell of freshly brewed coffee, soft morning light and gentle back massage. No, that’s not my boyfriend who is just going the extra mile to make my day so very special. He’s still snoring on the other side of the bed — some things just do not change. But it’s actually fine for me. Sensors in my mattress analyze my sleep and give the information to the coffee machine, roller blinds and massage elements in my bed automatically. So that I am being awakened in the optimal moment — without my deep sleep phase being interrupted.
Air conditioning, music, lighting — the whole apartment is connected, equipped with sensors and it not only can be controlled with simple hand movements or speech commands, but it also learns my preferences, recognizes my mood and adjusts to them. My closet automatically recommends clothing, which would be most suitable for the appointments I have in my calendar. My smart toilet analyzes my values to get the information which nutrients I might be missing. Based on that, dietary recommendations are passed to my smart kitchen. My breakfast is being prepared based on these recommendations, and on my known preferences.
While I’m having breakfast, my smart fridge provides me with suggestions for foods that I should buy, along with the recommendation of the cheapest provider. Actually it’s also possible to automatize this process completely, but I rather prefer to have a look at it. Everything is fine; I have nothing to add, so I approve the order.
On the way back to the living room — 15 years later, I’m still clumsy — I bump my foot against a closet. With a diagnosis app I can quickly make an X-ray scan. Fortunately, the toe is not broken, but only bruised, so I reject the recommendation to seek medical consultation (I know, I probably should follow this recommendation, but I just don’t like doctors).
My self-driving car has synchronized with my calendar and checked the traffic on the route to my morning appointment. Self-driving cars became approved also by the German legislation after all, so I finally can “drive”, since no license is necessary anymore. It looks good; I still have some time to buy a birthday present for my sister. This time it’s quite easy: the recommendations are based on my sister’s social media activity and the history of gifts I have already bought for her. I accept one of the proposals.
In the afternoon on the way back home, I stop by a fashion store. At the very moment I enter the store, I receive a friendly reminder from my personal finance management app: “Agnieszka, you’ve almost reached your spending limit for fashion-shopping this month. I recommend you to leave now and come next month again as the prices are expected to drop”. I reluctantly leave the store, drive back home and go for a run to get some extra points for my health-insurance discount.
Futurists as well as the science fiction literature and movies have predicted such scenarios for quite some time but the actual use of such technologies in the consumer mainstream is just starting.
Enabling anticipatory experiences
Artificial Intelligence is making computers smarter and more useful — as you might have heard, last year a computer have passed the Turing-test for the first time. Increasing memory and processor capacity enable processing more and more data faster and faster. The current iPhone has about as much processing power like the supercomputers from the late 90’s.
And of course, there are the five context-enabling forces, I was talking about at the beginning: mobile, social, data, sensors and location. They all make the rising of anticipatory solutions possible, as they are driving the change of the status of technology in the life of the users.
Technology is not just a productivity tool anymore, but rather a mobile, social and functional companion in our daily live.
Also mobile is not about mobile phone calls anymore, as it still was some 8 or even 5 years ago, even if we had some other features in our phones, PDAs and first smartphones. When you think about mobile devices today, it’s a completely different story.
Making phone calls is currently only the 6th most popular activity involving a smartphone and 40% of the smartphone users say, they could live without it.
But there is more about mobile. Mobile means by far not only mobile phones or smartphones, but also the whole variety of different wearable devices: smart watches, glasses, fitness tracking devices and so on. And there is more to come, in the near future we may wear smart contact lenses and have health monitoring devices implanted into our bodies.
This leads us to sensors, which are an essential component of smartphones and wearables. Starting with accelerometer, gyroscope and magnetometer for orientation, over proximity sensor and light sensor to thermometer and air humidity sensors, pedometer, hearth rate monitor and fingerprint sensors.
But what would mobile and sensors mean without location data? Meanwhile the availability of location information is so essential and obvious, that we can barely imagine functioning without it. The majority of mobile use cases involve location. The time when I moved to Berlin almost 10 years ago and was discovering the city on my bike, with a printed map, you know, on paper, appears to me like ancient history!
Mobile, sensors and location data can all be integrated within social media and enrich the highly personalized content, the user generates. This content allows technology to understand the context of who the user is, what is she doing, what is she looking and what is she likely to do next.
And here comes the last force: data generated by mobile, sensors, location and social. According to IBM estimations 90% of the world’s data was generated in the past two years.
According to Jennifer Erwitt in The Human Face of Big Data
“Now, in the first day of a baby’s life today, the world creates 70 times the data contained in the entire Library of Congress”
We tend to subsume all the data under the buzzword “Big Data” and to focus on the size dimension. But I see rather the small, highly personalized data as the driver of context and the anticipatory experiences.
What are these anticipatory experiences?
Our daily lives might not be yet as penetrated through anticipatory technologies, as I’ve shown in my hypothetical scenario, but we’re indeed not that far away. As you see — we already have the necessary components, the only question is how far and how fast are we going to implement and adapt them. Though, for now there is no common terminus for experiences generated by these components. We’re speaking „anticipatory computing“, „anticipatory Interfaces“, „proactive experiences“, „predictive technology“. “Right time experiences” is also a possible term (and a title of a great book about such experiences), or “contextual technology” based on the “five forces of context” in Robert Scobles & Shel Israels book “Age of Context”, which I also highly recommend. I’ll stick to „anticipatory experiences“ for the purpose of my talk.
These „anticipatory experiences“ are a kind of User Experience Design, where the devices or software the user interacts with, are actively “thinking along”. They know what the user is doing in this very moment, and predict what she’s going to do next. Without the user having to take action.
It’s as I’d write a text to my boyfriend, telling „hi darling, I’ll be home about 8 p.m. and I’m pretty tired and stressed out”, and he’ll prepare a relaxing bath, light some candles and pour a glass of wine. Without me telling him to do that. An ideal User Experience in an Anticipatory Design should be like an interaction with a best friend or a partner, who is not only able to complete my sentences, but can almost read my mind. Or with an extremely mindful assistant, who knows my habits and preferences so well, that she can handle most tasks without any explicit request.
Examples of anticipatory experiences
I probably shouldn’t compare my iPhone to my best friend or partner (although, no one is spending so much time with me as my iPhone does), but it can indeed complete my sentences sometimes. Even if it still does it only very rudimentarily and usually without any deep understanding of context and grammar.
There is also another simple application I use quite often, which goes in the same direction. When I receive a mail and read on my iPhone and this mail contains the phrase “see you tomorrow”, the software presumes, that I might wish to save it as an appointment in my calendar. So it converts the word “tomorrow” in a hyperlink, which links directly to the calendar. Although here again the context is missing.
As you see in this example, “Thursday morning” has been understood properly, but not in the context of „Thursday morning is no longer free“. A functioning anticipatory system should recommend me to delete this appointment, not to create it.
A nice example for an anticipatory system is Google’s Nest. Nest manufactures intelligent thermostats, providing personalized energy plans, based on the behavior of the residents, weather and electricity prices. Nest devices are able to anticipate which temperature settings are needed under which circumstances.
Luna, a smart mattress cover, can learn the users’ regular bedtime and set the bed to a comfortable temperature. It also tracks the sleep quality and claims to combine this knowledge about the users’ night with the information about her day and to use those insights to recommend what works best for her sleep. Luna’s smart alarm can identify the correct moment in the sleep cycle to the user up in the moment of light sleep. I’ll see if it delivers the promise in August, when it’s going to ship.
LGs intelligent fridge “knows” what’s in it and can prepare shopping lists on its own. You can also send a message to the fridge and ask how many bottles of beer are still inside. It can also send you an alert, when it’s time to buy fresh milk.
Another promising approach — a further application from the Google family: Google Now.
This app provides the user automatically with information relevant for his individual actions. Google Now learns the users routines and behavior and matches them with the external conditions.
Depending on the usual daily schedule and the current location of the user, Google Now can recognize if she’s at home or at work, what is the most probable next destination and if there is traffic jam on the usual route. With this information the application can provide recommendations without the user having to explicitly request them.
Also in FinTech there are some anticipatory approaches and I’m very much looking forward to their evolution. For example the Spendific App predicts how much the user can spend based on her income and fixed expenses and help her create and reach saving goals. (Side note: banks are actually naturally predisposed to lead anticipatory experiences, as they have all the necessary, highly personalized data and a high market penetration. Unfortunately my online banking still looks and behaves in the same way as 15 years ago, just showing me a big amount of red and green numbers, so the future will probably belong to other players. End of the side note).
Already these few examples are showing us, that we’re actually not that far a way from my hypothetical future scenario.
One more example: Amazon with anticipatory shipping. For some people Amazons “Anticipatory Shipping” was the first time they’ve heard of anticipatory Experiences. And they are not necessarily enthusiastic about it. This anticipatory logistics-management-tool shall be capable of identifying products, the customer will include into her next order. This prediction is made based on the preceding search and shopping behavior and other factors. The chosen products should wait at the shippers’ hubs or on trucks until an order actually arrives. If implemented well, this strategy has the potential to take predictive analytics to the next level, allowing the data-savvy company to greatly expand its base of loyal customers, as the delivery times can be reduced to minimum.
But something here sounds somehow fishy. Temperature regulation based on my preferences — sounds fine, personalized traffic information — quite convenient. But sending me something, before I have actually ordered it? Even if it’s not really being sent directly to me, it feels kind of spooky.
Anticipation and privacy
Or think about this aspect of anticipation. Wearables like Fitness-Trackers help you to live a healthier lifestyle by collecting data about your body. Or to get a better health insurance rate: last year several health insurance companies stepped up their game by rewarding those customers with a cheaper rate, that were able to prove healthy habits based on the collected data provided by their tracking-devices.
Sure, health insurance companies can’t stress enough, that monitoring those habits is optional, but possible concerns over malpractice and fishy new insurance policies are actually not that far fetched.
Just imagine that: as soon as the benefits of health-tracking get so appealing, that most of us are willing to provide these informations — those of us who remain reluctant to full disclosure suddenly seem like they have something to hide. Insurances then smell bad business, rates are being raised and the days of attractive conditions are over.
Or one other spooky story connected to personalization and anticipation. You might have heard this one as it went around in 2012. An angry man entered a Target-Store near Minneapolis and wanted to talk to the branch manager. He pointed at the booklet with coupons for baby supply and pregnancy fashion and was quite upset. The booklet was addressed to his daughter, still a teenager. The branch manager was quite embarrassed and he apologized.
But guess what happened few days later. It turned out, that it was actually nothing wrong with sending these coupons to this guys’ daughter, as she was indeed pregnant. And Target just get to know it a little bit earlier then the father of the girl. And the interesting thing about it is — she didn’t tell Target explicitly, that she was pregnant. She didn’t even buy anything explicitly connected to pregnancy. The coupons have been sent to her based on the analysis of her usual shopping behavior. Her behavior was compared with this of thousands of other women, and the sudden change was quite a strong indicator for pregnancy.
Don’t know how do you feel about it. I’m actually really excited about the possibilities of data mining and prediction and technological progress in general. But I’m not sure, that I like these examples. Even if they provide me a great experience at the end, first they make me feel observed and somehow uncomfortable.
I call this phenomenon „uncanny valley of prediction“. The term „Uncanny Valley” originally refers to robotics. It states that the acceptance of robots and avatars depends on the degree of anthropomorphism. The bigger the similarity with human beings, the more we like them, at least until some point, where the linearity stops.
The same principle can be applied to anticipatory software.
Approaches like these from Amazon or Target seem to be somewhere around the uncanny valley, as they give the users the impression to know more about them then they actually wish to expose.
And there is one more thing I’m asking myself. Shouldn’t the curve go down again at the end?
Wouldn’t such scenario as in „Minority Report“ be imaginable? For example the police in Bavaria, Germany is using a system called “precob” (quite similar to “precogs” from Minority Report) to predict and prevent future burglaries. Where exactly is the border between assistance and patronizing? Will my daily life, optimized based on my preferences and habits, be easier and more comfortable or will it be boring and predictable, as there is no place for serendipity. Can the anticipatory software develop so far, that it becomes independent, like the OS Samantha in “Her” did? Can it become uncontrollable for mankind?
It will take some more years until we’ll have to deal with these questions on a great scale. Nevertheless, it would be good to keep them in mind while designing anticipatory experiences. So, what can we do to create awesome experiences instead of freaking out the users?
• Being relevant, but not uncanny — sure, we’ll analyze and interpret user data to create relevance, but we shouldn’t give the users the impression that she’s having a stalker.
• Assistance instead of patronizing — the applications should behave like a mindful butler, who is holding back when he’s not needed, who is discreet and offers assistance in an unobtrusive way. But they shouldn’t be like an over-protective mother, who always knows better what is good for her children (even if they are long since grown up) and thus she feels entitled to make decisions for them.
• Try not to frustrate the user — an anticipation, which works perfectly 85% of the time could sound like a success. But it becomes relative, when you consider the loss of trust, which occurs in the remaining 15% of the cases, in which an incorrect recommendation has been made. You can see here an analogy to speech recognition. You can see the frustration with SIRI, when the user inputs are not detected. It’s an astonishing technology, and most user are amazed when it works, but as soon as it starts making mistakes, it can be quickly called stupid. You’ll have exactly the same problem with all sorts of proactive applications, that’s why we have always to keep it in mind and have strategies to overcome it
Anticipatory interfaces are still in their infancy. But all the big providers will expand their proactive capacity further and further. Apple, Google, Amazon and other companies already do possess many of the building blocks necessary for anticipatory solutions. Many predictions of technology trends for 2015 (like for example from Gartner) strongly suggest, that proactive experiences are increasingly entering our daily life. People like us, who are designing and developing applications, should contribute to this development in a positive way, by creating relevant, assisting and, if possible, not frustrating anticipatory experiences.