Searches, notifications & emoji: What data reveals about people

Main illustration: Harry Woodgate

Don’t mistake data for a dry collection of numbers – looked at in the right way, it offers an entirely new window into human nature.

Every time we share something on Instagram, buy something on Amazon, search in Google or swipe right on Tinder, valuable data is generated that reveals more about us, our attitudes, our interests and our needs. This information can be used to enrich our experiences and to improve the products and services we have access to.

Most people tend to think about the technical aspects of data science. For me there’s a much more human side to data science that is often overlooked – it gives us a unique lens through which we can discern aspects of ourselves that were previously invisible.

Through data about the collective, we can learn more about us as individuals. The three data explorations below show that data is fundamentally about people and their behaviors. By not considering this human side, we miss an opportunity to truly understand people and in turn build amazing experiences just for them.

Mobile data is a window into human behavior

In 2006 I was given the unique opportunity to work with a European mobile operator and analyze a large-scale dataset to understand how mobile users searched and browsed the Web at that time. The dataset was made up of 300 million URLs and 6 million search queries by 2.6 million people over one week in early 2006.

Remember that this was pre-iPhone. A clunky Nokia full of buttons was as smart as a smartphone got in 2006.

Mobile portals were the primary way of accessing online content. To search or browse the web you had to go through your mobile operator’s gateway. This involved a lot of clicking and waiting and scrolling through fiddly mobile interfaces and eventually you might find what you were looking for.

I was interested in understanding more about the types of information these early adopters of mobile search looked for because the behaviors of early adopters of any technology are normally strong indicators of what mainstream users may do in the future. If you think back to 2006, did you search or browse the Web from your mobile phone?

To understand the types of information people searched for, I classified a sample of the search queries into one of 16 topics.

Popular searches included things like “Britney Spears” and “ringtones”. The chart below represents the top 5 most popular topics.

many great image

As you can see, accounting for a whopping 61% of all queries was adult-related content – queries such as “free porn”, “porn”, “sex” and “sex stories”. I found that fascinating. Most of the content of interest here would have been images. All that effort, all that scrolling, all that pixelation and the image quality was likely to be so poor.

This early analysis highlighted something interesting about people and their behaviors. Mobile phones are highly personal devices – they tend not to be shared. Even today if I asked anyone to borrow their phone for a second, they’d likely let out a little gasp of horror and say no. And that’s because your entire life is on your phone – very intimate and private details of your life. The conversations you have with loved ones; photos of your most cherished memories. And the queries and websites you visit day to day. It’s likely that people felt more comfortable querying adult terms on these private, discreet devices compared to their PCs.

We’re highly attuned to smartphone notifications

Fast forward to 2013. Smartphones had evolved to beautiful, sleek devices and had become a constant companion to almost 1 billion of us across the world. And just to make sure we never missed anything important, apps like Whatsapp and Facebook started using proactive, push-based notifications using sounds, vibrations, icons and badges, to inform us about new messages or events, even when we weren’t actively using our phone.

At the time of our analysis, WhatsApp was handling over 50 billion messages per day and Facebook had more users connected via their mobile phones than their desktops.

At Telefonica Research in Barcelona, we wanted to understand how many smartphone notifications people had to deal with? Even more interesting to me was asking, “How did people deal with notifications and what does that tell us about them?” We didn’t have access to a dataset to answer these questions so we created our own.

We built a mobile app using Android that tracked all notifications received, the time they were received, the app they came from and the time between receiving the notification and viewing it. Then, we recruited a group of 15 people who installed the app on their phone and let it run in the background for two weeks.

These interfaces are fueling expectations of responding quickly.

After cleaning the dataset we ended up with 7,000 notifications. Almost half of them came from messaging applications like WhatsApp and SMS. This means each person received an average of 65 notifications per day – that’s one notification every 22 minutes.

The most interesting insights emerged when we looked at how people handled notifications, in particular their reaction times. We found that notifications that came from messengers were attended the quickest – 6 minutes during weekdays and 3.5 minutes on the weekends. 3.5 minutes! That’s quick.

But if you think about it, messaging apps like WhatsApp tend to be used to communicate with loved ones, family and friends. As a result there’s an increased expectation of a faster response. Because the “two blue ticks” in WhatsApp indicate that a message has not only been received but has also been read, our loved ones now know for sure when we’re ignoring them. And we now know when we’re being ignored. These interfaces are fueling expectations of responding quickly. “He knows I’ve read his message, I better respond.”

The most fascinating thing for me was when we looked at the effect of a phone’s ringer mode – whether it was silent, on vibrate or normal – had on people’s reaction times. We found that notifications were viewed significantly faster when the phone was in vibration mode compared to both silent and normal mode. And we found no evidence that putting the phone into silent mode led to slower response times.

This means that people proactively checked their phone to see if they received any notifications when their phone was in silent mode. “Has someone liked my Instagram photo? Did anyone retweet my tweet? Has she messaged about a second date?” This analysis highlighted the extent to which we have become connected to, reliant on and borderline addicted to our phones.

Emojis really do reveal our emotions ?

At Intercom, we’re always looking for new, rich ways to enable our customers to communicate. One such way is with emoji. We love emoji! So much so that a couple of years ago we added the ability for our customers to add emoji and stickers into their conversations because emojis help people understand the ways in which words should be interpreted.

At this point, most people know that the majority of in-person communication is actually done through body language, gestures, tone of voice and eye contact. Because messaging is stripped of that, it’s significantly easier to misinterpret what someone is saying. For example, “I’m fine ?” means something totally different than “I’m fine ?” or even “I’m fine.”

Understanding how emoji are currently used can provide insights into how people express themselves in contemporary online communications. And could help us design novel online communication in the future.

So we analyzed emoji trends abstracted from more than two million anonymized conversations that took place our customers (i.e. businesses) and their end users (i.e. customers) using our messaging platform during a 3-month period, from June to August in 2015 and 2016.

We found that emoji use in business messaging quadrupled to 12% in 2016.

The 20 most common emoji found in business messages in 2016 were quite similar to ones you might use in your personal life:

many great image

Over half of the top 20 emoji were facial emojis pointing toward the idea that people are using emoji to convey something that in messaging text alone can lack: emotion. Then we found objects like the rocket ship ?. And as you might expect in a business setting, money-related emoji also featured like the dollar emoji with wings ?

Mirroring is important in real life communication because it shows that we’re positively engaged in a conversation. Our body language, facial expressions, voice and tone subconsciously start to resemble each other, which indicates we’re both listening and understanding. And we wondered if businesses mirrored the same emojis used by their consumers.

In theory, if businesses were connecting with their end users, we should see them mirror emoji usage, and therefore the top emojis for businesses and consumers should be quite similar. That’s not what we found; instead this is what we saw: Consumers used facial emoji 30% more than businesses did (83% vs. 51%); businesses stuck to objects (18%) and money symbols (11%), neither of which show up in the consumer list.

Data provides us with an entirely new vantage point.

This analysis highlights that end users, you and I, communicate with businesses in similar ways to how we communicate in our personal lives. We’re highly expressive, personal and emotive. This makes sense. If you think about your own relationships with your favourite local businesses, it’s always nice when you can actually connect.

When the barista in your local coffee shop knows your name and your usual order and asks you about your day, you start to build relationships with them. Relationships are key to businesses meeting their customers’ needs and desires, but what we see with this data is that businesses don’t manage to mirror their customers’ approach – it’s subtle, but reveals an asymmetry that speaks to the challenges businesses face in the internet era.

Data is about people and their behaviors

Whether it’s understanding the types of information early mobile phone users look for online, or how people deal with distracting smartphone notifications, or even how emojis are used in business messaging, data is about people and their behaviors.

The reason why I work with and love data, the reason why I’ve devoted most of my career to working with data, and the reason why I now lead amazing data teams, is because I’m fascinated by people and how they behave. I’m fascinated by how they interact with technology and how technology shapes their lives. And data enables me to question and understand this.

Data provides us with evidence. It provides us with an entirely new vantage point from which to perceive people’s behaviour, a new prism through which we can understand society and attitudes and people’s likes, dislikes, needs and desires.

Once we understand we can improve. We can enrich. We can even predict. We can design and build new, amazing experiences that enhance the day to day lives of people.

Data is made up of numbers, sure, but when you know what to look for, what it reveals is very human – and when you’re striving to make business more personal, that insight can make all the difference.


This article is based on a talk I first delivered at Inspirefest earlier this summer.