Dropbox's Head of Design on the Dawn of Personalized Products

Dropbox's Head of Design on the Dawn of Personalized Products

As one of the first designers at Facebook and Dropbox, Soleio Cuervo has seen a lot. Now he's looking at the next trend: Personalization.

Soleio Cuervo has been thinking about how to tailor technology to users for almost a decade. And he’s had a pretty unique vantage point. As one of the very first designers at Facebook (and reported creator of the famous "Like" button), he was instrumental in the creation of many of Facebook's early products and features. Now leading design at Dropbox, he spends his time thinking of new ways for products to understand our needs and wants in real time.

“Personalized streams are everywhere we look now: Twitter, Instagram, LinkedIn,” says Cuervo. And that’s not even delving into the new movement toward predictive technologies like Google Now that serve up information before you even know you need it. But personalization doesn’t have to be the provenance of massive companies. Startups can take advantage of the same principles to build deeply engaging experiences.

After years of firsthand work and observation, Cuervo has seen four ingredients emerge that power personalized products. At First Round’s recent Design+Startup event in San Francisco, he explained each one and how newer and smaller companies can put them together to not only build great products but accelerate progress for everyone.


Cuervo is a firm believer that you need to know where you’ve been before you can decide where you’re going. In this case, it’s important for founders to understand how product personalization in computing has evolved and where it is today so they can decide what their next move should be.

It may be hard to believe that the concept of personalization is as old as personal computing itself, but that’s what skeuomorphic design was all about on the early Macintosh: Folders and files were fashioned to look and operate like their real-life counterparts to ease early adopters into digital work. Obviously since then there’s been a massive jump.

Over the past ten years, consumer technology has gotten to know our names, where we're located, what we’re interested in, and who we’re connected to —both socially and implicitly. It was at the start of this transition that Cuervo entered the scene, joining Facebook right as algorithms were making it possible to give individual users tailored experiences.

What early Facebook saw was the beginning of a system capable of improving itself the more people used it. “A virtuous cycle spun up as more people joined, as more people engaged with the product in new ways, and as we molded the UI to adapt to their usage. Now News Feed seems like an old idea.”

But there’s still a long, long way to go, he says. “Thousands of iPhones are built and shipped every hour around the world, but every single one delivers essentially the same onboarding process —every new owner of an iPhone goes through the same first-run experience no matter who they are.” In the future, he believes every device will conform to your identity the moment they are unboxed (not unlike the operating system in the movie Her).

So how can companies take all of these past learnings, recombine them and push them further to create paradigms beyond the feed? Already, Cuervo says, the startup ecosystem is presenting new opportunities for personalization, and the data and technology are largely there. This is where his four ingredients come in. If a startup can properly integrate each into their roadmap, they have a shot at doing something that will change the way millions of people interact with their products and each other.


Your identity is at the core of any personalized experience in this new age. It extends beyond your name, profile picture and the information you’ve filled out in your account. It’s a robust entity that evolves as you make decisions and take actions across multiple devices and platforms.

To clarify and define people’s identities, Facebook tweaked existing products and shipped new ones designed to do exactly that. When they first introduced photo tagging, for example, the feature was inherently social and created value not just for the person uploading the photo but also the subjects tagged in those photos. A photo tagged with multiple people not only contributed to the profile of the author but of the subjects as well; it also created a teaching moment for every person who received a notification that they were tagged in a photo. Friends taught friends about the feature organically — and that's just one example of the many ways Facebook encouraged people to connect and express themselves.

“When people can build an identity with your product over time, they form a natural loyalty to it.”

As a company, it’s critical to invest in this relationship with users. “Personalizing software is all about figuring out how to help people develop a sense of identity using the service, and then building features that let them accomplish that naturally.”

It’s equally important to recognize that people don’t have just one identity online like they do in the real world. Data shows that people define themselves sometimes very differently across services —they’ll have a different persona on LinkedIn than they do on Instagram than they do on Facebook and so on. Sometimes they want to project a professional identity. Sometimes they want to project an artistic, creative identity. Your product has to cater to the identity that serves your company’s goals best.

At Dropbox, Cuervo is part of the team working to serve identities for work and personal contexts. After launching as a consumer-facing service, the company is branching into the enterprise in a big way with Dropbox for Business. This means that users need to be able to distinguish their personal accounts (and personal identity) from their business accounts. It’s also motivated Dropbox to ponder the elements of people’s professional identities.

“People’s business identities are shaped by the company they work for, their role there, the team they work with, the people they work with the most, the information that they need to know to do their jobs,” he says. “There’s a whole different cadence to business collaboration, and organizations have their own identities and needs that we’ve had to learn to serve.”

The takeaway for startups is that you should actively manage people’s identities in ways that encourage the behavior you want —whether it’s getting people to buy something or converse more with each other or share media. The features on your roadmap should bring out these aspects of your users’ identities whenever possible. And if engagement is your goal, your product should take cues from how people are organically identifying themselves on your platform.

“Just look at all the ways people have tried to stand out on Twitter —everything from at-replies to retweets to hashtags just emerged out of the userbase,” says Cuervo. “The company realized that people were using these things to cultivate value and ultimately an identity on the service that made them part of an early community. Then Twitter started baking all of these features into the service to reinforce the behavior that was getting them traction and making the platform so successful.”

The first hashtag ever tweeted.


When people hear the word “graph,” they immediately think of their personal social network. The term has been popularized to describe everyone’s system of relationships as a communal asset. Your social graph doesn’t belong solely to you, it belongs to everyone in it. And the more people who join the network, the more valuable the graph becomes, Cuervo says.

Today, there are many graphs across many different platforms. Just like people have a variety of identities spanning the digital world, they are also members of a growing number of groups, communities and networks that may have nothing to do with the people they know in real life. They belong to interest groups, follow celebrities (Kim Kardashian has 15.9 million followers on Instagram), connect with professional contacts, tap into media sources for their news, and more. All of these systems of structured relationships act as graphs that can be used to deliver tailored experiences to individuals.

You might not think of it this way, but even the people you share files with on Dropbox represent a new graph; the people who re-pin the images you share on Pinterest do too. “As more graphs are being built and popularized, the relationships between people and data are becoming more dense and fluid,” Cuervo says. “Every day, there’s more information available about people, what they want and how they may act on their desires.”

To leverage this data, Facebook introduced Facebook Connect, allowing users to log in to completely different services and products using their Facebook credentials. “The idea is that now you can take all of your friendships and travel across the web with them to see what they’re doing in the context of a third-party service,” Cuervo says.

Some key examples include apps like Runkeeper, Lift, and Duolingo, which let you import your friends and see if they are also interested in running, self-improvement, or learning new languages. If so, these apps benefit by letting you track your progress together, compete and share —basically getting you more deeply engaged.

“The best graphs are the ones that align with the value you want to create for your users.”

As an entrepreneur, you want to select the graphs that make the most sense for your product. Instagram can attribute a fair share of its momentum to celebrity watching. “Companies that can cultivate unique, valuable graphs are the ones that are able to create truly novel experiences for their users,” says Cuervo. “If you can assemble a compelling graph for users, it can be a major source of differentiation and user acquisition, and retention.”


Mobile has added a whole new layer to the mix. Smartphones and tablets let users broadcast where they are, communicate on the go, and use products across a host of different contexts. Suddenly, it’s not enough to offer just one product —you have to create distinct experiences for every context that makes sense.

“Soon it’s not going to be enough to have one service that looks the same across multiple devices,” says Cuervo. “People want a seamless experience, sure, but what that really means is they want it to be easy to use your product at all times and they expect it to maximize for their convenience. Basically, your product or app needs to be increasingly cognizant of the devices each user owns and how they form habits on each device type.”

Of all the personalization tactics out there today, this may be the most important as the mass migration to mobile continues. “First we just had our smartphones and our computers —now we’re going to have our glasses and our watches and however many other wearables, our cars, our televisions — the list goes on.”

Touchscreens like the one in Tesla's Model S will become increasingly common and customizable.

“The more devices that emerge, the more nuanced your product will need to be.”

“You can look across the most successful companies that are out there today, and they are the ones doing exciting things across multiple platforms and devices.” Pandora is a prime example as an early entrant to the automotive space that has also focused on creating thoughtfully different web and mobile experiences. Similarly, winners in the enterprise sector will be those that embrace the way people establish workflows across multiple device types, he says.

“The amount of data that all of these services are accumulating per person per device is amazing,” Cuervo says. “Soon, our ability to manage people’s attention no matter where they are is going to be incredible. Already people are getting annoyed with how their mobile devices rack up notifications when they are away. They don’t want to scroll through to find what’s actually important to them at that moment. The new wave of innovation will be all about presenting the right information at the right time given the context of relationships, location, and device.”

Startups and apps that can filter people’s huge influx of information in a way that seems natural are set up to succeed in today’s climate. “If your product can tell the difference between someone sitting in front of their computer, and someone on their phone running late for a meeting, you’ll also have endless opportunities to expose the right information at the right time,” Cuervo says. “This has enormous implications for how contextual advertising will create legitimate value for customers over the next decade.”

In addition to landing more revenue, there are countless ways you can learn to improve your user experience, too. “Just look at the novelty ofUber. When you call a car, you’re told how many minutes away they are, and you can actually track the vehicle’s approach. All of this empowers the user to decide to hail a car versus wait for the bus, or sneak in a coffee shop purchase before the driver arrives, etc. It’s a richer experience that makes them feel in control while managing their attention very intentionally.”

This can sound boundless and intimidating, but we’re just at the beginning of this trend. Startups looking to launch personalized products should fix their sights on how tasks and needs vary across different devices over the course of a given day.

“How will people receive emails on their wrists? On their giant televisions? In their cars?”

“Software makers will need to form a much deeper understanding of each device type and how they function in a given user’s life,” Cuervo says. “If your use case simply doesn’t fit with a certain device, don’t force it, it’s not going to fly.”


User behavior is somewhat of a catch-all for the attributes discussed above. People’s identities, graphs and varying contexts all influence how they behave and use products —and this is where things get very interesting. “You can start to ask, ‘Okay, based on all of these things and a history of behavior, what is this person likely to do next?”

Cuervo’s most emphatic advice for startups: Don’t be afraid to be opinionated about how your users should behave. You’ve created a product to get them to do something. Get them to do it. “Be purposeful when it comes to driving a particular type of behavior —have a very strong viewpoint about how people should be using the service you’re providing.”

“Dare to be opinionated about what's important to your users at every moment.”

Nest is a fabulous example of a training UI,” says Cuervo. “It’s a thermostat that learns a person’s behavior over time. It quickly forms patterns around when you're home and when you’re not. Soon it’s turning on the AC a few minutes before you walk in the door. It does this by building a behavioral model that says, ‘There’s a 95% chance she’s going to get home by 8 p.m., and we know she prefers 72 degrees most evenings, so we’ll start cooling things down now.’ That kind of model is going to be essential for any highly-personalized experience in the future.”

Netflix is another interesting example. It’s gradually moving away from traditional navigation toward a UI of precisely-calculated movie recommendations. “This is a big strategic gambit for them,” Cuervo says. “But they have what it takes to know when their users will be in the mood for a particular genre or actor —basically to say, ‘We’re confident that this is what you want to watch at this moment in time given your past consumption patterns.’ The end result will be a much cleaner, streamlined service that feels effortless and perfectly tailored to the average Netflix user.”

Experiences like this have enjoyed the label ‘automagical’ —when predictive technology is so right on the money that you get what you want before you even know you wanted it. But not everyone is on board. Just as Facebook was inundated with complaints when News Feed launched in 2006, new technologies that seem to be too smart or too prescient have turned some people off. Entrepreneurs interested in working in this space will need to be able to address and assuage these attitudes.

“Determining how automagical your product should be is another opportunity for personalization,” Cuervo says. “The best form of software will be the kind that gives people the ability to decide for themselves when they get the algorithmic experience or when they decide to take control — something along the lines of, ‘Hey, I want to go to manual. Please stop switching gears for me.’ But it’s a very difficult balance to strike without introducing tons of complexity to the product experience.”

His recommendation is to build a product that starts out as very opinionated and directive, but that can still adjust to people’s reactions. “Find the right sporadic, contextual opportunities to prompt the user to confirm an algorithmic response. Like in the context of an email client, one could ask: ‘You archive emails from this person a lot. Do you want to automatically archive emails from this sender?’ Most importantly, make them feel in control of their own experience with your product. You want to tailor their future use, but only in a way that is consistent with what they want.”


“Building extremely personalized products begins with your design process.”

What tools are you investing in that allow you to personalize your product experience as you’re prototyping it? How are you sharing designs with peers and stakeholders for feedback?

“Sharing a quick screenshot isn’t going to cut it much longer. Make sure you can create prototypes that pull in your stakeholder’s graph data and personalize the experience for them. That way they can really see the experience from the perspective of a user,” says Cuervo. At both Facebook and Dropbox, when his team rolls out a new feature to colleagues for internal testing, they create personalized sandboxes for them to see how the thing will actually work in the wild. This is how you get the best feedback: When stakeholders experience the prototype through their personal lens.

Also, capitalize on the existing work of other companies that have come before. Resources and APIs like Facebook Connect can give startups a huge boost in the realm of personalization. Here, Cuervo points to a delightful app called Timehop that connects to your Facebook, Twitter, Flickr, Instagram, and Foursquare accounts to build a personal time capsule for every day. By pulling in all of the photos and content from elsewhere, it allows users to revisit old moments seamlessly.

Just remember that there are dozens of data firehoses out there, Cuervo says, and the way to capture the activity you need may be to depend on others. Don’t assume that relying on one or more of these platforms will keep you from building something special.

A final development tenet: “Remember that you’re not competing against other services. You’re competing against people’s habits. The companies that are truly out there disrupting things are the ones that drive a wedge into people’s habits. They say things like, ‘You used to walk out to the street to flag a taxi down. Now with Uber you can hail a car from your desk, so break that habit of leaving the building.’ As you’re creating your product, always be thinking about what people currently do and how you can very purposely create a new set of habits around your service. That’s what retention is —helping people build a routine around the utility you provide. Successful tech companies are built on new habits they helped form.”