Get a new interview delivered to your inbox every week!

Design at Facebook

Uncovering the challenges behind building a product at scale

Recently Amandah and I had the opportunity to visit the headquarters of Facebook and get an inside view of how a small piece of this very large tech company operates. With all of the news, figures, debates and punditry that swirl around Facebook on any given day, we saw this as an opportunity to tell a unique story. We wanted to give an objective “fly on the wall” view of the company and some of the people who bring it to life. It’s easy to forget that behind the likes, scrolling news feeds and shades of blue are real people grappling with decisions on how to improve the experience we’ve become so used to. Understanding the problems faced and how the teams at Facebook solve them is not something we often think about when we’re pouring over our friend’s posts. With this unique look inside, we’ll meet some of the people responsible for the design decisions and chat about some of the challenges they face making it all happen.

We pulled up to the Facebook campus and saw a large company sign that donned the iconic thumbs up logo. It said “1 Hacker Way” and it was draped in a rainbow tarp celebrating the upcoming pride festivities. We turned away from the original buildings towards the newly built “1 Facebook Way” facility. It was an incredibly large building that was designed by famous Canadian architect Frank Gehry and it had been completed just a few months prior. After checking in with security we scaled a flight of stairs into a massive atrium to wait for Brett and Maykel; two of the members of the Facebook Search design team and our guides for the day. It was a pretty impressive entrance and wasn’t short on any wow factor with vivid artwork, sharp angles and massive scale.

After connecting with Brett and Maykel, we were introduced to the rest of the Facebook Search design team and then headed out with a few of them to the cafeteria for a bite to eat. The walk gave us a sense of scale of both the company and the building. In either direction there were people, computer screens, offices, wires, beams and desks as far as the eye could see. We arrived at the cafeteria and it was setup as a self serve buffet. This wasn’t your average “mess hall” food as there was plenty of fresh food for all types of paletes, diets and so on. We filled our plates we headed outside to sit on the patio that overlooked the rest of the campus and Menlo Park. We got to know the team and chatted about what each of them did and what the general work life was like at the company. Little details stuck out like the ability to bring your laundry into the building and have it cleaned while you worked. We discussed how services like that allowed you to focus more energy on thinking about ideas rather than having to worry about clean clothes. This was also the case for the food. It was available for free to anyone in the building and took the problem of having to go hunt down something to eat off the table. We wrapped up lunch and took a pit stop for some soft serve ice cream before heading out.

As we walked back I spoke with Brett about some of the challenges faced when designing for a product with massive scale. Brett spoke about fringe cases; those situations where a small set of people have unique requirements that aren’t easily met by the product. In the regular world that small percent of users might consist of a couple hundred or thousand people. Given the relatively small impact, a lot of fringe cases get back burnered. At Facebook this cannot be done. In the world of a product at scale, 1% means tens of millions of people are having a poor experience with the product, and that’s a big problem. Brett explained that no matter how small the fringe case, the team needed to look into making the experience as positive as they could. Brett talked about how design teams have to consider the impacts that a broad range of network speeds and older mobile devices have on people’s experience. Given that Facebook is used worldwide, some locations won’t have the latest and greatest smartphones or high speed cellular networks. Those experiences need to be tested and accounted for on top of the best case scenarios. Brett added that these fringe cases add a unique challenge when designing new features that smaller scale products don’t have to worry about.

After arriving back at the team’s work area, I started chatting with Justin; one of the Facebook Search designers. I asked him what it was like working on a design that hundreds of millions of people would engage with. He explained that it could definitely be a little overwhelming if you let it get to you. He showed me one of the widgets he was making and said that it would typically get tens of millions of interactions on a good day in the US alone. Justin explained that making a design change that resulted in the usage numbers dropping a couple percentage points could definitely keep you up at night. He added that it reinforces the urge inside every designer to make the experience perfect at all times. If you didn’t, millions of people would potentially have a degraded experience. He then concluded that you need to move past the anxiety and do the job you’re capable of doing and know that sometimes things will ship to millions of people in an imperfect form; it’s just part of the job. The team learns and iterates change quickly, so if something doesn’t work as expected, it wouldn’t be long before it was fixed.

“Any time we are pulling content for you, we‘re searching Facebook, whether you typed something in or not. Whether we show you personalized recommendations or Trending Topics, which show you what people are talking about, the results come from search.” —Justin

I noticed Justin was working on user interface elements that were located on the right of the news feed and I asked him what that had to do with search. He explained that search at Facebook wasn’t what you typically think of when you think of search. It wasn’t your classic text field where you type in a query and it gives you results. Facebook is very much a personal experience and Justin explained that Search will explore topics that are trending on the platform on a given day and share those with people who might be interested in them.

After grabbing a few more photos it was time for the design critique, or “crit” as it was referred to. It was a meeting where the entire team gathered to review their work in progress. We all headed into a large boardroom and got situated. Brett called for everyone’s attention and asked who had work to review and how much time they would need to present it. He wrote down names and then gave each designer a block of time to speak in. Each presenter then led a portion of the meeting and showcased their latest work to the rest of the group. Other members offered feedback and ideas were batted around on general improvements or next steps. After the meeting I talked with Brett about what the life cycle of a typical feature was. He mentioned that there wasn’t a formal process that they followed and in a lot of ways it was a positive thing. He explained that strict processes can often be too confining and because of this they frequently get subverted. He added that the entire design team is made up of broadly skilled product designers and they bring lots of ideas to the table on how to make the product better. Team members are grounded in a simple principle of “how can we create more value for our users” and this guides how features are born. After an idea is formed, it’s put into a visual mock up and then brought to the team for critique. As the feature moves through the cycle, designers will keep the engineers abreast of what they’re doing and will often collaborate with them to prototype the idea with real data. Brett explained that having developers aid in prototyping is a big advantage vs. other places he’s worked in the past. Seeing the feature work within the context of their own Facebook experience is a huge aid in making a feature add value. After having the feature designed and vetted with the team, a designer will bring it through further implementation and review with management.

“There isn't a cut and dry process, which I think is a really good thing for us because I think for most of the processes that I‘ve tried generally they don‘t apply to every single situation.” —Brett

The next step brings the implemented feature into a testing and analysis phase. Depending on the size of the feature it may go through rigorous usability testing with someone from the research team but for smaller features they will jump right into A/B testing with a small segment of real users. From here designers and researchers will meet with a data analyst on the team to review metrics and see if the feature makes the numbers move in the right direction. They tend to dig deeper than the top level numbers because sometimes a metric moving in a positive direction isn’t necessarily a good thing. For the Search team, a metric that shows people are searching more would appear to be a positive thing, but if you dig down it might mean that users aren’t finding what they want. Digging deeper into the metrics usually reveals whether a new feature is actually creating value like the team intended. After they understand why the numbers moved the way they did, they will put the feature in queue for a production launch. Brett noted that the web platform will see updates on a daily basis while the iOS and Android apps are typically updated a couple of times per month. He explained that the company runs in a very iterative way and that design is never really considered complete. There is a constant push to move the product forward and create more value for users as quickly as possible.

After we got back to the Search team’s work area, Amandah and I connected with Maykel who was going to give us a walking tour of the building and answer any other questions we had. As we walked through the laneways in the open concept core of the building I noticed a lot of messaging about design, hacking, and “moving quickly and breaking things”. These values seemed to ground everyone who was working within the company; from signage, to t-shirts, to stickers covering laptop lids to vending machines filled with USB sticks, batteries and other things hacker-types might need in a pinch. The hacker culture within the organization felt very authentic and it really permeated through the building. I asked about the culture and how they have managed to maintain the startup vibe during their growth. Both Brett and Maykel gave credit to the company leadership for instilling their values on everyone and not wavering from those as the company grew.

“I think your company’s culture stems from the leadership and their values. So I think the way we operate is in keeping with Zuck and other core leaders and what they value and ultimately that‘s making Facebook better for people as quickly as possible.” —Brett

Maykel brought us over to a cafe that was located within the building for a quick pitstop before heading onward with the tour. Amandah and Maykel were talking about product ideas and how they evolve to become full fledged products within the company and he mentioned that a lot of ideas are tested and ultimately scrapped for various reasons. Maykel added that within the company if someone has a big idea and they can prove its value via a prototype, they can bring that idea to management and pitch it like a startup founder would. If it passes, the idea then makes its way up the management chain for review. Ultimately, if good enough, the idea could make its way to Mark Zuckerberg, Facebook’s CEO, for a final review. If the idea passes, it’ll get prioritized within the current structure of the company and potentially brought to life. Maykel mentioned that he valued that aspect of the company because it meant that one person could have a very large impact if the idea was good enough.

After grabbing a couple lattes, Maykel asked if we’d like to go up and see the rooftop park. We both gave him an enthusiastic yes and we headed up through a stairwell to the roof. As we came out of the building it didn’t feel like we were on top of anything, instead it felt like we were standing in an open meadow. As I panned around taking it all in I could see a few windows, lamp posts and small structures that gave me clues that we hadn’t actually stepped into a teleporter. It was incredible to say the least.

As we wandered through that natural wonderland we started discussing some of the bigger picture concepts around Facebook and what it was trying to accomplish. I asked Maykel about the vision of the company and where it was hoping to take us. He said the vision was pretty simple and that it was to create a more open and connected world. He added that it was about people and bringing them closer to the people and things they cared about. As we talked, we passed full size trees 20-30 feet in height and I kept having to remind myself that I was on top of a building and not wandering through the countryside. The rooftop park had no discernible edges and it appeared to completely blend into the natural landscape. The path we were walking on was a 1/2 mile circuit that meandered through the green space and was no doubt a great place to walk and think through a problem.

“If we fuck up, people use our product less, so we try and do this as little as possible.” —Maykel

After I came back to reality, Maykel spoke about how at the scale Facebook operates there isn’t much they couldn’t do from a product perspective but ‘just doing anything’ wasn’t part of the modus operandi within the company. How they evolve the platform must follow the core vision and be something that adds even more value to its audience. I asked Maykel about some of the criticisms that swirl around the company and he took pause. He thought for a moment while he stared downward at his coffee and then looked upward at us and simply said “you know, if we fuck up, people use our product less, so we try and do this as little as possible.” The response was an honest account of what it must be like working on a product with such scale. So much of what the people do within the company gets scrutinized with one of the biggest public lenses in the world and the company’s motivations are always put into question. When I was speaking with Justin earlier in the day he said that he is a person before he is a designer and then asked me the question “why would I design a feature that would make me want to use the product less?” I felt a little silly after asking the question because the response seemed so obvious. The conversations with Maykel and Justin reminded me of when Steve Jobs had to defend the iPhone 4’s antenna design after it came under huge public scrutiny. Near the end of the ‘Antenna-gate’ press conference Jobs said “When we fall short, and we do sometimes, we try harder. We pick ourselves up and we figure out what’s wrong and we try harder. And when we succeed, [people] reward us by staying our users, and that makes it all worth it.” Jobs went on to add that when the work they do comes under fire, everyone at Apple takes it very personally. I got the sense that the same thing happens here at Facebook. Public scrutiny is definitely something you’d have to be comfortable with even though at times it can be fairly disheartening. This constant review and judgment from billions of people has to be one of the most human challenges faced by the people who work at Facebook.

After our walk through the rooftop paradise, Maykel took us over to see the original Facebook campus that was across the street from where we were. The vibe in the original campus was a little different than that in the new building. It had more of a university campus feel with walking paths between numerous buildings and courtyards. There were burger joints, an arcade and other areas where you could go to chill out; there was even a Philz Coffee. As we walked down the pathways we got a sense of the same hacker fervor on this campus as well. Classic pirate flags flew above buildings, hacker signage adorned buildings and there was a large courtyard with the word ‘hack’ spelled out in the stone underfoot. We stopped by the Facebook Analog Research Lab which was a place for all employees to go and get involved in making non-digital things; from woodworking to print making. Hacking, it seemed, didn’t just involve a computer and keyboard at Facebook.

After a short trip through the Instagram building, we all hopped on the Facebook transit and headed back to the new building. Between the old campus and the new one you could feel a clear evolution in the culture of the company. The older campus had a more youthful feel to it whereas the new campus felt a little more older and wiser. No matter the feeling the common values were still clear.

We got back into the new Gehry building and walked back to Maykel’s work area. We said our goodbyes to the team and thanked them for their time. Brett then showed us back to the entrance where we had entered earlier in the day. We stopped at a massive model of the building that laid near the exit. It gave a sense of the size of the building that was easier to comprehend than walking through it. The model put into perspective just how large of a company Facebook has become and how successful it’s been at achieving its vision thus far. After getting a glimpse into the inner workings and being witness to its culture for the afternoon, I was left with a sense that the company had a goal and was moving quickly in uncharted territory towards it. All of its users have hitched themselves to the train and have voted with their likes and personal information that they approve of where Facebook is taking them. Like both Brett and Maykel explained throughout the day, it’s really about the users and what they value and if Facebook wasn’t creating something people wanted, we certainly wouldn’t be standing in the massive building we were in.