Why is Dropbox reinventing itself?

A chat with Dropbox VP of Product Adam Nash and CTO Quentin Clark

According to Dropbox CEO Drew Houston, 80% of the product’s users rely on it, at least partially, for work.

It makes sense, then, that the company is refocusing to try and cement its spot in the workplace; to shed its image as “just” a file storage company (in a time when just about every big company has its own cloud storage offering) and evolve into something more immutably core to daily operations.

Earlier this week, Dropbox announced that the “new Dropbox” would be rolling out to all users. It takes the simple, shared folders that Dropbox is known for and turns them into what the company calls “Spaces” — little mini collaboration hubs for your team, complete with comment streams, AI for highlighting files you might need mid-meeting, and integrations into things like Slack, Trello and G Suite. With an overhauled interface that brings much of Dropbox’s functionality out of the OS and into its own dedicated app, it’s by far the biggest user-facing change the product has seen since launching 12 years ago.

Shortly after the announcement, I sat down with Dropbox VP of Product Adam Nash and CTO Quentin Clark. We chatted about why the company is changing things up, why they’re building this on top of the existing Dropbox product, and the things they know they just can’t change.

You can find these interviews below, edited for brevity and clarity.

Greg Kumparak: Can you explain the new focus a bit?

Adam Nash: Sure! I think you know this already, but I run products and growth, so I’m gonna have a bit of a product bias to this whole thing. But Dropbox… one of its differentiating characteristics is really that when we built this utility, this “magic folder”, it kind of went everywhere.

That’s why you hear these giant numbers quoted, you know, of more than half a billion users, exabytes of content, etc… but the thing is, it allowed us to reach customers of all sizes.

So we have individuals using Dropbox, small businesses, and large enterprises. I think in the last few years Dropbox has been learning a lot about how people get their work done with Dropbox. How can we make Dropbox an even better environment — a smart workspace for people to get their work done together? That’s a lot of what you see motivating these announcements.

Kumparak: What had to change from a technical standpoint to make this possible?

Nash: The good news is Dropbox has some pretty phenomenal technology roots. Dropbox is one of the few companies that are scale that actually build custom infrastructure. Like… we built out our own cloud services around our own data centers. Magic Pocket [Note: Magic Pocket is Dropbox’s custom architecture for handling many exabytes of data behind the scenes] is one of the reasons that we can store this content so broadly, globally, and have the economics that we do. That “magic folder in the cloud” that seems so simple to the end-user really has a lot of technology and infrastructure behind it — so we had that as a base.

But probably the biggest change was moving to being a foreground experience. I’ve used Dropbox for probably the better part of a decade. Dropbox, being invisible in the background, is exactly what you want for that magic folder in the cloud. You don’t think about it, it’s always working, it syncs everywhere; if I have to check it from my phone, or from my work computer, or from home, it’s everywhere. I get to it, I can preview it, I can share it. That isn’t going away.

But when we talk about collaboration, and talk about integrating things like Slack and Zoom, etc., to bring all those pieces together you really need an application interface. So when you hear us talk a lot about the new Dropbox, and the new desktop application… that’s why. There’s only so much that you can help people with when you’re a menu… behind a menu… behind a right-click, you know?

When you look at Finder, or at Windows Explorer today, what don’t you see? You don’t really see people. You see files, you see folders… but you don’t see people, you don’t see their activities, you don’t see the interactions around it. So that’s what we’ve tried to bring into the new Dropbox.

new dropbox

Kumparak: Some have complained that you’re sort of complicating that initial promise of Dropbox, of a simple shared folder. How do you respond to that?

Nash: I have to admit that this was a little surprising to me, only because we’ve put such a high priority on making sure that everything that people love about Dropbox continues to work the way it did before. So if you’re a Mac user, and you have this work pattern where you open up the folder, you drag files around using the MacOS finder… all of that still works.

So hopefully the new Dropbox just means that you can now get more done with Dropbox, but you don’t have to use these new features. If you love that magic folder in the cloud, that will not only keep working the way you love it, it will keep improving as well.

But the new Dropbox is about us embracing, I think, an even bigger challenge. How can we go beyond just solving what we did in the past, and add more value to our customers? I think that’s the motivation of our software; we’re always looking for more to do for our customers.

Kumparak: I think a common user hesitation is the idea that an app they’re already used to, already familiar with, is going to be using up more memory, or CPU, or screen real estate. What’s the logic behind shipping this as part of the existing Dropbox experience, rather than building it out as a separate tool?

Nash: I think the reason is that we really don’t see it as a separate tool. We see it as an evolution of the experience.

But I would agree with you. We’ve all been through… well, I don’t want to name names, but we probably all have a history of going through different products and services, where they changed things to where maybe you have to relearn a certain workflow, or relearn the way we get things done.

I feel like every time there’s an operating system upgrade, or an upgrade to your favorite smartphone, there’s always a bit of “What changed now?”. So I understand the anxiety people have.

This focus we have on productivity is serious. I think when people play with the product, they’ll see that we’ve tried to elegantly do it in a way that makes sense. For example — the new tray, the new “For You” tab. People already have a [Dropbox] tray, they already use it. We’ve worked really hard to make sure that it feels natural so when you open the tray, everything you did before is still available — but now those new capabilities are elegantly integrated. So now there’s a search box, for searching content both native and cloud. That new “For You” tab, where we’re intelligently recommending things; that wasn’t there before, but we’re adding it to a surface we already had.

The desktop application is probably the most significant change, and I think what people are going to see is that for all the workflows they were using with the old Dropbox, that magic folder… it still works the way they want it to.

Kumparak: What’s had to change from a privacy standpoint? [With the new Dropbox] you’re asking people to hook up their G Suite, their Office 365, their Trello, and their Slack. That’s pulling a lot of data from generally protected third party services. How do you go about that?

Nash: First of all, with our business model…  in being a subscription service, people pay us for the software. This really aligns us with our customers. They trust us with their content, first and foremost, and we try to earn that trust every day, every week, every month, and we take it really seriously.

One of the reasons we’ve taken the approach of doing these deep partnerships [with these other services] is that we can structure the agreements to make sure that all the content that’s exchanged abides by the high standards that Dropbox has set for privacy. I feel like we have a good reputation for that, for a reason.

Nothing has changed on that front; content that you’ve stored in Dropbox, content you share with Dropbox, it’s still under the same policies we had in the past. As more and more customers become attuned to issues of data and privacy, we feel like Dropbox has a strong role to play there.

Kumparak: Does bringing in this sort of data potentially make Dropbox more of a target for malicious parties?

Nash: I don’t know how to rate it “more” or “less”, that sort of thing.

It’s one of the reasons I joined the company: Dropbox has Internet scale.

Dropbox already, long ago, hit half a billion customers. When you hit anything that rhymes with billion, you’re going to be a target for folks. But that’s why we’ve invested so heavily into our infrastructure, that’s why we’ve invested into the technology we have.

I think you’ll find in the community, on the talent front, Dropbox is one of the few companies of our scale that has the reputation it does around infrastructure, around security.

For better or worse, that’s because we hold three exabytes [Editor’s note: that’s 3 billion gigabytes] of content now. I think we’re now up to over 550 billion pieces of content trusted to us by over 600 million users. I mean, it’s a really huge volume.

But we see that as a strength. If you can build a platform that people trust with their content, we then think that’s the right part to build great collaboration and focus and productivity around. A lot of companies, a lot of products, struggle to get users to trust them with content — and we don’t think that problems going to get easier as people are getting more attuned to different business models, and who they’re sharing their data with. So we see this as a huge advantage that already so many people trust Dropbox, and we spend a lot of time making sure that we earn that trust.


GettyImages 1177090465

(Photo by Matt Winkelmeyer/Getty Images for Dropbox)

I also got a chance to chat with Dropbox CTO Quentin Clark, who joined in September of 2017 to lead the company’s engineering efforts. We talked about his perspective for the company’s focus shift, how they’re protecting the new data they’ll be handling from their integrations with Slack/Trello/etc, and why they’re getting into things like image recognition.

Greg Kumparak: How’d the launch go?

Quentin Clark: Really well. As a product person, I feel like I’m, you know, bearing our art, bearing our work, and getting judged for it. It’s always a little nerve-wracking, but I think it’s gone really well. We’re really happy.

Kumparak: How’s the response been so far?

Clark: Positive. I think people want Dropbox to continue to be successful, so we have the brand permission to do this. I think people are starting to get a sense of… like this content [we store] is really valuable. It’s like the heart of what we do at work. If there’s a way to make it easier to stay focused on that, it seems like a good thing. So it’s resonating.

Kumparak: It’s a bit of a shift for the company. Can you tell me about this new focus?

Clark: It comes out of this recognition that, as we’ve watched our users use Dropbox, and as we’ve done research […] with business schools and whatever else to understand the changing nature of work… it comes to this standpoint that we have this fairly rich position and tooling around content.

When you look at the changes in the modern SaaS world, how teams work there — I’m gonna draw a little diagram [he tears out a sheet of paper, drawing a Venn diagram with three circles] — how teams work there has really evolved. Teamwork is really at the intersection of three activities: the content that people work on and build together, the communication they do around their work, and the coordination between people. Dropbox Spaces is really focused at the nexus of these things.

Kumparak: Okay. How many users does Dropbox have now?

Clark: We have 600 million registered users.

Kumparak: A lot of those users signed up for this idea of this very simple folder, this Dropbox, this shared folder. Some of the early feedback so far has been that this complicates that.

Clark: It doesn’t have to.

I’ll say two things. We’ve been very careful about the design of this launch. If users just want to stay anchored to the classic file sync-and-share job that Dropbox has done a great job with for years, they can. There’s a place in the user experience where they can go and say ‘No no no, don’t give me the Dropbox foreground app, just stay in the finder/explorer/etc, I’m good, thanks.” We’ve been very careful about that.

The other thing I’ll say is that we’ve continued to make deep, deep investments into that core functionality. Core synchronization of content, and sharing of content, is still the foundation of what we do. We’re never going to walk away from that. I’ve been making big engineering investments into our synchronization engines, so that we can continue to scale, so that we continue to adopt the new operating system releases, so that companies of all scales continue to embrace Dropbox. That takes continued vigilance and investment, and that’s not going away.

Kumparak: So the integration of Dropbox directly into the OS as a finder/folder isn’t going anywhere [in favor of the new interface]?

Clark: That’s correct. And we think that’s important. Not just because like, oh that’s something we should do. But because we think it’s important. Those workflows that people have already gotten used to, and trained on, we need to support them. Eventually, they’ll find their way to the new value.

Kumparak: If I’m using that Finder interface, and the rest of my team is doing things in Dropbox Spaces.. how will I know that there might be stuff going on that I’m not seeing?

Clark: You’ll get some notifications that there’s more to see here. If you’re @ mentioned in a folder review, you’ll get notifications for that — so like “@greg, take a look at blah blah”, you’ll get the notification, you’ll go look in the folder, you’ll go look at blah blah. You go on the web, you’ll see those folder reviews there as well, so that might be a place you can choose to engage instead of changing your existing desktop experience.

So we’re kind of providing these multiple alleyways and avenues to get to things, but if you’re hardcore about “No, on my desktop, I just want Dropbox back there synchronizing”, we’re going to support that.

Kumparak: With everyone running 50 different apps for work, a main concern is efficiency; how much memory an app is taking up, how much CPU it’s eating up in the background. Is there much of a difference between “new” Dropbox and “old” Dropbox?

Clark: No. There isn’t. I’m actually very proud of the Dropbox engineering organization; I’ve only been here for two years, but I inherited a really talented engineering organization.  They’ve done a really good job there.

The modern desktop software takes advantage of a whole bunch of things — modern design paradigms, development frameworks, operating system capabilities… and these apps, over the years, relative to ten years ago, continue to consume the resources available to it. That’s pretty normal stuff.

The good news is that the underlying technology, in chip architecture, memory, operating systems, etc, provides a bounty of opportunity to do more for users. And users expect it! They expect high performance, and this real-time nature, and a UI provided in a modern way… they expect all of this, but none of it is free resource-wise.

But no, we’re on par with where we were before because most of the frameworks we had before to support the [old Dropbox] tray.

GettyImages 524746433

Image via Getty Images / TCmake_photo

Kumparak: With the new integrations — Trello, Slack, G Suite and what not — how much of that flows through Dropbox’s servers? How do you protect that data, and does it make you any more of a target?

Clark: When content, like if a Google Doc inside of Dropbox is accessed in any way, shape, or form, obviously, we see all that. If you have a channel connected up to a Dropbox space, and new content is created in that channel, obviously we see that content — we pick up on it, we drop it into a Space, we showed that in the demos. In so far as the Dropbox Space has been wired into these other applications, then yeah, we’re seeing the activity around those tools. Which is good! Because it means the search works, and the context is there. The promises we’re making around providing an experience that gives people an awareness of whats going on, getting people back into flow — it means we can make good on those promises.

That data is being managed the same way we manage user data. So it’s not finding its way to some different infrastructure, where we have to re-harden. Our very first cultural value at Dropbox is “Be worthy of trust” — it’s, like, the number one value of the company.

It’s a statement that started from engineering. “We have people’s stuff, do a good job!”. So we designed security, I kid you not, from the hardware architecture all the way up to the user experience. So those investments are valuable because they carry into all of the data we have, not just the file data.

Kumparak: And does bringing in that data make you any more of a target?

Clark: I don’t think so. It’s an interesting question. With all of the big SaaS services, there’s always sort of a constant… probing, that we’re always vigilant about. We have ways to keep on eye on this; are there different sorts of things that people are trying to knock on our door? So we have the ability to see that signal if it comes.

But we don’t anticipate that a lot of those things are actually of any more “value” [to a malicious party] than the actual content. What’s more valuable — the undisclosed financial statement, or the fact that the CFO and the CEO were having a side conversation about the wording of this paragraph? Probably the actual financial statement.

We’ve always had an incredibly high bar on protecting things, and I don’t think any of this stuff extends beyond that bar.

Kumparak: Can you tell me about the new computer vision stuff with regards to image search? [Note: the new Dropbox app lets users — currently business/pro users only — search their images for things it recognizes within the image, like clothing, or animals]

Clark: Yeah!  Image search is valuable because businesses need it. In the consumer world, of course, we like this. Type “Christmas tree, snow, dog, 2007” and like, okay, there’s a picture of the dog and a Christmas tree.

But in the business world, rich media is an increasingly constant part of how businesses operate, and there’s more and more and more of this content being created. Drones flying, taking pictures of oil fields. Industries like fashion, or media, etc — they’re obviously so visual content heavy.  So we got serious about this as our business customers started getting serious about it, and started asking us how they could do a good job with that.

So we’ve built technology … it’s state of the art, but the state of the art is understood. It’s no longer rocket science to create a machine learning kit that can understand object recognition; it’s engineering work now, not rocket science.

Kumparak: That computer vision — is it happening locally on device, or in the cloud?

Clark: It’s happening in the cloud.

Kumparak: Is there any human involvement in the object recognition?

Clark: No. Our training sets come from standard training sets that are available — you can go search on this, and find whole Wikipedia articles on visual object training sets. There’s some human involvement on those training sets, I’m sure, on tagging [training data]. But on the end user’s stuff, no.