The Red Streaks Report

When Darius joined Dropbox, they were in the tens of millions in ARR and growing at record speed. And they were RAPIDLY iterating – so much so that they were pushing changes to their marketing site and to onboarding on a daily basis. Things were breaking, and it was taking the team sometimes weeks to diagnose and fix issues that were negatively impacting revenue. Darius immediately recognized this as an area to improve – to bolster revenue by playing great defense.

That’s where the Red Streaks Report came into play. It was designed to quickly and highlight anything that could have broken in production. The beauty of the Red Streaks Report is that it’s incredibly simple - something almost anyone can implement.

It started in a spreadsheet (of course :) ). Each column was a day, and each row was a different source of revenue traffic. For example, one source was “shared folder access” and another represented revenue customers being over quota (e.g. they needed to buy more storage) browser, traffic source. A cell in the table would go red if it was down from a certain % or $ amount from the prior days average (over the last 3 weeks). Then what they’d look for were streaks of red – typically defined as 3 cells red in a row – to indicate that there was a decent chance they’d broken something.

“If it’s red for four or even three days in a row (a red streak), probably something’s wrong and worthy of an investigation. That’s a lot faster than what we had previously… about two weeks.”

Sounds really simple, right? The core value is its power to drastically speed up the troubleshooting feedback loop, from red flag to resolution.

A Stronger, Faster QA Pipeline

It’s hard to overstate how much the red streaks report transformed Dropbox’s QA pipeline.

“Previously we just looked at total weekly revenue.” After two weeks of revenue dropping, they opened up the system to see what, if anything, was wrong.

You can see the difference in time and lost sales: identify a broken channel in three days, or start investigating a total revenue drop after two weeks. It not only saved time on identifying the problem, but it saved a lot of engineering hours by going straight to the impacted channel.  

Darius mentioned an interesting downstream effect of the Red Streaks Report: quicker, more confident product launches. Darius said the engineers loved it because people weren’t as scared of the knock on revenue that a glitch could cause a new product.

“[Before] We were a little more hesitant to launch stuff. Like, ‘Hey, if this thing’s got any kind of error in it, it could take down revenue, and we might spend months debugging.’”

Once the time to resolution looked more like days than months, potential issues weren’t as daunting. The result was a faster launch and a more agile response to any issues.

“If you fix it in three days, it’s probably not going to affect quarterly revenue.”

Quick Fixes, Steady Growth

Darius mentioned that while the team loved the Red Streaks Report and used it often, at first the company didn’t appreciate just how instrumental it would be. After a lukewarm reception, he just built it on his own. Only after a year of steady use, and plenty of bullets dodged, did leadership realize how much of an asset it was.

That speaks to an important pattern I’ve seen when speaking to many great growth leaders and analysts. Strengthening your current products may not be the sexiest game in growth, but it’s what the best companies do to stay ahead. Don’t try to outgrow your problems – fix them!  

I loved this talk, and Darius’ simple but transformative idea, because they speak to a key idea we’ve heard from other industry leaders like Ray Ko and Ed Park: strong growth comes from playing great defense. Create a strategic framework for addressing problems as they come up, revisit it often, and dedicate a team to fighting those issues quickly.