Making Job Search Easier by Finding the Great Companies First

Find a
Title/Key­words Com­pa­ny Name
Where Search
City, state or zip (option­al)
City, state or zip (option­al)
Job title, key­words Com­pa­ny Name Only


How cloud storage startup Dropbox engineered its way from 2000 users to 175 million

Here’s a ques­tion that the more for­ward think­ing prod­uct man­ag­er would like an answer to: how does a tech start­up scale up from 2,000 users to 175 mil­lion in just five years? As recent news reveals, that’s exact­ly what Drop­box has done – and its for­mer head of ser­vice engi­neer­ing, Rajiv Eran­ki, has been explain­ing the engi­neer­ing that lay behind its explo­sive success.

A hap­haz­ard learn­ing curve

Eran­ki joined the fledg­ling Drop­box team in 2008. That was in the days when it had just 2,000 users, and things were dis­tinct­ly glitchy. His job was to help scale up the plat­form, and he had just one full-time col­league to help him.

Reas­sur­ing­ly for prod­uct man­agers look­ing after a new tech offer­ing, in the begin­ning, Drop­box was, as Eran­ki puts it, “bug­gy” and it need­ed to have “joins” across its data­bas­es sep­a­rat­ed. But Eran­ki found the hap­haz­ard, inef­fi­cient efforts of the ear­ly team to be ulti­mate­ly high­ly pro­duc­tive, because he thinks it led to some key engi­neer­ing ben­e­fits that wouldn’t have been found otherwise.

Joins across data­bas­es were easy to accom­plish, as it hap­pens, and the Drop­box team did them as and when they need­ed to. And because they could put queries read­i­ly into MySQL, bug fix­ing was actu­al­ly a breeze. And even if a prod­uct man­ag­er might be skep­ti­cal that a soli­tary data­base machine and a soli­tary front-end serv­er could facil­i­tate scal­ing upwards, the team would dis­agree: it meant they only had one log to study.

Toward 1bn?

Com­bined, Eran­ki says, these fac­tors allowed for “tremen­dous flex­i­bil­i­ty and scal­a­bil­i­ty” and a good deal of learn­ing to boot. Like, for exam­ple, find­ing out that they could use Python for every­thing: even when Drop­box had acquired its first mil­lion users, Python ensured that the plat­form would run on just a few hun­dred lines of code. As Eran­ki puts it, Python meant “we could get to 40m users with­out hav­ing to write thou­sands of lines of C code.”

When Eran­ki and his col­league ana­lyzed what Dropbox’s real core users were doing with it, they worked out how the plat­form could be devel­oped into a real mon­ey-mak­ing busi­ness. And he now believes it’s pos­si­ble to scale up from 175 mil­lion users to a bil­lion with­out any difficulty.

That silence you’re hear­ing is the sound of a prod­uct man­ag­er being impressed.

Print Friendly, PDF & Email