Hacker News Reader

Comment by dbg31415 | Hacker News Reader

How many people does it really take to run Twitter?

If they need more than 100 people I'd be really shocked.

It's hard for me to understand how Twitter has so many employees when:

a) The site has relatively few major features beyond displaying lots of text in chronological order. b) The site is so overwhelmingly shitty.

They had a great thing going for themselves years back when developers were really excited about building on the Twitter API and experimenting with it in all kinds of ways and creating great new apps daily. That stopped when Twitter started running wild and harassing devs and making investment into the Twitter ecosystem practically impossible. They could have kept their costs down and just needed a small number of employees to keep the infrastructure up.

I'd wager a bit more than that to maintain existing services. But they haven't given up on growth yet, disappointing news doesn't = throw in the towel. If these rumors are true, best of luck to those affected.

Even to grow, does twitter really need 3,800 employees? The site hasn't exactly seen much change in the past few years.

Jonathan Blow gave a talk to Berekley college students a few months ago, complaining about the poor quality in modern software.

He picks on Twitter in one part, showing the graph of the number of employees by year, challenges their claims to the complexity of implementing Twitter, and mocks a UI problem he just had.

Twitter starts at 7 minutes: https://www.youtube.com/watch?v=k56wra39lwA

http://i.imgur.com/U4mHId3.png

Thanks. Good video.

How do you even grow like that? The current staff during that spike must have spent most of their time interviewing people.

That spike is misleading because the chart jumps from "Jan '11" to "Dec '13", 23 months. The other columns are all 12 months apart.

still, 100+ joining every month, that surely was not an easy task by any means.

But 3,750? Does anyone think Twitter needs that many people? Scale back to the bare essentials and regroup. Clearly the people there haven't been good for ambitious growth, time to shake it up, find a buyer (which will be easier with less overhead), or close the doors.

they should focus on being THE platform for key influencers, "thought leaders" and celebrities, mover-and-shaker types, kind of like Yahoo did with Katie Couric!

And how many people do you need to monetize Twitter?

Well, either a lot more or a lot less.

Whats app comes to mind as a comparison. Far more complex. A billion customers. 35 engineers.

Nah, WA started as a 1:1 direct messaging service with next to no server storage - so all the backend had to do was to distribute the incoming messages to the recipient ASAP in order to keep storage load low.

Twitter's backend is far more complex, you have a 1:n distribution (with n occasionally drifting into the double-digit million counts, like President Obama), and all the messages have to be archived and live searchable.

Agreed. The keys to making whatsapp work were tuning an XMPP server for and the Erlang VM for performance, which takes a few very smart people, but not a large quantity.

And targeted ad sales, the complexity of which should not be underestimated.

Reply on news.ycombinator.com