Why Twitter just raised $200m.

http://swombat.com/2010/12/15/why-twitter-raised-200-million-dollars

by swombat • 16 years ago

29 5 16 years ago

swombat.com

gordonbowman• 16 years ago

Great points. These funds will surely be used to turn on the money faucet.

Also interesting is the fact that they brought David Rosenblatt on to the board.

As peHUB notes,

"Also joining Twitter’s board is David Rosenblatt, who was long the CEO of DoubleClick, then chief of Google’s display ad sales for roughly a year after it acquired DoubleClick for $3.1 billion in 2008. (He left Google in April of last year.)

Certainly, though, for a company that looks to supercharge its revenues, bringing on the longtime CEO of one of the world’s biggest ad serving networks looks very smart."

http://www.pehub.com/90776/kleiner-muscles-its-way-into-twit...

ig1• 16 years ago

That's property prices in London for you...

kondro• 16 years ago

I would find it interesting if many mobile networks in the world that support many more users and many more messages per second will have even spent $460m on their SMS/MMS infrastructure.

I understand that Twitter is different, but 65 million tweets a day is only what, 2,000/second burst @ 3.2Mbps and 13GB per day of new tweets (4.7TB/year). And that is only their recent volume (2010-06-08: http://techcrunch.com/2010/06/08/twitter-190-million-users/).

I know that Twitter made the decision to put the data at the location of access (in this case, that makes it the end-users 'inbox') however, that means that the more popular a user gets (say Gaga with 7.3m followers), every tweet they make needs to be delivered 7.3m times and cause, say 1.5GB of data/traffic to be generated, especially when we're only talking about 5TB/year of new, raw data.

Surely a model more like Google's search infrastructure makes more sense. Store tweets in clusters. Send a search request for all tweets less than X-minutes old to the clusters, the search clusters each return their results in parallel to the caller who then merges & orders the results before returning to the user.

The clusters would replicate data based on popularity (e.g. Gaga may have 10 copies of her data throughout the cluster, compared to me with a couple dozen followers having just one). 5TB + replicas is more than enough to store in memory among a couple-hundred servers and the load generated on the total system would be based on those accessing data rather than producers with a large amount of followers tweeting.

To build this infrastructure to date they have raised $160m and, unless this $200m is completely unnecessary, they have spent most of it.

Anyway... I seem to have gone a little off-topic and gotten a little ranty however, the total amount of money being spent on a company with no long-term, sustainable business model just seems ridiculous. It's as though the VC's have already spent too much money on building a massive network that they need to keep propping it up so they don't look like fools.

wowfat• 16 years ago

" more popular a user gets (say Gaga with 7.3m followers), every tweet they make needs to be delivered 7.3m times"

- Do they really need to duplicate the same tweet 7.3 million times? The tweet cant be edited too!

helwr• 16 years ago

they will sell it, probably to Google

just need some money to survive a winter or two

Why Twitter just raised $200m.

Related Stories