Series E Terms for Twitter | Hacker News Reader

Series E Terms for Twitter

pedatacenter.com

23 pointsvcexperts14 comments

Out of curiousity, I tried to figure out the current capitalization table based on the listed values for price per share and amount raised for each round, along with a current valuation of $1B:

  round    # shares    fraction ownership
    A       14.6 M            23%
    B        8.2 M            13%
    C        8.4 M            13%
    D        8.3 M            13%
    E        6.3 M            10%
  non-VCs   16.7 M            27%

Why is running Twitter so expensive? What proportion of that money is going to towards engineering talent and servers?

Running Twitter might not be calamitously expensive, but positioning it and executing a plan to make it lucrative might be. For instance:

* It might require mainstream media promotion to push it into the mainstream market.

* Their plans to monetize it might involve M&A.

I agree on the second point but I think once you've hit the Oprah show, you qualify as mainstream.

Getting to the mainstream and staying there are two different line items on the budget.

It's realtime, there's a lot of users, and there's no follower limit, so Twitter tends more (than say facebook) to the O(n^2) case where everyone follows everyone.

>> "It's realtime"

The trend to label websites "Realtime" as if it's something amazing and new is really pretty funny.

The only real concrete feature I've heard about google wave is that you can see characters as users type them.

Did we just go back in time 20 years?

The technology twitter is simple, anyone could build it. You could likely run twitter on 10 servers max. The brand and community they've built is what is valuable.

"Realtime" is amazing and new, when it's in conjunction with web-scale data. Remember when Google used to take weeks to update its index?

> The technology twitter is simple, anyone could build it. You could likely run twitter on 10 servers max.

I'm not sure whether you're being disingenuous or naive.

I'm just stating a fact :/

It's a simple message passing service. It's pretty much as complex as a router that supports multicast UDP. eg not very.

Multicast UDP is so easy that it has never successfully been deployed at Internet scale.

Building a router that supports multicast UDP is simple. Getting THE INTERNET to all simultaneously upgrade is the hard bit.

Maintaining group-based routing tables for individual pieces of content as opposed to hosts that serve many pieces of content is a hard problem, and no vendor has solved it.

You are naive. It's a dead simple system.

If your company has an incredibly high valuation, it might make sense to take all the money people will give you at that valuation and figure it out later.

Reply on news.ycombinator.com