Internal status email: Twitter at 3 months old. 160 Users & 100 Tweets per day | Hacker News Reader

Internal status email: Twitter at 3 months old. 160 Users & 100 Tweets per day

anarchogeek.com

104 pointsrabble49 comments

This is remarkable, even to me. The main point: if you're doing a startup, things take a long time. Longer than you'd expect. And overnight success is anything but.

That's actually kind of a relief, especially considering all the news you hear about 10000 users in 72 hours, it seems like if you don't explode onto the scene you're a failure.

Personally 160 users sounds like a lot to me. None of the unfunded services I've bootstrapped had reached that.

Mind you, these aren't paying clients or anything, just a 160 twitter accounts. At the time you had to have it created FROM a mobile phone, so the onboarding process was clunky. We were showing it to our friends, tell them how cool it was. A team of 10, with 3 months to pressure their friends and family in to using the service, got 160 users.

"Timing, perseverance, and ten years of trying will eventually make you look like an overnight success." -Biz Stone (co-founder of Twitter)

http://www.bizstone.com/2010/09/timing-lessons.html

Unless your startup is Facebook.

Does anybody have any information about the types of things that twitter did at first to market themselves?

This is a problem for me doing thingist (http://thingist.com)... the people who have found and signed up for it seem to have really liked it, but I'm having trouble figuring out how to get more of those people.

(And I am an admittedly horrible marketing person)

Care to provide a bit more on thingist? I clicked through and at first glance have a few ideas of what it is but not totally sure what I would use it for. Without an about page or any precursor I would not sign up for something on a whim. Interested in checking it out though!

Also side note: The bottom pagination links do not work for me. I can click next page and get next page but if I select a number be it 2, 3, or 75 I get an empty page?

Sure :)

Thingist is kindof like status updates, but with more context (and more available space [You can include a description of why you're posting it). Instead of a big bucket called "status", you separate things into different lists.

For instance, I have a list called "Good Bars in Phoenix" (http://thingist.com/t/list/65/)

or a list called "Adventures in Nerdery", which is mostly projects I'm working on (http://thingist.com/t/list/139/)

There are some global lists like "songs I like" (which get added to people by default when they register) (http://thingist.com/t/global/38/)

The example for a use case I always use is finding music that you've posted into the internet. I've been posing songs I like into facebook for 5 years, but as far as I know there isn't a simple way of pulling them back out.

Thingist [tries to] solve that.

But it's more than just status, at least it's a new approach to it. Somebody joined a few days ago, and started posting "things on the menu at a nerd themed restaurant". This resulted in neat sort-of back and forth between us: ( http://thingist.com/t/global/2944/)

(also: fixed the pagination buttons, oops)

How about telling us (and everyone) what thingist helps us with. That'd help!

This is an excellent point. "Create an about page" is at the top of my list of things to do as soon as I get off of work.

I made this video: http://www.youtube.com/watch?v=i1rOJJWyU2s

but it was wayyyy too long.

No – don't make an about page. Well sure, make one, but it probably won't help a great deal.

You need to be showing people what it's about, the instant they visit your site.

I saw your page, thought "hmm, not quite sure what this is" and left. I'm not going to stick around to read an about page.

How about a list of reasons why thingist is awesome, or a list of what it can do that will solve or fill some need, promoted to the front page?

If I were you I'd add a facebook like button to each thing and allow people to create lists of things they love just by clicking the facebook like button (no registration required, just save their list to a cookie).

Then if they signup they can save their list. This makes it easy for users to jump right it in, and may bring in a lot of traffic from facebook. Of course, some users may not want to share each thing they like on facebook, so you'd probably want a normal "love this" button as well for logged in users.

There are like buttons on the item pages... (like here: http://thingist.com/t/item/4720/) I tried putting them in the feed, but they looked really out of place (it made things look really really cluttered)

I really like the idea of [optionally] importing facebook likes... Hmm

(As a sidenote: can I just point out that facebook's API is freaking awesome?)

> (As a sidenote: can I just point out that facebook's API is freaking awesome?)

No you may not. Facebooks API is horrible. I was just implementing the like buttons on my site and realized that they have to be iframes (even the FBXML or whatever its called creates iframes). This turns out a bit ridiculous on pages such as yours (if you put them on the feed page) and mine where you Twitter on the other hand allows you to just gives you a URL which points users to a link to share the page.

Well, you could write your own, but you'd have to have users authenticate the application. In my opinion, this isn't much different than twitter, since both actions would require two clicks:

facebook: click your custom widget -> facebook prompt -> perform action

twitter: click your custom widget -> twitter prompt -> perform action

(Or just do like I do, and change the size of the iframe to 450x20)

as far as i'm aware, you can't really write your own. I was watching the network traffic for a submit and it was pretty ridiculous how many transactions it did when you clicked "like" (about 4 if i remember correctly). Having the users authenticate the action would actually be good for sites like mine but facebook doesn't seem to provide that option.

Currently you're in a bit of a chicken-and-egg situation, where new visitors probably need to see more fresh content to get interested, but your content only comes from having lots of visitors. If you can bootstrap the content, you may find that user acquisition gets easier.

It seems a little bit like you're trying to create a more structured version of what people are already doing on Twitter with many of the trending hashtags. (e.g. today's #failedmcdonaldsproducts[1] or #100factsaboutme[2]) Rather than fight the momentum Twitter already has there, perhaps you could take advantage of the Twitter API to automatically pull these trending topics in as Thingist Lists. When a new list is auto-created from a trending topic, you can automatically tweet out a link to the Thingist list with the same hashtag, which will get you exposure among people following that term.

The hard part then is making sure Thingist provides a more compelling experience than Twitter for browsing these sets of things. I'm not sure if you're there yet, but there is definitely an opportunity to do something better than search-by-hashtag. If you do pull this off, you'll have plenty of fresh content to keep new users engaged until they start creating lists right on Thingist.

[1] http://twitter.com/#search?q=%23failedmcdonaldsproducts

[2] http://twitter.com/#search?q=%23100factsaboutme

"We receive about 100 status updates per day". 5 years later, the average number of tweets per day is 140 MILLION.

(According to http://blog.twitter.com/2011/03/numbers.html)

Loved that! It's great to see that even great big things start small.

Google Cache Link : http://webcache.googleusercontent.com/search?sclient=psy&...

How did you find that? I found the original article in Google but there was no link to a cached version.

Take the URL you're trying to pull up and search for

cache:<target URL here>

In Firefox, you can also bookmark the URL

http://www.google.com/search?q=cache:%s

Then give it the keyword 'cache'.

From now on, from the URL bar, you can just type cache <whatever> and it will go straight to a search for the cache of that.

Very cool! Thanks for the replies, good to know.

I'll be the first to say... Thank you for putting the vowels back in your name. "Twttr" looks horrible.

The name has always been "Twitter," but they were wanting to get the SMS short code "twttr." Unfortunately for them (fortunately for us?) someone already had it. :)

It was always PRONOUNCED twitter, but we did write it twttr, everywhere, back in the day.

Right, that's what I meant. :)

Yes, the twatter thing, wasn't something we thought of, but many other people had it pop right in to their heads. It was, unfortunate.

but be honest, how many people did you show that to that said, "what's Twatter?"

They probably couldn't afford twitter.com back in the day.

Ev had a bundle of pre-IPO google stock, so he could afford it. Also Odeo stil had cash in the bank, but he didn't want to buy the vowels until he was sure there was something to the idea.

Obvious was supposed to be a company which hacked out many ideas, a kind of incubator. The problem was one of the first batches of ideas turned in to a major run away success.

Link is dead or down, I can't wait to read this!

Not sure what happened, sshing in to the server and poking around.

The subject line is so tempting, I upvoted while waiting for you:)

Thanks! G'Luck

The hard drive filled up and mysql corrupted a table. It's fixed now.

Probably your Web server OOMing the box? Check dmesg.

Apache-specific advice: if you're on a small VPS or other box with not much RAM, and didn't tweak the distribution defaults for MaxClients and friends in its configuration, a HN-size load will OOM you rapidly, especially with mod_php. The defaults for prefork MPM are nuts in the major distros (as in, assuming 4 GB of RAM or more for mod_php).

Since you're going up and down, I assume this is the case. I suggest dropping MaxClients to 5 for now and letting us wait a bit.

Yeah, i thought i had it back up, went in to a meeting, came back and it's down. When i went in and looked at it there were a ton of apache2 processes. It's my blog server, so i don't pay a lot of attention to it, i probably should.

I dropped down the number of preforked apache processes, we'll see if that helps. It's also where i run some rails experiments, so we'll see. Load's back up to 7...

I had no idea hacker news got a lot of traffic.

Having just launched Momentomail and watching our userbase slowly grow despite quite a bit of effort to get coverage, it's been a struggle to keep positive.

This is a great story that put a bunch of hot air in our balloon tonight as we realized we're actually not too far off this track.

We have no expectation of being another Twitter by any possible stretch of the imagination, but finding out that at this early stage we're tracking with a reasonable expectation is a great point of validation.

"Older" hackers take note: Biz Stone, Evan Williams: 30+ founders club. They were both over thirty when twitter started.

Ev started his first company when he was a teen, and kept starting them over and over again. Biz had founded Xanga, many years before. Sure, they were over 30, i think Jack and Noah were too, when twitter got started. All of them had started earlier companies in their teen's and 20's.

Jobs also started NeXT when he was 30 after founding another well known company years before. My point is age shouldn't be a barrier to starting a company you believe in, especially not the imaginary "30" line.

Interesting how he signs of with "keep twttring". I that email was written today it would have been "keep tweeting".

Nobody liked the word tweet. It was a term created by the user community, and only slowly accepted by the team.

For nostalgia's sake (if it's suitable to reveal), what was the original number that Twitter used?

The first email i've got from jack, saying he's got a prototype running on his laptop is from March 14th 2006. He called the app "twttr stalker" in that email. I remember it being "friend stalker" as a joke name thrown around as well.

Reply on news.ycombinator.com