The personal ramblings of Jon Carlos mobile & web developer and technology enthusiast.

Tuesday, 7 June 2011

Anotomy of a twitter experiment

00:12 Posted by Webmonger 1 comment
This weekend I spent quite a bit of time reading. I found some new RSS feeds and like the good little reader I started flicking through them picking out stuff I found interesting. Gradually as I did I began to think some of this stuff would be interesting to my twitter followers so I started to line up tweets for posting and started releasing them about 4 or 5 and hour.

I set this up to run for a couple of hours and then once they had completed I read some more and set-up some more tweets to send out.

To be honest it really just started as a test of TweetDecks Scheduled tweet function. I'd never used it before and wanted to see how it worked but after a little while I wanted to test a couple of things out.

1. How many clicks could I generate through to things I posted?
2. How many followers would follow my tweets and what would the quality be like?
3. How many posts could I make in 2 days?
4. Would people RT and Fav the tweets I sent out?

Sorry no snazzy infographics here just numbers but I hope you find the numbers alone interesting.

1. Clicks. I use Pro and have my own domain set-up so this was reasonably easy to check.
  • Saturday 99 Clicks
  • Sunday 176 Clicks
  • Monday 92 Clicks
  • Total: 367
2.a. Followers. I used TwitterCounter for these details
  • Friday 390
  • Saturday 397: +7
  • Sunday 401: +4
  • Monday 407: +6
  • Total Growth: +17
2.b. Quality. This is a difficult one I've guessed at the quality of the followers by checking their profiles.
  • 8 seem to be Human / Companies I mentioned
  • 9 bots / hash tag followers
47% of followers I would class a quality. This was a little less than I had previously hoped but I would say not a bad number and I'm sure as you post more interesting things in more specific areas this number would grow after the initial bot on slaught.

3. How many Tweets. I've excluded some all the non link out tweets.
  • Total: 56
4. RT and Fave. I used FavStar for this
  • Fave: 17
  • RT: 10
So I would say all in all I think it went very well. I got more followers in a short space of time that i have ever before. I got more Fav posts than ever before and more RTs in such a short space of time than ever before.

I tweet more than most but no where near as often as some people and to be honest just the shear number of articles I read over the weekend took it out of me let alone trying to write something interesting about all the links I was posting.

As I discussed on twitter earlier today I think more focused and much less is the way forward but it was interesting to see how it went for a short space of time.

Can you annoy people and loose followers? Yes! Again earlier today and after all the tweets had been sent I asked what people thought of my experiment. Here are some of the responses.

Me: I spent the weekend posting a loads of links to stuff I found interesting. Did everyone enjoy/find it interesting or should I stop?

: It was interesting but overwhelming and very wide-ranging. I would have to unfollow you if you keep up that rate

: +1

: agree with . Was interesting, but a little too much at once. Find a focus, and just plan to tweet less but regular.

: stop. You came across like a spam bot. Although you gain some followers are they quality ones. Nice experiment though.

So there we have it. Don't piss off people you like. Don't Tweet like a spam bot. Keep your tweets in context or at least focused.

Last but not least
It seems the most interesting post I made was about user testing. I guess everyone else has the same problem with this one.

: Great list of UX Testing tools #usability #usertesting #cardsorting 

8 x Fav
6 x RT
59 Clicks 


Webmonger said...

Yeah I believe the socialization of the tweets really does matter when posting these things. I'd not really thought of time as such an important thing but I guess you're right.

During the experiment I was very aware that more people would see what was being posted when the US was awake. I don't have the number break down for the times of day unfortunately.

I'll continue to experiment and if I find anything else that's interesting I'll do an update.