So Leo Laporte, Jason Calacanis, Obama, Britney Spears, Shaq etc.. all have hundreds of thousands of followers, in some cases because they add to society and have interesting things to say, and in others because people often behave moronically towards decisions of who to follow.
I have 880 followers at the time I’m writing this; a very small number for some, but more than a lot and close to what I find to be a useful number of people to have receive my updates and to receive the updates from. This number is growing by ~30 people a day so in a week or so I’ll have over a thousand and in two months at this rate I’ll have over 2000 people; at which stage I may have to look at changing my policy of following back those who follow me.
I love twitter; it took me a while but I now find it to be an incredibly useful tool for answering random questions and getting advice. It was through Twitter I decided to see Transformers today instead of ‘The Hangover’; that yesterday I discovered why all Americans refer to the ‘#’ sign as ‘pound’ and how I arranged to how dinner with a fellow technology-entrepreneur in London last week.
This isn’t because I have 880 followers; this is because I have 880 english-speaking, non-bot, interesting, like-minded followers who are on Twitter for the same reasons as I am; to enrich, and be enriched. I have friends with thousands of followers who have a very poor twitter experience filled with repeats of bot posts, spam invites to increase twitter followers and hash-tag spamming. (e.g. Software MP3 DVD Movie and many more http://ow.ly/h47O +++#ihavetoadmitMark Webber Chuck Norris Michael Jackson Pirates Bruno Iran) – can you guess what the trending topics were at the time?
So, I have tried various Twitter tools. Twollow is actually very effective; if you leave even 5 keywords on the ’15 followers per batch’ setting, you will get new targeted followers, quickly. Or at least they appeared targeted; but there were a number of problems.
- At the beginning all Twitter users spoke English. Twitter is now a worldwide phenomenon that spans many countries and languages however it offers no translation features. Mix that with the fact that Apple iPhone is spelt the same way around the world, as is Microsoft Windows, ASP .Net etc… and they keywords you’re hitting match up, but their posts are in Spanish, French, German, Chinese etc… this isn’t useful!
- Keywords are spammed; iPhone is mentioned in lots of inappropriate tweets because it’s one of the most searched words, as is start-up and many social media terms.
- A post can be totally misrepresentative. For example I had one user that said “Just returned my iPhone, was awful” in a tweet. The rest of her tweets were about Makeup. In every sense she is someone who I probably have very little to say to (although she did seem to know a lot about Makeup!!).
So I’ve now stopped using most automated tools; I still auto-follow people who follow me, and hit a few very specific keywords, but I monitor that list and evaluate them. I also have a script that I’ll be releasing soon that matches trending topics at the time of tweets with the content of tweets. If all someone hits is Trending Topics then they’re probably a spammer, but even if not, then I can get it elsewhere; it is ‘Trending’ after all!
How do I gather followers? I write interesting tweets that often link to blog posts, interesting articles or images. It actually isn’t hard to gain Twitter followers; follow anyone that you are even vaguely interested in, and @reply to people who say interesting things and have lots of followers. For example, if I @reply to leolaporte, and someone searches for leolaporte they will see my tweet and if I’m saying interesting sensible things, they may follow me.
But why can’t this all be done automatically?
I believe that it can, but that none of the tools are there yet. I’m talking to a few people who are already in the ‘automating twitter’ space and am adding some things of my own and trying to build a tool that will completely automatically and sensibly manage your twitter followers.
There are a few points I want to hit:
- Number of followers they have (safety can usually be taken in the fact lots of other people follow them)
- Language (not just in one tweet, but in many)
- Context – a basic idea of the context of tweets – positivity, negativity (i.e. I hate my iPhone wouldn’t necessarily keyword match to iPhone)
- Groups – i.e. Social groups could have auto-follow groups that will add a user to everyone’s followers / followees every time a new user joins. e.g. a forum; all users could want to join a ‘Neowin’ twitter group, that will then link up all members.
- Recommendations – recommend users to follow to a user from those whom people they follow are following and on keywords they’re monitoring.
Obviously this is a complex problem and Twitter Karma and Twollow are doing a lot of good work; the main feature that I would like to see added quickly is language awareness so that will be the first additional feature offered by this service when it goes live in the next month or so. The others will follow soon after.
I’ll post again once I have a name and domain set up, looking for beta testers, or email ahead if you’re interested at firstname.lastname@example.org.
Let's talk about it.
Head on over to Twitter and let's talk more.
Find me at @chrismaddern.