My agency, Receptional, has been involved in some interesting projects over the years. It seems that when other agencies or consultancies are asked a question that just cannot be answered without a bit of clever programming, they bow out of the race. We find ourselves building ad-hoc tools for clients which do not cost the earth, but produce interesting data designed exclusively to answer the question at hand.
We think that this is something we would like to get involved in further. It’s interesting stuff. To that end, I am going to want to really start looking at technologies which “scientifically” try to make predictions and forecasts in new ways.
To that end, I will be starting a new blog soon – called PredictionTech.com and the first post is nearly ready to go. The point of the blog will be to report on new predictive analytics technologies and if you want to be in RIGHT AT THE START then you could do worse than following @PredictionTech on Twitter. If you have technologies that you think PredictionTech should be reporting on and investigating, then send them through via that Twitter profile for consideration. If you want to write for PredictionTech (It’s unpaid, but may get you to some cool conferences) then also let us know.
Do you think the project will have legs? What would you like PredictionTech to look at?
Gab · 25th September 2011 at 3:24 pm
I think a tool that could accurately predict your CTR from Twitter would be extremely valuable for a Twitter ad product. Currently it’s somewhat of a crapshoot, with historical data giving some indication, but pricing being arbitrary relative to that and inflated follower counts obscuring things…
admin · 26th September 2011 at 8:24 pm
Yeh – nice idea – although “Click Through Rate is not really a three letter abbreviation that I would use, because one Tweet doesn’t really give me an idea how many people see the tweet. So how to proxy this idea? “Engagement” is a similar metric, which I think people like Klout are trying to measure. The challenge with Twitter is access (or not) to the full Firehose data. Reputably $1 Million a year and that’s just the start because you need to handle the data. Without that, it would be reasonably easy to work out your OWN CTR from Twitter