Am I talking to myself? TweetReach may have the answer

I’ve been rather busy lately moving house, hence the lack of posts and even a reduction in my Twitter activity. To get back in the swing of things a quick post about TweetReach.

Stephen Waddington highlighted this tool to me at the North East CIPR Awards on Friday (congrats to all the winners by the way).  It is pretty straightforward to use.  Put in a term, a url or hashtag and it calculates the following three measures:

Reach – the number of different people who follow people who have talked about the search in some way.

Exposure – how many times someone could have seen a particular reference to the topic e.g. if 4 people have tweeted it who all have a follower in common then that follower will have been exposed to it 4 times but will only count as 1 in the reach figures.

Impressions – the total number of occasions to see (effectively the sum of the “Exposure” figures for everyone “Reach”ed)

Note that the Reach figure measures the number of different people. Assuming this is the case (and I have no practical way of checking this) then this is good stuff as it ensures that duplication in networks is taken care of. It also tells you what proportion of the relevant tweets were retweets or @replies.

So a very good tool in theory for understanding the extent and likely penetration of a conversation. However unfortunately the bad news is that it is limited to the last 50 tweets, unless you pay TweetReach $20 and even then it is limited by Twitter’s API to the last seven days or 1,500 tweets. The seven day limitation also means that you MUST carry out the analysis close to the time of the relevant conversation as you can’t go back historically. These factors weaken its role as a measurement tool significantly unfortunately IMHO.

Perhaps now that Google are indexing our tweets the tool could be expanded?

6 thoughts on “Am I talking to myself? TweetReach may have the answer

  1. I think it might come in useful if you’re doing ongoing monitoring though. I.e. if you can measure the reach as you go.

    Long time reader, first time commenter. Thanks for updating the blogroll. :)

  2. Agreed it might well be useful for ongoing monitoring, however the 50 tweet limit without payment is a bit of an issue even for that purpose. Even my little old blog has had a couple of posts tweeted a lot more than 50 times nevermind those of the great and the good such as yourself :-) But yes it could certainly provide an ad hoc indication of impact.

    Oh and by the way this isn’t your first comment Oh how quick they forget me… :-)

  3. Adam,

    I’m the founder of Appozite – the company responsible for TweetReach. I appreciate the in-depth post. Just wanted to drop you a quick comment to let you know that the reach number is indeed a *unique* number measured by counting each follower of each twitterer that sends a tweet only once.

    As for the 1500 tweet/7 day limit, that’s a frustration for us as well. For customers that want to measure results over a longer period, we tell them to run a report a week. It’s not optimal but it allows us to get around those Twitter-imposed limits.

    Feel free to drop me a line (@hayesdavis) if you have any questions or additional comments.


  4. Hi Hayes

    Thanks for stopping by. Take your point about running weekly reports and realise that the limit isn’t your doing, but this does mean that for each search a user wants to analyse they are looking at $80 a month (if I have my sums right) and I suspect that level of cost per term might be out of the reach of many. Perhaps you could offer some sort of deal whereby users paid the $20 fee once and paid a lower rate for future queries on the *same* term/url/hashtag?

    Nice tool though, despite the API limitation, and the deduplication of followers must take a bit processing? :-)

    Good luck with it.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>