The short answer? between 1 and 4PM and between 7 and 10PM.
Twitter analytics is a game changer in many ways. If you are looking to increase visibility of your tweets (or brand), the dashboard gives very interesting information to help you optimize your tweeting time (assuming you want to increase the number of people who view your tweets), target your tweets effectively (with audience analytic reports), or even measure how effective your twitter campaigns are.
Now that I have a new blog (and everyone need to know * insert appropriate smiley face here* ). I wanted to know at what time I should be tweeting to maximize the number of people who will see/read my posts. Obviously, there are other ways of improving visibility, including #hashtag hijacking, URL baiting, mentioning etc (which work). These methods seem unorthodox for me (plus, I think they are counter productive in the long run). So my immediate solution was to find out the best time to tweet; meaning the time in which I am most likely to get a higher number of people viewing my tweets (and hopefully reading the post). I could then use these time windows to send my “new post” tweets.
Now, I needed data to do this. I am not a regular tweeter so my data was very small, and so I asked @blackorwa and @princelySid to help out. The data used was a three month data covering period from September to December period. Combined, this was data about 1892 tweets. A few precautions I made was, first, to exclude all tweets in which there were #hashtags and mentions/replies (they tend skew impression and also they are driven by demand (not random).
My final prepared data looked something like this:
Because the data was from three different people, I first needed to standardize the impressions across all the contributors. Since your tweets are only presented on timelines of your followers, I used the followers to weigh the effect impression (i.e divide impressions by the number followers). The value of this new measure represents the number of times a tweet reaches each follower. Lower values means very few of your followers have seen the tweet. Higher values shows a good number have seen your tweet.
Now, I needed to know two things. What time should I tweet. And second, what is the size of the effect, compared to other times.
The chart below shows when I should tweet.
You can get the code from my github account.
See! three pence talk!