Is FollowFriday even Valuable?

I’ve been tracking FollowFriday for a while now and using justSignal to power FollowFridays.com – and today it really, really looked like FollowFriday had jumped the shark.

Here’s the thing. If you compose your FollowFriday tweets like:

#followfriday @joe @mary @steve @beth

You really aren’t adding any value to FollowFriday – as a matter of fact I’d argue you are just creating noise. Here is why. The simple reality is that I only follow about 10 or 15 people who I know well enough, trust enough, and have enough of a complete relationship that I would simply follow someone they told me to. And odds are I’ve been following those people for some time now and follow nearly everyone they would recommend.

So your @username missle for FollowFriday has no effect on me. I’m not following them… my guess is very few others will either.

But, if you compose your FollowFriday tweet like:

#followfriday Follow @micah because he came up with this and @strebel because he has mad design skills

I can determine if I want to follow those people based on WHY you follow them… not just because you said so. Even more powerful is the ability to head over to FollowFridays.com and enter a filter on the Tweet Stream (click on “filter this content”) for “cool” or “design” or whatever it is you are looking for and see JUST the FollowFriday Tweets with that word in them.

Today, based on what I was seeing and hearing it appeared that the majority of the tweets for FollowFriday looked like the first example. If that were the case – I’d have to say FollowFriday had outlived it’s usefulness.

Conveniently – because we us justSignal to track FollowFriday – we have access to each and every tweet sent about FollowFriday. So I decided to rely on data… and here it is:

NOTE: Tweets of Value is defined as any tweet that does not contain all @names and hash tags. Raw hour by hour data posted after the jump.

As you can see – while a significant number of the tweets (about 20% overall) were just hash tags and @names the vast majority actually contained useful words (hopefully) describing why we should follow the people being recommended.

I’ll grant you that we did not perform any kind of semantic analysis on these tweets trying to determine intent to state why someone should be followed, but I’m still pleasantly surprised that a consistent 80% of the tweets were not all @names and hash tags.

Make sure you tell everyone – only do FollowFriday recommendations with reasons… it is much more effective and keeps FollowFriday valuable.

One other thing I’ll take a moment to mention (shameless plug) – How cool is it that you can think of something interesting to discover from Social Media data and immediately be able to go answer that question? That is part of the power of justSignal… check it out.

Raw Hour by Hour Data:

Hour: 0
There were 812 tweets that only contained hash tags or @ names.
There were 2962 tweets that contained actual words.
78.4843667197 of the tweets were valuable as recommendations.

Hour: 1
There were 843 tweets that only contained hash tags or @ names.
There were 3050 tweets that contained actual words.
78.3457487799 of the tweets were valuable as recommendations.

Hour: 2
There were 824 tweets that only contained hash tags or @ names.
There were 3327 tweets that contained actual words.
80.1493615996 of the tweets were valuable as recommendations.

Hour: 3
There were 918 tweets that only contained hash tags or @ names.
There were 3962 tweets that contained actual words.
81.1885245902 of the tweets were valuable as recommendations.

Hour: 4
There were 1385 tweets that only contained hash tags or @ names.
There were 5753 tweets that contained actual words.
80.596805828 of the tweets were valuable as recommendations.

Hour: 5
There were 1939 tweets that only contained hash tags or @ names.
There were 7805 tweets that contained actual words.
80.1005747126 of the tweets were valuable as recommendations.

Hour: 6
There were 2536 tweets that only contained hash tags or @ names.
There were 11982 tweets that contained actual words.
82.5320292051 of the tweets were valuable as recommendations.

Hour: 7
There were 2874 tweets that only contained hash tags or @ names.
There were 14391 tweets that contained actual words.
83.3536055604 of the tweets were valuable as recommendations.

Hour: 8
There were 2714 tweets that only contained hash tags or @ names.
There were 14701 tweets that contained actual words.
84.415733563 of the tweets were valuable as recommendations.

Hour: 9
There were 2394 tweets that only contained hash tags or @ names.
There were 13266 tweets that contained actual words.
84.7126436782 of the tweets were valuable as recommendations.

Hour: 10
There were 2322 tweets that only contained hash tags or @ names.
There were 13896 tweets that contained actual words.
85.6825749168 of the tweets were valuable as recommendations.

Hour: 11
There were 1940 tweets that only contained hash tags or @ names.
There were 11279 tweets that contained actual words.
85.3241546259 of the tweets were valuable as recommendations.