Twitter, Buffer and Patterns of Attention

I’ve been a bad digital marketer, I must confess. Though I have been working on SEO initiatives for clients for nearly 4 years, I only recently started using Twitter to try to do some proactive reputation management, and ensure that the Search Engine Results Pages (SERPs) show the picture I want to show when someone Google searches my name. I’ve been using Twitter to run social media campaigns for clients for years- even before we had any analytics at all and had to guess at the ROI of social media- but just never took the plunge personally, until recently. I’ve been using Twitter with a nice little free app/Chrome extension called Buffer (http://www.bufferapp.com) which allows you to trickle out your tweets in well, a buffer, so that you don’t bomb your followers with 15 links back to back in their Tweet streams, as you read your RSS feeds over coffee.  The nice thing about Buffer is that you also get statistics about how many times your links were clicked and re-tweeted- this got me to thinking about analyzing some Twitter data to see what effect Tweeting on different days of the week, at different times of day, using different categories and hash tags and personalizing my Tweets would have. The data is interesting and a little surprising, too.

First, here’s a little background about how I conducted this (admittedly non-scientific) test. Here’s my Twitter biography:

SEO Account Manager @WebMetro, where I work to build value for clients through Internet Marketing. I’m also a Daddy, Disneyphile, WWE fan and avid PC gamer. http://franirwin.com

So, this is definitely a public, promotional Twitter account and I watch what I say on here. Very seldom will I curse, I’m quick to thank people for re-tweets, and I keep things industry/work related and squeaky clean. The main reason I have this account is to drive traffic to my blog and also maintain an active social media presence for services like Klout, which gives me cool perks for sharing my opinion, and of course, just to keep up with the news for myself! I frequently am asked “how do you SEO’s keep up with everything that’s going on, especially with Google changing the algorithm so frequently”? The answer is, with Twitter and with a good RSS reader (Google Reader, in fact).

The period of time we’ll analyze runs from September 1, 2011 through September 30, 2011. Now, sometimes, I will actually use the Twitter interface to manually tweet about things and send messages- those types of interactions aren’t covered here. This piece refers solely to tweets scheduled with bufferapp to go out during business hours on weekdays. I do this purposefully- Tweeting is not something I want to do full time or even in my free time; so I have a pattern set up with Buffer to start tweeting at 5 AM PST- this is for my West Coast followers, so they have Tweets to read as they come in to work- and ending at a little after 5 PM PST, at the end of my workday. I tweet approximately once an hour, with a higher concentration around the beginning, middle, and end of the day. Sometimes, I will be at home reading something that seems worth Tweeting but, if it seems like it will “keep” to the next scheduled Buffer trickle time, I will just add it to my buffer, which has a max capacity of 10 scheduled Tweets. I’m too cheap to pop for the full version J (Actually, the free version works just fine for me; though you may want to pay the very inexpensive fee to get unlimited space in your buffer.

I mainly tweet about SEO related stuff, lots of Google, Facebook, Twitter and Klout posts, with some personal stuff about my hobbies- namely going to Disneyland and playing PC video games- sprinkled in. It’s about 90% “professional” Tweets though- I am mainly cultivating others’ content. I try very hard not to put things out there that I’ve already seen- for that reason, I’m always looking for the next obscure RSS feed loaded with great information to source. I don’t have a lot of followers, but I did gain about a 30% increase for the time period I’m discussing here.

Let’s summarize my Tweet habits, shall we? First, I sorted my posts as granularly as I could without being ridiculous, into the following categories (followed by the number of Tweets and total % of Tweets for the topic, in September):

Google

44

18.26%

Blog

3

1.24%

SEO

33

13.69%

Foursquare

3

1.24%

Industry News

27

11.20%

Hulu

3

1.24%

FaceBook

21

8.71%

Netflix

3

1.24%

Gaming

12

4.98%

Amazon

2

0.83%

Social Media

11

4.56%

Yahoo

2

0.83%

Twitter

11

4.56%

Advertising

1

0.41%

Electronic marketing

10

4.15%

Bing

1

0.41%

Self-Promotional

10

4.15%

Business

1

0.41%

Personal Interest

8

3.32%

Microsoft

1

0.41%

YouTube

7

2.90%

Nestle

1

0.41%

Technology

6

2.49%

Panda

1

0.41%

Video

5

2.07%

Politics

1

0.41%

Deal

4

1.66%

Sports

1

0.41%

Groupon

4

1.66%

We’re looking at a pool of 241 Tweets that got a total of 907 clicks and 3 re-tweets, an average of 3.73 clicks per Tweet and about a 1% re-tweet percentage. The raw numbers themselves don’t look great, but we’re going to dive a little bit deeper and look at what posts scored “applause”, meaning they somehow got attention from my followers beyond my adding them to my Twitter. I’m glad to see that most of my posts are on topic, with “Google”, “SEO”, and “Industry News” speaking for 42% of my total Tweets. Looks like I’m doing a good job, staying on topic.

For the rest of this post, Clicks will be the main metric by which I measure engagement. I’m making the assumption that if someone clicks a link, they are interested in what lives behind it. Here are the times when the most clicked Tweets were…tweeted:

2:20 PM

92

5:11 AM

91

11:42 AM

90

4:17 PM

89

5:09 PM

58

7:05 AM

49

12:47 PM

44

1:05 PM

33

5:11 AM

33

9:12 AM

16

Remember when looking at these numbers that I am on the West Coast of the USA and a lot of my followers who actually KNOW me personally are on the East Coast. There doesn’t appear to be much rhyme or reason to correlating the time of the post with its popularity, but we can see that I’m getting my heaviest concentration of clicks in the times around the beginning and end of work, and lunch breaks. This makes sense on its face because that’s when most people have some time to spare for social media, running down their Twitter stream in the elevator on the way to lunch or while standing in line for a latte. I have timed my Tweets specifically to try and hit these lulls in activity, and that’s also why I don’t Tweet on weekends. High rates of adoption aside, I just don’t know a lot of people that turn to Twitter to kill 5 minutes; they are using Facebook or playing a game during most of those times.

Next up, let’s look at personalization. In other words, does it make a difference if I add a personalization to the Tweet, or is the page Title and bit.ly link (which is what buffer tweets when you use the Chrome extension if you don’t modify it) enough? Basically, are my followers following my content or my comments? Let’s take a look at the most clicked tweets again:

Personalized?

# of clicks?

NO

92

NO

91

NO

90

NO

89

NO

58

NO

49

NO

44

NO

33

NO

33

NO

16

At least in my stream, it doesn’t matter if I add a witty bon mot or comment, or answer the question in the headline- my Followers want to read the content, not my take on it. In fact, here are the numbers for all of my personalized Tweets:

Personalized?

# of clicks?

YES

5

YES

5

YES

3

YES

2

YES

2

YES

1

Yes

1

YES

1

YES

Yes

Yes

Yes

YES

Yes

YES

Yes

YES

Yes

That’s…discouraging. My followers seem to prefer robotic re-tweets of the content that interests me. I don’t blame them, I get up from my desk for 5 minutes to get coffee and when I get back there are sometimes as many as 30 new Tweets in my stream. I’m not going to read all of those and it seems unreasonable to expect my followers to.

Now, let’s move on to days of the week. Here’s the breakdown by Day of the week, number of Tweets, and number of Clicks:

Tweets

Clicks

Ratio

Saturday

2

2

1.00

Friday

57

90

0.63

Sunday

2

5

0.40

Thursday

54

215

0.25

Tuesday

41

177

0.23

Monday

37

186

0.20

Wednesday

44

232

0.19

We have to take this data with a grain of salt as I really don’t tweet on weekends, but if we toss Saturday and Sunday out as having too small a sample size, it seems that Friday is the best time for me to Tweet something and get it clicked- this probably correlates with most offices’ relaxed Friday atmosphere, and the fact that people usually have more time to devote to secondary activities like social media as the week winds down. I also seem to Tweet more on Fridays, probably for the same reasons. Wednesday looks to be the worst day for me to Tweet if I want my message to resonate- maybe due to stress from the mid-week crush people just don’t have time for Twitter?

Next, I wanted to see if my followers are picking up the messaging I am putting out there. To check that, I will compare the number of clicks with the number of “on-topic” Tweets- that is, Tweets where I speak specifically about my professional area of expertise and where a stranger might be compelled to follow me, were they to see those Tweets.

Posts Clicks Ratio
Industry News

27

70

0.39

SEO

33

86

0.38

Facebook

21

86

0.24

Google

44

200

0.22

Total

125

442

0.31

 

So for Tweets that I am actively trying to get clicks for, namely, the self-declared topic of my Twitter, I have about a 30% click rate which I feel is very good. However, this is right on pace with the overall click rates on my general pool of Tweets, so I can’t (as much as I’d like to) attribute this to being an outgrowth of my fabulous Twitter content curation skills.

Next up, hash tags. I want to ask two questions here about hash tags: does using a hash tag make a difference in your click through rate, and which hash tags get the best click through rates? (I have an inkling on the answer to the second question but we’ll let the number play out and then I’ll make an observation later. Of the 241 Tweets I made in September, 176 of them or 73% contained at least one hash tag. Those tweets resulted in 556 clicks, whereas the posts without hash tags resulted in 351 clicks. That’s a rate of 2.3 clicks per Tweet, compared to 5.4 clicks without a hash tag. It seems that not using a hash tag will actually result in a higher click through rate!

Facebook- 20 Tweets resulting in 86 clicks

Google- 42 Tweets resulting in 94 clicks

SEO- 31 Tweets resulting in 73 clicks

Twitter- 8 Tweets resulting in 183 clicks

The takeaway here is that if you’re posting information about Twitter, on Twitter- you’ve got a highly engaged audience! Social Media topics do well on Twitter because of the service’s high rates of adoption among those that live and breathe the topic. Speaking of Social Media topics on Twitter…


Hey, that’s not cool! At least 30 of my followers liked my Tweet well enough to click on it twice, but not one single re-tweet?!?! I’d be willing to bet that at least a few of them simply copy/pasted into their own Tweet stream- but savvy social media Users, like the ones that would click on a link about Klout, appear to be pretty stingy with the re-tweet love! This spotlights one of the problems with Twitter- people aren’t as altruistic as they tell you you should be. I will re-tweet things that I think are particularly funny or amusing, or that I haven’t already seen before- but I’m generally not LOOKING for things to re-tweet, either. It’s sort of like no-following a link on a page- sure, the link out doesn’t hurt anyone, but why give away my expertise/time/attention currency for free? Next time I get a similar topic, I’ll try an experiment- I’ll put the same tweet out twice, once with “PLEASE RT” added to see if that makes a difference in people’s generosity.

According to my (admittedly limited in scope) research, my best chance for getting a Tweet clicked is to post non-personalized Tweets that are specifically about Internet Marketing industry news, with no hash tags after lunch on Fridays.  What do your Twitter statistics look like, and what do they tell you?

No comments

Leave a Reply

Your email address will not be published.

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=""> <s> <strike> <strong>