How Many Tweets is Too Many Tweets?

As Community Managers, we’re constantly trying new things. Whether it’s a new network or a new CTA, experimentation is in the job description. Unfortunately, we, as Community Managers, are also really busy. Trying new things can be daunting, especially when you have a strict plan to hit your numbers. But, that shouldn’t deter you; The risk is often worth the reward. How do you know what works, what doesn’t, or what could be better? Test.

The Question: How Many Tweets Is Too Many?

downloadI’m not sure where I heard it, maybe it was a conference or a blog post, but at some point it was put in my head that Tweeting a lot can really peeve your community. Now, I wasn’t sure what “a lot” really was. Does that mean I should only tweet once a day? Or once an hour? I kept Tweeting and my numbers continued to improve. The community didn’t seem to mind, there was no negative feedback or slump in engagement. In fact, my key metrics got better. Much better. So, I decided to formally test if my frequency of tweeting affected my engagement, engagement per tweet, and site traffic.

The Experiment: Testing Tweet Frequency

To test the effects of Tweet frequency, I Tweeted for one week in 15-minute increments, and another week in 30-minute increments. I ran this test in December, and again in January. I pulled all the data from each week using the Simply Measured Account Report and Google Analytics, put it all into Excel, and compared the weeks side by side. Sounds good, right?

The Experiment… Again: Back to the Drawing Board

I presented my findings to my team, confident about the results. But our Conversion Manager, former analyst, Nate, poked some serious holes in my logic. See, after 5pm on a business day, my Tweet frequency would change. I only Tweeted in 15 or 30 minute intervals between 6am and 5pm, but my reports pulled for the full 24 hours. My sample set wasn’t specific enough. So, I went back to the drawing board. I adjusted my samples, pulling data only from within business hours, and what I found was really telling.

The Findings: Tweet Your Heart Out!

Website Visits
Screen Shot 2014-02-11 at 2.44.18 PM
After compiling the data from both weeks, I found that Tweeting in 15-minute increments increased traffic by 31% and engagement by 89%.
Engagement Per Post
Screen Shot 2014-02-11 at 4.05.50 PM
The most interesting finding was the changes, or lack there of, in engagement per post. This was shocking because the natural inclination would be to assume that the more you post, the more spread out engagement will be and, thus, the lower your engagement per post will be. But, engagement per post barely changed between the two. In fact, 15-minute increments even averaged a tiny bit higher (1.4%).
Total Engagement
Screen Shot 2014-02-11 at 2.41.04 PM
Total engagement was the least surprising disparity to me. It stands to reason that when I Tweeted twice as much, total engagement followed suit. My one fear was that there’d be a diminishing return as followers got sick of constant updates, but that didn’t prove to be the case.

What We Learned
Aside from the obvious findings about Tweet frequency, this test led me to tons of other findings. For example, when I segmented our engagement and traffic for exclusively between 6am – 5pm, I saw how much was occurring after business hours. Stacking more tweets, more powerful CTAs and richer content in the evenings could potentially drive our numbers up even further. Our audience is not only receptive during the day, but they’re actually tuned in all the time. Perhaps it’s because of the space we work in; social media geeks seldom turn it off. In the future, this is something I will definitely be able to capitalize on.
Conclusion
One thing that you probably shouldn’t do after reading this blog post is automatically start tweeting more. Just because it worked for me, doesn’t necessarily mean it will work for you. Every audience is different; some will respond better to seeing your handle pop up in their feed multiple times a day, while others won’t. But what I strongly recommend you absolutely do after reading this post is run this test for yourself! The best way to find your sweet spot, is to test, test, test!

Want to test this out for yourself but have questions? Leave them in the comments or Tweet me!

Jade Furubayashi

Jade Furubayashi

My name's Jade and I'm a social media addict from Seattle, WA. I'm the Community Manager at Simply Measured, so my job is basically to talk to you fine people everyday- which is awesome. Aside from that, my favorite things are dogs, tacos and Beyoncé.

  • Jen

    Curious to know if you received more unfollows or follows in the time period where you were tweeting more frequently.

    • Kate Hutchinson

      An excellent question.

      • Jade

        Hey Jen & Kate,

        This is definitely something I would want to put in my next phase of analysis. I’ve taken a quick look at the data and couldn’t identify any major peaks in follows or unfollows, but it is definitely something that requires a deeper dive. Stay tuned!

  • http://www.rishona.net/blog/ Rishona Campbell

    Great write up Jade. I like how your emphasis is to encourage others to do the research for themselves and take the approach that works for them. Each audience is different….the best person to gauge what content they appreciate and what they respond to is you :-)

    • Jade

      Thanks Rishona! You’re exactly right. Since everyone’s audience is different, there’s no one size fits all strategy.

  • LeBlancly

    This is interesting, but I can’t help but feel that some element of it is going to go back to what the content was. Did you have any sort of plan to post similar types of content for your experiment to standardize the results in some way? For example, posts with links vs not, posts that ask a question vs not, etc.
    Either way, completely agree that the only way to know is to test, test, test!

    • Jade

      Hey LeBlancly,

      I definitely hear that. Variations of content certainly have an impact and that would be a great thing to look at in my follow-up analysis!

  • D. Kruegel

    A 15 Minute interval is probably still long enough for a tweet falling down the stream unnoticed for many followers until the next tweet comes in. I have read somewhere else that a 20-minute interval is ideal, as the longevity of a tweet is barely 20 minutes in average. But that surely depends on the average amount of accounts a follower is following and the size and resolution on the screen (in other words: how many tweets fit into the screen of the follower’s device). The lower the amount of followed accounts and the larger the screen, the longer the life of a tweet on their screen.

    • gmart

      It just amazes me that 1) people have that much to tweet; and 2) people tweeting that much can get any other work done. More power to ya!

      • Alan Martin

        That’s where BufferApp comes in;-)

  • heyitskarri

    Hi Jade,

    Great experiment! We found similar results when we dove into how B2B marketers engage on Twitter a couple months ago… the people who get mentioned and retweeted most generally tweet a heck of a lot more than the average person! We also found some other “best practices” that might make interesting experiments in future.

    Slides from that webinar are available here: http://www.slideshare.net/dnnsoftware/how-to-engage-b2b-marketers-using-social-media

    Looking forward to your next post!

  • Juan

    However, if you’re ever at a Twitter office, their slide deck will say the ‘Tweetspot’ is tweeting 4 (yes, four) times a DAY. The reality is that it always depends on your audience and your niche. Measure and figure out what’s best for your audience. It’s good to see a community manager that can create bar graphs though :)

  • sabbadoo32

    I believe the people who tell you “don’t tweet too much” should instead say “don’t tweet too much of the stuff that we write.” Generally we tweet a lot, as there are a lot of unique things going on. If it’s well-written and interesting, are one thing, putting it out so people can see it is another. Timing-wise we space them around 20 minutes apart, not at set intervals (sometimes 18 minutes, sometimes 21 minutes, etc). We look at our day in terms of morning (10 am to 1 pm), afternoon (1 pm to 4 pm), and evening (8 pm to 10 pm). Weekends 8 am noon, and evenings 9 pm to 11 pm. We’ve grown from 2100 to close to 7000 followers in a little over 2 years.

    • gmart

      I am always curious about people specifying time periods. Are the people you tweet to only in your time zone?

      • sabbadoo32

        Yes, we are the third largest county government in the US.

  • http://inspiretothrive.com/ Lisa Buben

    Great stats, I believe too it matters how many followers you have – I find when I get new followers if they only are following a few hundred of less they may stop following me because of how often I tweet. I also agree with others that the niche and type of tweets make a difference too :) Love to see your follow up on that as well Jade.

  • http://www.ThoughtfulRealEstate.com sunshineseattle

    You at SimplyMeasured write some good articles – I could cope with one every 15 minutes. But when people are telling me that they need coffee or they had great sex last night – those are the ones that go to my “loudmouth” (private) list. I need the loudmouths but I don’t need to hear them. I no longer read my tweets directly I segment them into lists and read the list (or not).

  • Robert H. Woodman

    Pardon me, please, if this is a dumb question, but when you say “Tweeted” in 15/30 minute intervals, I take that to mean that you originated 1 tweet every 15 minutes or 1 tweet every 30 minutes. My question is, did all of these tweets contain novel content, or were some of the tweets reworded tweets of things already tweeted earlier? Followup question: if you tweeted both novel content and reworded material, how did engagement vary? How does time spacing of reworded tweets affect engagement?

    I also would be interested in your follow-up analysis relative to LeBlancly’s question about content.

    • Robert H. Woodman

      Thanks, Jade. Very interesting information.

  • http://blog.socialmediahq.com/ Nick Robinson

    This is a really interesting experiment, and I’ve seen similar results on some tests. What type of business recommendation would you make based upon these results? Staff up on your social creation team (which requires budget)? Staff up on your tweet scheduling team as well (also requires budget)?

    Basically, my point is, all of this activity isn’t free and would require someone to schedule tweets all day long without doing anything else. Are you assuming that you will be doing all of this activity into the future, or will you require help from another community manager? How much money will this cost? How will this affect your cost/engagement?

    • Jade

      Hey Nick,

      We actually wrote a post addressing this problem. The concept is getting a greater ROI on your Tweets and time by finding the most effective times to post: http://simplymeasured.com/blog/2014/02/27/are-you-tweeting-at-the-wrong-time-for-site-traffic/

      I would recommend running both of these tests on your own brand, find out what works best for you, and budget within your constraints!

      • http://blog.socialmediahq.com/ Nick Robinson

        Thanks Jade! Answered my question :)

  • http://2013fashionnnews.blogspot.com/ Fashiontechno Top
  • http://www.twitter.com/netsjetsfan njfan

    I would like to retweet this article but if I did would that mean I risk tweeting too much?

  • http://jeffbullas.com Jeff Bullas

    That is some insightful data Jade.
    I have been tweeting every 15 minutes for the last 3-4 years (I have a tool that does it for me) and had been accused of tweeting too much (especially when I started)
    But I have always treated Titter as a stream and not an inbox and if you don’t keep tweeting you are forgotten.
    On the type of content and the the quantity challenges.
    You need to create evergreen content so you can keep tweeting and you do need to have a lot of good content.

    • Jade

      Hey Jeff, thanks for taking the time to read this post! I definitely agree with you about Twitter being a stream, rather than an inbox. One thing I pointed out in the post that I think is also particularly relevant is how your audience defines “too much Tweeting.” For example, an audience of a B2B social media company or a social media influencer might be less irked by frequent Tweets (if they’re relevant) since they’re interested and invested in social media marketing.