The Fallacy of Timing Blog Posts & Social Media Updates
Studies that claim to tell you when the best time to publish a blog post or share content socially are, in my opinion, mostly a load of crap.
That’s my opinion, and I’m sticking with it.
You’ve probably seen these studies and statistics. The folks at Bit.ly (a service I like and use) just shared some of their data and we wrote about it on Marketing Land. In a nutshell, Bit.ly is saying that the best time to post stuff on Twitter and Facebook is early afternoon, U.S. Eastern time.
I don’t mean to pick on Bit.ly because, as I said, that’s just the latest in a pretty long line of articles and reports that try to suggest that there’s one magical time that you should be publishing and sharing content.
There’s not.
Why There’s No Magical Time To Publish
Here are a few problems with those kinds of articles/reports:
1.) When is YOUR audience online?
Those studies try to look at Internet activity as a whole, and it may be that retweets or likes happen more often on certain days and at certain times. But chances are good that, if you’re reading this, the entire Internet is not your target audience.
Those studies don’t take into account when your audience is online, when they’re active and when they’re most likely to be receptive to your content. Danny Sullivan said it well in that Marketing Land article I mentioned earlier:
“…ultimately it will be up to your own individual findings to know the optimal time to post content.”
2.) If everybody starts posting at the optimal time, no one will see anything.
Social media is noisy enough, right? If you’re trying to get your content seen by potential prospects and customers, talking at the same time that everyone else is talking probably won’t work.
3.) These studies don’t account for quality of the content.
Things are read and shared because they’re good, not because they were timed correctly. I’ve yet to see a study on timing blog posts or social media updates that takes quality into account. (Did I miss it? Leave me a comment, please.) It would be tough to account for quality in any of these studies because quality is subjective — it depends on the audience (see No. 1 above).
I like what Jason Falls said earlier this year about how quality relates to timing:
“…does day of the week or time of day even matter? We don’t know because we didn’t first see if the critical success factor was something different … like quality of the content.”
4.) You can’t predict how anyone will respond to your content no matter when it’s published.
Humans are often unpredictable. You don’t know who’s online at any given moment. Last August, I accidentally published an article here on Small Business Search Marketing at about 11:00 pm on a Monday night. I meant to schedule it for Tuesday morning when more people would be online, but I hit “publish” by mistake.
Look what happened:
That huge spike near the middle there? Yeah. That was my “mistake.” Published at 11 pm (Pacific time, even!) on a Monday night. I wrote about this on my personal blog and called it the greatest blogging mistake I ever made.
Most of my audience wasn’t online at that late hour, but one person with an enormous following was and when he retweeted me, all hell broke loose. I could never have predicted that.
Final Thoughts
I’m not trying to throw these studies (not the people/companies that publish them) under the bus. The data is interesting and informative, and has value at a general level.
But they’re too general. They don’t take your audience into account, and they don’t take the quality factor into account.
Test what works for you and your audience. Publish blog articles at different times of day, and on different days. Do the same when sharing content on Facebook, Twitter, or wherever you’ve found your audience online.
And, even after you run your own experiments, don’t be surprised if that perfect/magical time isn’t always perfect and magical. Because it’s really hard to predict human behavior.
Great perspective and points, Matt. The study data is interesting to read and internalize; but, most importantly,businesses need to superimpose the info with’their specific situation,’ to truly benefit respectively.
Also, show the marketing world a rule and we’ll show you an exception; as you state, “humans are often unpredictable.” As a corollary, that makes marketing norms unpredictable too.
Good stuff here Matt, and I totally agree. I published my Author Search post on a Saturday night (what is often called a “bad time” to publish), and it went hot on HackerNews and in the Twittersphere because there was low competition (and well, it’s a good post).
Thank you for this dose of sense.
Atleast in partial answer to point 1, there are tools like http://www.socialbro.com which analyse your twitter followers to determine the most active times they are online. They suggest you reanalyze your audience at least once a week.
Matt – what time did you publish this?
(Because that’s when I am going to publish everything from now on 🙂
Ha! Chris – too funny. Time stamp in admin says it was 12:16 pm, so there you go. Ironically, I was hoping to publish this earlier in the week because … well … more people are online. But I obviously got sidetracked into a Friday afternoon post.
Daniel – thx, and I think there are a couple other tools/services that can analyze Twitter followings like that, too. But I still would argue that there’s no formula to be had purely in terms of timing. There are too many other factors at play.
Anthony, John – good stuff, thanks for sharing.
As you point out, one problem with these studies is they assume your target audience is “the internet”. The problem with the SocialBro, Crowdbooster, Timely, etc approaches is that they are focused on your audience and your results.
Your 11:00 PM example is great, and highlights the reality for most people: it isn’t your followers that matter, they are a drop in the bucket. There are a small handful of individuals that can really amplify a post. Anything that works on averages, even averages for your audience, misses that fact that the average person picking up your post isn’t what matters.
Great reminder, thanks for sharing!
Matt and Eric, would you suggest this is an issue of insufficient analytical methodologies or that the social aspect of social media implies an element of chaos which is impossible to predict?
Daniel, if I could take a stab at that question…
If you imagine your own sharing habits, I’d be willing to wager that you would not be able to isolate one hour out of every day when you’re active. Speaking from my own personal habits, I’ll Buffer items while I’m making coffee in the morning, waiting for the check after lunch and maybe during a commercial break in the evening.
The key is to engage the right people, while they’re in the right mood and that have enough amplification to make a difference. I’m not aware of an app for that yet… (Bottlenose is trying,) but I’ll sign up as soon as it’s available.
So my vote would be for the “element of chaos” you so accurately described.
I actually pay very little attention to “optimal” posting times, but in theory I don’t see what is wrong with using this data as part of your considerations. All things being equal, it probably is best to post things to social networks at the times when it is more likely to get more engagement. As you mention, this needs to be specific to your audience, which is why generic statistics may not tell you the whole story. They are just a starting point that can be used in your planning and testing.
However, just because you bias posting at times when statistically you are most likely to get high engagement, it doesn’t mean you should become an automaton and never vary things. Why not generally try to post at optimal times but have a little variance/flexibility to take advantage of chaos events?
Not only tweet and blog times, but “best day and time to email” aren’t always the best for your situation.
We had a large commercial bank for which we were doing a series of email messages, most directed to small business CEOs. We had one scheduled for the week before Thanksgiving week — but it got bogged down in approvals, and on the day before Thanksgiving, the marketing director finally gave us the go-ahead.
We talked about waiting until after Thankgiving, but the marketing director wanted the email sent NOW. The email was sent on Wednesday afternoon, the day before Thanksgiving. The bank got better response on that email than on any other previous email in the series.
Our new motto: “when in doubt, send it out!”
It’s all about your audience and it’s specific to every business. It is hard to generalize however, if you have a large customer base people do start becoming predictable and timing can be more accurate. The smaller the crowd not so much.
It probably falls into the category of knowing your target market, but I never see any info on the differences in type of media being posted. Whether its video, images, or articles.
Another thing to consider, particularly for social media, is that your most valuable readers/clients/customers might be paying attention at times that the majority are not.
Both of those are good points, Josh – thanks. I’ve also not seen any deeper looks at how different content types perform at different times of day. Anecdotally, and speaking only for myself, I don’t have time during the week to sit and watch videos that are shared socially. Maybe on weekends. Maybe.