Twitter   The value of crowdsourcing can be more than mining social media for sources. nie.mn/JLgvek
SHARE
Fuego
Nieman Journalism Lab
Nieman Journalism Lab
Pushing to the future of journalism — A project of the Nieman Foundation at Harvard

Tweet late, email early, and don’t forget about Saturday: Using data to develop a social media strategy

Tweet more, and embrace the weekends.

That’s according to Dan Zarrella, a social media researcher (with 33,000 followers himself). Zarrella works for HubSpot, mining data on hundreds of millions of tweets, blog posts, and email newsletters to help marketers find trends. News organizations should pay attention, too.

Zarrella says the right Twitter strategy depends in part on what your goals are. Want to accumulate as many followers as possible? Then tweet a lot: Twitter’s A-listers — those with the most followers — tweet an average of 22 times a day, and more tweets generally lead to more followers. But if your goal is to drive more traffic to your site, you should show a little more restraint; accounts that share two or more links an hour show a dramatically lower clickthrough rate than those who share no more than one.

It’s an inexact science, but at least it’s an attempt at science where so much social media strategy is driven by intuition. (Zarrella complains about the the “unicorns and rainbows” strategy: “Love your customers, hug your followers, engage in the conversations. It sounds like good advice, and it’s hard to disagree with,” he says. “But generally, it’s not based in anything substantial.”)

After collecting more than two years of data, Zarrella shared his findings Tuesday in a webinar called “The Science of Timing.” That science is less about when and more about when not — what he calls “contra-competitive timing.” The trick is to reach people when the noise of the crowd has died down.

It turns out that time is often the afternoons, when blogs and news sites are slower, and the weekend, when they’re all but asleep.

Retweet activity is highest late in the work day, between 2 and 5 p.m., and the sweet spot (tweet spot?) is 4 p.m., Zarrella’s analysis found. Late in the week is most retweetable, too. Zarrella created TweetWhen to tell Twitter users what time days and times yield the most retweets. (Our @niemanlab tweets get the most retweets around 9 p.m. and on Saturdays. Go figure. That’s our hour-by-hour chart up top.)

On weekend mornings, when most news sites see substantial drops in pageviews, Twitter clickthroughs spike, he says. Comment activity also jumps dramatically: Users have more time and attention to devote to content on the weekend, even if the content isn’t fresh, and fewer distractions compete for attention. On Facebook, Zarrella says, the effect is even more pronounced: Facebook participation on weekdays is infinitesimal in comparison. (He thinks it might be because so many companies block Facebook at the workplace.) Facebook does not reward frequent posting in the same way Twitter does, however, and it’s much easier to flood (and annoy) Facebook fans than Twitter followers. Postings on Facebook also tend to “stick around” longer, re-emerging when people post a comment or like.

Not only does Zarrella recommend tweeting more — he recommends tweeting the same links two or three times a day. Don’t bother calling it a “rerun” or apologizing to people who might have seen it before. Simply wait a few hours and change up the language. Only a fraction of your followers will see it the first time. Even an organization with thousands of followers won’t reach most of its audience most of the time. (As an experiment, social media star Guy Kawasaki once repeated the same tweet every day for nine days, and found the clickthrough rate remained high each time.)

Zarrella recently performed a similar deep dive into data for email newsletters, working with MailChimp to analyze 9.5 billion messages. A lot of the same lessons apply, he says: Email more, and embrace the weekends.

Most people who unsubscribe do so after receiving their first email. Send 30 emails a month or send five — it makes little difference, he found. (“Unsubscribers are doing you a favor,” he says — they don’t want to hear from you anyway.) The most important time to reach subscribers is right away, especially in the first couple of days after signup. The average click rate for the average user drops to almost nothing after four months. Like with blogs and social media, readers are more likely to open an email newsletter and click links on weekends. For all days of the week, early-morning hours (between 4 and 7 a.m.) are the best times to reach readers — before they get caught up in their to-do list for the day.

Here are the slides from Zarrella’s webinar:

The Science of Timing
View more presentations from HubSpot Internet Marketing

                                   
What to read next
Josh Marshall
Adrienne LaFrance    May 17, 2012
TPM’s Josh Marshall says he now views what was launched as a website as a bundle of knowledge and expertise that “exists inherently on no particular platform” — which is why he’s moving big into mobile and video.
  • http://www.company2keep.com Cathie Guthrie

    Brilliant summary of today’s webinar! Kudos Mr. Zarrella and Mr. Phelps.

  • http://www.company2keep.com Cathie Guthrie

    Brilliant summary of today’s webinar! Kudos Mr. Zarrella and Mr. Phelps.

  • http://www.company2keep.com Cathie Guthrie

    Brilliant summary of today’s webinar! Kudos Mr. Zarrella and Mr. Phelps.

  • http://twitter.com/UniqueVisitor Jeff Pester

    There are few things more irritating than people/accounts that are tweeting the same links two or three times every day. And changing the headline makes it even worse. In that case the reader is expecting new content, only to find out they’ve been duped in to clicking through to content that isn’t new at all.

    We’ve got to get away from these incessant “blasting” techniques that are designed solely to amass followers and/or get noticed – it’s maddening and it only adds to the increasing avalanche of Social Media static. In fact it’s worse than mere static – it’s digital litter. And we’re getting tired of having to wade through all this trash.

  • Pingback: Is @acarvin the First Twitter Anchor? Is NYT’s Paywall Like NPR? When is the Best Time to Tweet? | Ebyline Blog

  • http://twitter.com/cdevers Chris Devers

    Hear hear.

    My personal favorite example of this is @ebertchicago. I’ve been a fan of his film criticism for years, and I happen to generally agree with his politics, too. (You may or may not; that isn’t the point.) At first I thought his Twitter feed was great, until I started realizing that he tends to post the same handful of things over, and over, and over. On any given day, he appears to have a shortlist of topics he wants to flog, and he’ll circle back to them again, and again, and again, sometimes for hours, sometimes for days or longer.

    Ebert is obviously doing something right, as just under 400k people are subscribed to his account. But I’ve come to think that either these people aren’t really reading everything, but just skim (which this constant re-posting would work well for, of course). That or, more cynically, these people have a much larger appetite for repetition than I do.

    This way of using Twitter has a role, I suppose. But personally, I have a hard time picturing being a long-term reader of anyone that uses the service this way. I know my unfollow-button trigger finger starts getting real itchy, real fast when I get the feeling that someone I’d otherwise be interested in reading is going to be doing this.

  • http://twitter.com/cdevers Chris Devers

    Hear hear.

    My personal favorite example of this is @ebertchicago. I’ve been a fan of his film criticism for years, and I happen to generally agree with his politics, too. (You may or may not; that isn’t the point.) At first I thought his Twitter feed was great, until I started realizing that he tends to post the same handful of things over, and over, and over. On any given day, he appears to have a shortlist of topics he wants to flog, and he’ll circle back to them again, and again, and again, sometimes for hours, sometimes for days or longer.

    Ebert is obviously doing something right, as just under 400k people are subscribed to his account. But I’ve come to think that either these people aren’t really reading everything, but just skim (which this constant re-posting would work well for, of course). That or, more cynically, these people have a much larger appetite for repetition than I do.

    This way of using Twitter has a role, I suppose. But personally, I have a hard time picturing being a long-term reader of anyone that uses the service this way. I know my unfollow-button trigger finger starts getting real itchy, real fast when I get the feeling that someone I’d otherwise be interested in reading is going to be doing this.

  • Pingback: When is the best time to tweet? Late - Lost Remote

  • Pingback: Op zoek naar de juiste Twitter-strategie voor je redactie | Nieuwslab

  • Anonymous

    How many people do you guys follow that you even notice multiple tweeting?

  • http://twitter.com/damienadupont damienadupont

    I listened to the webinar yesterday and thought there was a lot of good in it. But I couldn’t help thinking – I wonder how much his data is affected by people (I think he called them ‘blogerati’) who *schedule* tweets and blog updates to occur later than they are written? For example, he thinks 11 pm might seem a wise time to reach out to bloggers because they post in the early AM. And that its best to blog early in the morning to get bloggers to link to you. But do we really have any idea why they appear to be posting/linking in the early AM? Couldn’t this be skewed by scheduled posts?

  • http://twitter.com/UniqueVisitor Jeff Pester

    Hi Dan,

    I don’t follow that many people in my personal account but out of that limited set I’d half a dozen or so (5%) of them have engaged in some form of multiple tweeting of the same content within the same day. I’ve dropped a couple of them, and more importantly, my professional opinion of the others has been greatly diminished because of this tactic of over-saturating a channel (Twitter) with the same content in an obvious attempt to get more followers and/or more attention.

    It’s a bit like littering. One person throws out one candy wrapper and you think “it’s only one little wrapper, I bet no one even notices”. But then more and more people do it because they see someone they respect do it and all of a sudden it becomes a real problem. That’s what is possible here.

    While “same channel/same content saturation” might be a good tactic to expose yourself to new potential followers, encouraging more people to do the same will only exacerbate the static problem.

    You have to respect the entire stream, not just the part that runs by your property.

  • http://twitter.com/saptarshi_ Prathamesh Murkute

    tweet reasonably!

  • Pingback: Nieman | Tweet late, email early, and don’t forget about Saturday: Using data to develop a social media strategy « Newsplexer

  • Pingback: links for 2011-03-31 « Sarah Hartley

  • Anonymous

    It is exactly a tragedy of the commons issue. Via @TweetSmarter I repeat a very small number of tweets one time only, but I mark each one. http://blog.tweetsmarter.com/twitter-tips/should-you-ever-repeat-the-same-tweets/

  • http://twitter.com/UniqueVisitor Jeff Pester

    In what way is this a tragedy of the commons issue? Tragedy of the commons involves the effects of multiple people acting in their own self-interest resulting in damage to the community. If you’re referring to my littering analogy, I don’t think you can equate someone littering to acting in their self-interest. In this case we’re talking about selfish individual behavior unrelated to anyone else’s participation.

    Repeat Tweeting is intended primarily to drive greater exposure and increase the base of followers. That’s it. It’s a customer acquisition tactic. Period. Arguing that it’s main purpose is anything other than that is just intellectually dishonest.

    As you say yourself in your blog post: “Repeating Tweets is unpopular with readers”.

  • http://www.dolcepress.com Dolce Letterpress & Design

    Like your analysis! The comments on weekends is definitely true.

  • Pingback: This Week in Review: Navigating the Times’ pay-plan loopholes, +1 for social search, and innovation ideas » Nieman Journalism Lab » Pushing to the Future of Journalism

  • http://themeasurementstandard.com Bill Paarlberg

    I wonder how much variability is being lost by analyzing all this data in large groups? I suspect there are many different audiences with various goals, each with their own characteristics of when and how they like to get information. For instance, it is clear from the comments here that there is a subset of Twitter users which feel strongly negative toward the practice of tweeting the same links two or three times a day. This subset could be quite large and influential, but that effect is swamped in the aggregate analysis.

  • Álvaro Ordóñez Gómez

    Brilliant!

  • http://www.qaqn.com Daniel M. Clark

    There are two factors at play that have lead to your irritation at what many call a fair use of Twitter. First, you follow few people, as you said in another comment. When you’re only following a hundred people, and only a fraction of them are very active (statistically, that’s probably a safe assumption) then you’re simply not reading very many tweets at all and repeats stand out. The second thing is that you may (or may not, but it’s possible) be watching Twitter too closely. If you’re noticing that someone tweeted the same link twice in a day, that might be a sign you’re a little too active.

    By way of comparison – I follow more than 500 people, most of them from the affiliate marketing industry, and they are very, very talkative. It’s rare I see repeats, but occasionally when someone is launching a new product or trying to drum up interest in a cause, I’ll see a little. I also only check Tweetdeck about three times a day, sometimes less. The odds are very much against seeing repeats with the number of people I follow and the amount of time I spend doing it.

    Lastly – Twitter is little more than an attention-getting tool. To call it anything else is disingenuous. If someone wants to have private conversations with people who never, ever, try to get more followers or try to get you to go to their sites – even passively – they can jump on IM or find a chat room somewhere. Twitter is, by its very design, intended to connect as many people with as many other people as possible. You (the general “you”) don’t put a tweet out there hoping that nobody will follow you as a result. You might not care if you don’t get new followers by the hundreds, but if your follower count was zero, you wouldn’t bother using Twitter.

    Just a thought. I could be wrong.

  • http://www.qaqn.com Daniel M. Clark

    Ebert’s a great example, though – here’s a person whose career lives and dies by the number of people who are paying attention to him. He’s a public figure. Twitter is just another channel for him to get his content to people, and it’s a way to build his fanbase. I get why you don’t dig the method, but like you said, 400k people… of course they’re not all hanging on every word, but many are – and that makes his methods successful, I think.

  • http://twitter.com/UniqueVisitor Jeff Pester

    Hey Daniel, appreciate your perspective. I’m pretty sure I didn’t suggest that Twitter is an appropriate place to have private conversations – it’s not. I use it primarily as an RSS replacement and to follow those whose updates I’m sincerely interested in keeping up with. Maybe that makes me an outlier.

    BTW, if anyone is following 500 people who are very, very talkative then although they may technically be following them, they’re not paying much attention to them. And that’s ok. The beauty of Twitter is that it lets each user define what value they want to extract from it.

    And while I agree that “Twitter is, by its very design, intended to connect as many people with as many other people as possible”, posting of identical content multiple times per day is obviously a calculated attempt to attract new followers without regard to how it affects existing followers.

    That’s my point. As long as there is currency/influence associated with # of followers we’re going to continue to see a lot of users do whatever they think is necessary to get more followers, no matter the tactics.

  • Pingback: The Best Time To Tweet (And More Secrets To Social Media Timing)

  • Pingback: The science of timing: When to post on Twitter and Facebook | Web Up the Newsroom

  • http://www.linkedin.com/in/cshunt312 Courtney Hunt

    With all due respect, I think people should ingest this pseudo-science with a massive grain of salt. I’ve listened to one of Dan Zarella’s webcasts and was surprised by how many people accepted his data – and interpretations – without question.

    Bill brings up an excellent point re: aggregation effects. Also, one question I’ve never seen adequately addressed is “which time zone?” The community with which I engage is global, and I’m in the central time zone in the US. So, 9 am for me could be 3 pm in the UK. I also think it’s critical to remember that the past is no predictor of the future, especially when it comes to Twitter. Data from even 6 months ago may not be relevant today, due to changes in users and their activity. Finally, if we all want people to take social media seriously, then we should promote using it responsibly rather than encouraging people to engage in gaming techniques.

    I agree it’s important to measure metrics and use that feedback to refine your own engagement, but the sample used for benchmarking should be your own activity as well as the activity of relevant others. This is definitely a situation in which more is less.

    BTW, I recently wrote a post about Twitter “worst practices” that has been widely read (by my humble standards) and well received. Folks who are interested can find it at http://tiny.cc/SMinOrgsTwitterPost.

    Courtney Hunt
    Founder, Social Media in Organizations (SMinOrgs) Community

  • Pingback: Fenton Must-Reads: « Fenton | Progress Accelerated

  • Pingback: Questione di… timing « Totoromano's Blog

  • Pingback: Twitter for a Cause – Part 2 – Building a Buzz | Rush Essay Discount Code

  • Anonymous

    Useful stuff indeed!

  • Pingback: The early bird may get the worm, but the late tweeter gets the followers. | SocialMediaFluent

  • Pingback: How Often Should You Update Evergreen Content - Graywolf's SEO Blog

  • Pingback: How Often Should You Update Evergreen Content | WebProNews

  • Pingback: Online Marketing, SEO, PPC Agency » How Often Should You Update Evergreen Content

  • Pingback: TJantunen.com » How Often Should You Update Evergreen Content

  • Pingback: How Often Should You Update Evergreen Content? | Popular Events

  • Pingback: How Often Should You Update Evergreen Content? | SEO News Greece

  • Pingback: Texten fürs Web » Blog Archive » Der TfW-Praxistipp: Die optimalen Zeitfenster fürs Twittern

  • Pingback: Se devi twittare, fallo di notte. È il momento migliore « GOOTENBERG

  • Pingback: Timing Tweets « EJO – European Journalism Observatory

  • http://davidglarson.com Dave Larson

    He points out obviously non-causative correlations (analyzing single words out of common phrases, for example) in order to present bite-sized factoids of little or no relevance…but which often create dramatic, easily tweetable points.

    In other words, he will use any one of the following three to get more traffic: ▬ Lies; ▬ Damn Lies; ▬ Statistics.

  • Pingback: Dobry czas na ćwierkanie « EJO – European Journalism Observatory

  • http://www.siliconvalleywatcher.com/ Tom Foremski

    The problem becomes that if you don’t repeat Tweet (I use a DejaView label) everyone else is doing it and to punch through the noise you have to do it too to a certain degree.

  • http://twitter.com/UniqueVisitor Jeff Pester

    Hi Tom; So in order to cut through the noise, we need to create more noise, to cut through the increased level of noise :)

    I think more than anything this desire to rebroadcast content repeatedly in the hopes of catching people whenever they happen to be on the channel highlights one of the shortcomings of the realtime channel. It’s a little like having a TV that only gets Headline News. 

  • Pingback: Bullet Points: The Downside of Low Turnover; Making Your Company Resilient; the Point of Social Media Experts | The Hired Guns Blog

  • Pingback: e.politics: online advocacy tools & tactics » Quick Hits — August 14, 2011

  • Pingback: Feedcast Media Group » Blog Archive » Social Media News Round Up

  • http://pulse.yahoo.com/_G5PMGNXOG55F2G4H66GFHF4YRE Carlos

    The big question that many marketing professionals are asking today is. Can Twitter will be effective as a prime time tool in mainstream marketing!!??..

  • http://www.sparringmind.com Gregory Ciotti

    Adding something like “In case you missed it:” or a similar warning seems to be a fair compromise.