How do the best-performing Facebook pages get all that amazing engagement?

And how can you and I do the same?

The Facebook pages that are doing wonderfully well with likes, shares, and comments on their posts have so much to teach about new tactics and worthwhile strategies. Our friends at BuzzSumo analyzed 500 million (!) of these Facebook posts, and we’ve learned some amazing takeaways that you can implement on your page today.

(I’ll be joining BuzzSumo’s Steve to chat with you live about this great research! I’d love to invite you to join us this Thursday.)

I’m excited to share with you what we found and the great new tests to try in order to boost engagement on Facebook!

facebook research

tl;dr – Introduction and Takeaways

We had the absolute privilege to work with the BuzzSumo team on gleaning insights from 500 million Facebook posts from over 3 million Facebook pages worldwide, collected over the past eight months. It’s been one of the biggest Facebook research studies to date, and we’re really excited to have learned (and to be able to share) our takeaways from the data.

A couple of notes beforehand:

  • With a dataset of this size, there’s a fair bit of figuring out to do when it comes to the many ways to analyze the data. We were able to collect average interactions per post for a great deal of different categories and post types, though even still — given the wide variety of pages in this study — we aren’t 100% confident that those averages are meaningful. Large, popular pages with enormous reach can tend to influence the data strongly.
  • Whenever possible, we chose to share the “most popular” findings rather than the “best” findings, choosing volume and popularity over interactions and engagement.

In terms of the specifics, here are some of the data findings from the study that I’ll be happy to go into more detail with further below:

  • The most popular time of day to post is the late morning and lunch hour.
  • The least popular time to post is late evening and into the night.
  • Tuesday and Wednesday are the most popular days.
  • Sunday and Saturday are the least popular.
  • We found 7 popular post types: Photos (with no link), Question Posts, Video, Giveaway Posts, Links (with or without image), Coupons/Promotions
  • The seven types of questions to ask: What, Why, Who, When, Which, Would, How
  • The most popular ways to post to Facebook are through Facebook itself, Hootsuite, Instagram, Twitter, and Buffer

Note: When we use the word “engagement”, it means likes + shares +comments. We’ve not included clicks in this metric since some of the updates here did not have links.

1. Test a new strategy around the most (and least) popular times of day to post

One of the first places we looked for insights into the Facebook data was noticing when these 500 million updates were published.

As you might expect, the majority of posts are published during the workday, and far fewer — up to 1/3 fewer updates — into the late evening and late into the night.

The most popular time to post is between 10:00 a.m. and 1:00 p.m. Eastern.

The least popular time to post is between 11:00 p.m. and 2:00 a.m. Eastern.

most popular time to post to Facebook

Here’s the complete breakdown of posting volume at different times of the day:

  • 1:00 a.m.: 3,992,938 posts
  • 2:00 a.m.: 4,252,584
  • 3:00 a.m.: 4,793,412
  • 4:00 a.m.: 5,467,362
  • 5:00 a.m.: 6,016,886
  • 6:00 a.m.: 6,705,801
  • 7:00 a.m.: 8,057,272
  • 8:00 a.m.: 9,867,708
  • 9:00 a.m.: 11,626,048
  • 10:00 a.m.: 12,649,227
  • 11:00 a.m.: 12,991,817
  • 12:00 p.m.: 12,440,312
  • 1:00 p.m.: 11,680,244
  • 2:00 p.m.: 11,200,169
  • 3:00 p.m.: 10,673,145
  • 4:00 p.m.: 9,711,319
  • 5:00 p.m.: 8,533,286
  • 6:00 p.m.: 7,561,044
  • 7:00 p.m.: 7,050,473
  • 8:00 p.m.: 6,405,524
  • 9:00 p.m.: 5,682,640
  • 10:00 p.m.: 4,939,076
  • 11:00 p.m.: 4,387,240
  • 12:00 a.m.: 4,011,171

Some possible takeaways

There are many different directions you might go with looking at this data — many different perspectives even among your teammates. We found that my interpretation differed quite a bit from our co-founder Leo’s.

When I look at this data, I see a huge opportunity for added engagement by posting when no one else is posting — the Late Night Infomercial effect where your post might thrive with less competition.

Posting when there’s less competition seems to me that it would help surface your post in the News Feed of more people. And from what I know about the Facebook News Feed algorithm, one factor for virality and reach is getting likes and shares on your content. The more likes and shares you might get during the night, perhaps the better head start you’ll get on News Feed content during the day.

Leo’s take was the opposite. He saw the most popular times to post and made a great inference that those times are likely also to be the times when the greatest number of people are actively engaged with Facebook.

Popular times equal large audience.

Both paths are wonderful options, made all the better in that they’re based on some great data here.

If you truly want to test the Late-Night Informercial effect, you can look at the numbers here and see when exactly is post volume at its lowest: 1:00 a.m. Eastern (and then schedule your post in Buffer, so you don’t have to stay up all night.)

If you’re interested in trying your posts when the audience is largest, you can check the data and see that 11:00 a.m. Eastern is the most popular time.

A couple further clarifications

1. We did look at the average engagement per post sent at each time and found there to be a strong rise toward the end of evening with a peak from 10:00 p.m. to 11:00 p.m. Eastern. We’re excited to dig into this data further as right now we feel it might not quite account for the varying reach numbers of the pages in this study – i.e., if posting volume is lower in the late evenings and a majorly popular Facebook page like Business Insider posts then, it’s quite possible that engagement could be inflated as a result.

2. The times listed here are in Eastern time, though it’s possible that the posting volume generally follows the same pattern for most any time zone. Eastern time figures to appear here as that’s where a large number of these Facebook pages exist.

When we generalized a subset of this data to look into posts sent just from French Facebook pages, we found that the numbers in this small subset matched those in the Eastern time zone.

So if you find yourself several hours removed from Eastern time, it could be great to test not only a schedule that fits with an Eastern time audience but also one that assumes your local posting volume would match the data here.

2. Which day of the week is most (and least) popular?

The next layer of the most popular time to post is looking a bit broader at the most popular day.

As you might expect, the vast majority of updates are sent during weekdays.

Tuesday and Wednesday are the most popular days of the week.

Saturday and Sunday are least popular, with about 40 percent fewer updates being published.

most popular day to post to Facebook

Some possible takeaways

Again, similarly to the above takeaways for the most popular time of day to post, when it comes to the most popular day of the week there could be a couple different directions here as well:

Experiment with posting on the most popular days (Tuesday and Wednesday), when you might assume the most people are online to see your update.

Experiment with posting on the least popular days (Saturday and Sunday), assuming there’s less competition and a greater chance to shine on those days.

You could even combine the two data sets to see the ultimate most popular time to post: 11:00 a.m. on Wednesday.

Same for the ultimate least popular time: 10:00 p.m. on Sunday.

A couple further clarifications

As with the time of day stats, we did notice that average engagement was highest on the days when post volume was lowest. We imagine there could be many different explanations for this, including the weighted influence of popular pages that publish during these non-peak hours.

3. Ideas on what type of Facebook post to try next

You can share a lot of amazing content on Facebook in a lot of amazing ways. And these different post types — text updates, video, images, links, etc. — may impact the amount of engagement you can expect to receive on an update.

Looking through the 500 million posts in our study, a few main categories emerged of different post types on Facebook. In our study, the most popular categories we saw were:

  • Photos (with no link)
  • Question Posts
  • Videos
  • Giveaway Posts
  • Links (with or without image)
  • Coupons/Promotions

Some possible takeaways

When I see this list, my mind already starts spinning.

How many of these post types do you have in your Facebook strategy?

I often end up bundling multiple categories here together. For instance, I’ll share a photo with a link and I’ll ask a question in the update text. That’s three post types in one!

What we found in the study is that people had success posting from each of these different types. It might be worth testing and experimenting to see which type does well for you and your page or seeing what’s possible and available for mixing up your main posts.

A couple further clarifications

In thinking about the Facebook News Feed algorithm, we’ve noticed that visual content has been a big winner when it comes to getting more reach. Looking at the results from our study here, we see that visual content is indeed some of the most-engaged content, along with questions.

What this data doesn’t quite tell us is the effect that popular pages might have on the types of content here. If, for instance, CNN is sharing super high-quality photos and videos with its big audience, you might expect to see the interaction numbers being high for photos and videos. And if lower-quality pages are sharing lots of coupons/giveaways, you might expect interactions to be low there.

4. The 7 Different Ways to Ask a Question on Facebook

Question posts were one of the more frequent post types we noticed in this study, and many seemed to do quite well (our early data suggested that they might be the most popular post type for pages).

Breaking down the questions even further, we found there to be seven main ways being used to ask questions in Facebook posts.

Popular questions on Facebook

  • What
  • Why
  • Who
  • When
  • Which
  • Would
  • How

Some possible takeaways

The list here reminds me of the Power Words we often reference when thinking about ways to craft our social media posts. The specific words you use — even when it comes to a quite general one like a question opener — can make a big difference in capturing the reader’s attention.

Secondly, a thought comes to mind here that one reason in favor of questions like “who” and “which” and “would” is they are simple questions that are easy to answer (often with a multiple choice or yes/no), and they also help strengthen the identity of the person commenting on the question—a key element of the psychology of sharing to Facebook.

5. Could a longer or shorter update work for you?

Seeing that Facebook’s character limit per post (reported to be as high as 60,000 characters) leaves tons of room for experimentation, we were eager to explore what length of posts tended to be most commonly used.

We ended up with seven different buckets of character counts:

  • 0-50 characters
  • 50-100
  • 100-150
  • 150-200
  • 200-250
  • 250-300
  • 300+

Some possible takeaways

In thinking about potential action steps here, I find it interesting to note where these buckets fall. I’m sometimes tempted to either go really long or really short with my Facebook posts – either multiple paragraphs or a short-and-sweet update. What I’m seeing here is that most Facebook pages tend to experiment on the lower end of the spectrum, 300 characters or fewer.

What does a 300-character post look like? Here’s one from Brain Pickings that comes in near 275 characters (another line of text would have likely made it 300).

brainpickings post

Thinking of this another way, could it be better to write shorter Facebook posts that fit with the majority of content being published by other pages?

Or might it be worthwhile to experiment with longer posts that are a bit less common?

6. The most popular tools people use to post to Facebook

One of the really interesting bits of data we were able to pull from our study was the applications used to post to Facebook (e.g., Hootsuite, Buffer). Here were the top 20 applications used to post to Facebook, ordered based on market share:

  • No application: 75.64%
  • Hootsuite: 5.12%
  • Instagram: 3.14%
  • Twitter: 2.74%
  • Buffer: 1.31%
  • WordPress: 1.08%
  • dlvr.it: 0.9%
  • Pages Manager for iOS: 0.88%
  • Sprout Social: 0.77%
  • Facebook for iPhone: 0.72%
  • Photos: 0.66%
  • Links: 0.65%
  • IFTTT: 0.56%
  • twitterfeed: 0.55%
  • RSS Graffiti: 0.44%
  • NetworkedBlogs: 0.38%
  • Facebook for Android: 0.37%
  • SocialFlow: 0.33%
  • Pages Manager for Android: 0.31%
  • Share_bookmarklet: 0.29%

Tools used to post to Facebook

Far and away, the most popular way to post to a Facebook page is with Facebook.

 

A couple further clarifications

One thing that this data does not account for is the impact of popular pages on these numbers. In looking at such a big dataset, we ended up reviewing both enterprise businesses and small businesses, each of which has its own engagement levels and own tools it uses.

For instance, here were the top applications used by individuals/small businesses, ranked by engagement.

  • Post Planner: 921.8 average engagement
  • Buffer: 180.67 average engagement
  • Sprout Social: 166.28 average engagement
  • Edgar: 83.53 average engagement
  • Hootsuite: 78.45 average engagement
  • CoSchedule: 75.53 average engagement

Here were the top applications used by enterprises (bigger pages, more reach), ranked by engagement:

  • Adobe Social: 1420 average engagement
  • Spredfast: 1270.89 average engagement
  • Sprinklr: 991.26 average engagement
  • SocialFlow: 881.11 average engagement

Overall, posts that did not use an application averaged 216.766 average engagement, while posts that did use an application averaged 113.88 average engagement. (Buffer posts had an average engagement of 180.67).

Does this mean that you should rush out and get Adobe Social? It’s a great product, for sure; my hunch here is that the engagement numbers are so large because the tool serves some really big pages with lots of reach.

7. How to think about the articles you link to

One additional tidbit we were able to pull from the data here was looking at the articles people clicked to read. Did linking to longer articles or shorter articles lead to more interactions on a Facebook update?

Sounds like a fantastic test to try. 🙂

In our study, we noticed four lengths of blog posts that were linked to most often:

  • 0-1000 words
  • 1000-2000 words
  • 2000-3000 words
  • 3000-10000 words

(For reference, the article you’re reading here is about 2,500 words long.)

The early data seems to suggest that longform content gets the best shares on Facebook, though we’d love to examine that a bit more before making any confident assertions.

For the time being, this feels like a great one to test and experiment with for your own page, noticing the effect on likes, shares, and comments when you send folks to short-form versus longform articles.

Even more: Further tools and talks

We’d love to share even more about this data and research and to make it really easy for you to start experimenting right away with some Facebook strategies.

We’re hosting a webinar with the BuzzSumo team this Thursday at 11:30 a.m. Eastern!

I’ll be sharing some further info about the study as well as answering any questions you might have. It’d be a fantastic privilege to have you join! (We’ll make the full replay available also if you RSVP and can’t make it.)

Facebook webinar button 2

Check out out our complete guide to the Facebook News Feed, including all the available information (up-to-the-minute) on how the algorithm works.

Try out BuzzSumo’s brand new Facebook insight tool, SumoRank. Enter any Facebook page (yours or others’) and see data on their Facebook rank and content engagement.

Over to you

I’d love your thoughts and input on this research.

What findings stood out to you?

What did you learn here that might shape your upcoming strategies and experiments?

I’ll be hanging out in the comments, eager to hear from you!

Image sources: IconFinder, Pablo, Pexels

Looking for a better way to share on social media?

Schedule, publish & analyze your posts across the top social networks, all in one place.

Get started for free
Written by Kevan Lee

Director of marketing at Buffer, the social media publishing tool for brands, agencies, and marketers. We’ve got a new podcast! 🍟

  • Tim Manning

    Great research guys!

    • Thanks so much, Tim!

  • Fantastic write-up and findings Kevan (and to the BuzzSumo and Buffer teams)!

    I’m curious to see what your findings are re: #7 and the length of posts.

    Neil did a great post about this last year and found that once a post’s word count exceeds 1,500, it’s in the “golden share zone”.

    Here’s the link -> https://www.quicksprout.com/2014/03/31/how-long-should-each-blog-post-be-a-data-driven-answer/

    • Hi Chris! Wonderful to hear from you! Yes, for #7 we found some really cool information on the early data that suggested longform content does better with shares on Facebook. I believe 2,000 to 2,500 was the sweet spot! Sounds quite in line with Neil’s findings, too. I’d love to vet the data a bit more fully. Early signs are promising!

  • Very interesting! I’ll have to start posting more on Tuesdays and Wednesdays!

    • Hi there Angela! Awesome! Would love to hear how that schedule goes for you!

  • Bryan Davis

    Great post, Kevan and co. Excited to see two of our favorite tools teaming up for a great report.

    • Wonderful! Thanks so much, Bryan! Yes, it was awesome to partner with BuzzSum on this!

  • I’d like to preface my comment by offering my sincerest gratitude for all the research and analytics you’ve provided – for free! Awesome.

    Okay, so the majority of the businesses I work with are small manufacturing companies and have less than 5,000 followers. I find that the best times to post for both reach AND engagement is M-F between 6-7PM across almost all of our clients. Sunday nights around 9PM earn high stats, as well. So imagine my surprise when I read that the most popular time to post is late morning and lunchtime!

    My initial thought when I saw the stats about the most popular time for pages to post was that this time slot is most convenient for THEM, not their audience (especially for those that do not schedule posts in advance). Then I realized that many large brand pages are catering to a much more diverse audience that that of a small manufacturing co. Our audience is very niche, and we cater to the decision makers who maybe don’t have time to check FB at lunch, for example.

    So, I guess the whole reason for this comment is to suggest that the results are segmented by industry or audience size. I don’t know if that’s even possible, but the stats would be invaluable. I’d definitely be willing to dole out more than a few bucks for it, and I’m confident I wouldn’t be the only one! So, I’ll be keeping an eye out for even further data segmentation. (I read your blog like, every single day. No lie.) If that’s not possible, then I just want to send out a huge THANK YOU for the constant gold insights. Yall should be proud.

    • Great point. There are certainly big differences industry to industry and page to page. It’d be nice to see how our own look.

    • Hi there Miss Jules! Thank you so much for the comment! I totally see what you’re saying here, it’d be awesome to segment the data in this way to make it as valuable (and actionable) as can be. Would love to maybe do a followup post on what we find there!

      Love your point also about the most popular times of post in consideration of how few people might be using a scheduling app. Makes sense then that the most popular times would be the times when most people are at work or logged onto a computer!

  • Lots of interesting data here. Thank so much for putting it together.

    Leo’s take on point 1 is great. Between 9am and 1pm CST is the time people are most likely to engage with a page, according to Facebook. It’s the time most people login and interact with each other and brands.

    Was this data pulled using BuzzSumo’s new Facebook tool SumoRank (or through the same means it uses)? It doesn’t appear that it can differentiate between purely organic posts and those that were posted organically and then had paid put behind them. If so, this could have a very big impact on the data set for a couple reasons.

    Putting paid budget behind a post can greatly increase engagement. Even a post that no one would have engaged with will see great engagement if you put money behind it. Amount of spend, targeting and more will all effect these numbers.

    Paid spend also makes the post time and day irrelevant. You can post any time day or night, on any day of the week and then put paid spend behind it and get good engagement numbers. SumoRank would note that the successful post was published at a specific time but the paid spend and engagements could have come much later on and at completely different times.

    Very cool when Buffer digs into this kind of data. There’s so much to be learned here. Looking forward to the webinar. Any chance you’d open the data set for others to dig into and see what they may find?

    • Hi Ben! Thanks so much for the comment! Love your deep analysis into the numbers here, and that’s a really great point about organic vs. paid. I’d love to check with the BuzzSumo team for the specific answers to that question. I believe the SumoRank tool works off of a similar setup that was used to pull all this great data. Would love to double check on that!

      • Thanks so much Kevan. Really interested to see what else you guys find by digging into all this data. There’s certainly a wealth of insights to be had here. 🙂

  • This was such a fantastic post and so informative! I tend to think like you — seeing opportunity to post in those slower times. Then again, most of the social media clients I work with are small businesses so I prefer to post at times where there is less competition. Thanks for sharing all of this information, I have some great things to test now! I also really enjoyed the SumoRank tool – what a fantastic snapshot of data.

    • Thanks so much, Stephanie! Fun to know we’re of the same mind on these things. It’s so interesting to me how many different ways folks might see the takeaways here! Would love to know if you end up experimenting or see any new results!

  • Cool post Kevan! Ideal posting times are always hard to predetermine and assume that overall data like this is good for all pages. Page owners need to look at their own insights and see what works best for them.

    Lost in this data is popular is not always the best IMHO. I say post as often as you can get engagement!

    I love seeing that Post Planner is crushing it for engagement for small businesses, which is exactly what we want to do!

    • Hi Scott! Great to hear from you! Yes, congrats on the awesome Post Planner results! You all have a great tool and some really great strategy behind it. Love your thoughts here also on popular does not always mean best. Very true!

  • Great post Kevan! Lots to take onboard…

    • Thanks so much for the comment, Graeme!

  • Kirsten Grosman

    Excellent analysis thanks for sharing. Makes for a very interesting read.
    Thanks for the doing the hard work Buffer and BuzzSumo 🙂

    • Hi Kirsten! Thanks! Yes, big thank you to BuzzSumo for laying some great groundwork here!

  • I absolutely see better results when posting at less popular times. My go-to time is 2:15am my time (4:15am EST).

  • Thanks for the great post (as always), with incredible detail. It’ll certainly be helpful for planning our next set of campaigns.

    Would you be able to expand on the following items with some more data?

    1. Post lengths – Which were the most linked type?
    2. Update length (character count) – what was the most popular /most shared?

  • B-SimmS

    How do you get your website to show up on SumoRank? Is there a certain level of engagement?

  • This is a power-packed post, @kevanlee:disqus and BuzzSumo! Thanks for taking the time to do this 😀 I just started a Facebook Page for a side-project, so this is really helpful. I’m still trying to figure out the best times to post, being on the other side of the world and all, and am experimenting. Your analysis (contra Leo’s) will be helpful in helping me make sense of what I find. Thanks so much!

  • Marnie Swedberg

    Terrific article. Excellent, actionable content. Thanks!

  • Interesting study. I use Buffer to schedule to all but FB. I’m a small business creating handcrafted unique jewelry and slowly building my page to 2,5k. What you omitted are inspirational & motivational quotes. Maybe the large B2B companies don’t need to use them, but fans of small entrepreneurs LOVE this type of post. Everyone seems to make great graphics using Canva & despite other types of posts, these are the ones that we need to use to get engagement.

  • Great post Kevan!

    I’ve always found that insights show more people are online in the evening so the findings here make me think of two things:

    1) The data here is wrong 😉
    2) Most Pages (businesses) are posting when they’re ‘in business’.

    I think (and I’ve tested it) that most users (profiles) are online in the evening. I have several B2C Pages that I Buffer to and they’re set to post around 8pm.

    That’s working well for me and actually the Sunday point interests me; we get MOST interactions (or engagements as you call them) on Sunday evening!

    I think posting when less people are posting works but I also think that so many business Pages post in the day and that has skewed this data.

    What do you think? I’m going to try the Wednesday at 11am though (UK time) just for sh*ts and giggles 😉

  • cgwellness

    Great article, thanks. My page http://www.fb.com/CoachMD was averaging 5k – 7k new likes a week, with a reach of 3 – 7 million, and a people talking about of from 250k – 500k. This was steady and growing for the first 6 months of 2015. In July, those numbers plummeted to about a tenth of what they were – same content/same strategies. It seems for the past 6 months, I have been spinning my wheels trying to regain my reach. I recently joined PostPlanner which is very convenient, but I am not so confident that it will bounce me back to where I was. Any thoughts would be greatly appreciated. Best, Charles

  • Braden Allred

    I love this kind of stuff. Analytics and finding out the how and why behind the what… Thanks for sharing this!

  • Neat post…. Although it’s the kind of post that gets my clients asking when’s the best time to post. Which is fine but….

    When page is getting between 1000 and 2000 organic reach but paid promotion is getting it seen in the 50k range…”what time to post” becomes kind of pointless right? ( not saying all pages are like his)

    Good content gets shared any time it’s posted, and Facebook weights their algorithm to share that content more widely. This counts way more than what time to post.

    Focus on better content, not the ” what’s the best time to post” game. It’s really low level thinking.

    Most pages are posting when they’re in business, so this swings the results quite heavily too.

    Still a great post… I love he in-depth content from buffer 👍

  • dimosthenis karaflos

    if you want a quick post analysis checj this out

    http://www.facebookbi.com

  • KB

    @kevanlee:disqus – wondering about your numbers for engagement in “a couple further clarifications” section. Did you factor in the follower size of the social properties, or the algorithm score of the properties who were being tested in the engagement score. I agree that the Adobe Social users probably have larger followings, hence the larger engagement numbers. It would be interesting to see the engagement ratio numbers, that is — number of engagements compared to each social property’s followers. That would seem to be a more accurate test on if 3rd party software affects reach and engagement. Interesting article.

  • DLT

    Re: Facebook being used the most to actually post as opposed to a third-party SMMS: I wonder how many businesses or organizations are ONLY using Facebook and no other forms of social media? That may sway the results slightly, the same way Etsy businesses might ONLY be using Pinterest. Many small businesses and nonprofits have not explored other platforms or feel comfortable branching out to other platforms, and Facebook offers the most variety for a business or organization in terms of promotional value.