Social Media Research

12 Facebook Success Factors Based on Real Data

Social Media MagicWe track real-time data on nearly 20,000 Facebook business pages at Rival IQ, and I used this data set to determine several best practices for Facebook engagement.

This is not a random sample. Most of these pages we track are based on requests from RivalIQ users, but almost all major brands are represented.

What Does “Good” Mean on Facebook?

So how do you tell who is being most effective out of all those pages? There are several methods you could imagine. For this experiment we first eliminated any page with fewer than 5,000 fans. We then filtered for the top 100 pages based on Average Engagement Rate per Day (Total Engagement/Fans/Days in period) for the period of April 15 through May 14, 2014. Engagement rate allows us to compare a page with 10,000 fans to one with 1M fans based on the likelihood that fans interact with content on a regular basis

The spreadsheet (View in Google Docs) shows all the raw data. I then removed all non-English websites and a couple of duplicates and applied a category description. For clarity, I removed the non-English pages, not because they aren’t examples of good performance, but just because I can’t actually evaluate the posts without a ton of Google Translate time. This resulted in a final list of 60 Facebook pages (View in Google Docs) across these 15 categories:

Observation #1: News sites seem to be the dominant category with strong showing by Retail, Outdoor Recreation/Sports, and Fashion.

Observation #2: Virtually all of the top 100 pages ranked this way were what I would call B2C sites. Only a couple (Boom Social by Kim Garst and Volunteer Spot) would be targeting businesses vs. consumers.

Observation #3: The really big brands don’t seem to float to the top. There wasn’t a single top brand in the top list. The most recognizable included, USPS (likely promoted-see below), and several news outlets such as Huffington Post, Hello! Magazine, etc.

So what works on Facebook?

The number one overall page was “ICICI Prudential Life Insurance” – say what??  I hadn’t heard of them, have you?  Turns out as I’ve written about before on Convince and Convert contests can really jack up engagement and this is an example of this phenomenon.

Observation #4: Contests can drive significant engagement on Facebook.

Other FB Pages in the top 10 and what we can learn.

Company Name


Avg Eng Rate/Day


Engagement Type


Self Improvement



Feel good & Inspirational quotes





Nostalgia, Inspirational quotes










Humor – Promoted?

My Fox Orlando




Feel good stories





Shoe Pictures




Photos and feel good stories

Inked Magazine




Tatoo pictures – posts very frequently





Cute pet photos with frequent ask for share or like

Observations from the rest of the top 10 and what they can teach us:

  • People respond to positive quotes and human interest stories

  • Images work – only 8% of the top 50 posts did not contain a photo and all of these top pages were heavy with images

  • Ask for the share/like sometimes. We’ve probably all heard it, when you ask for the RT or Like, it happens more often. It is easy to overdo this but PetFlow demonstrates that it can be effective – or at least in combo with a cute pet photo.

  • While Facebook doesn’t tell you if something has been promoted, clues can include an off-balance like-to-fan ratio with no clear indication that the post went viral as indicated by lots of shares.

  • A weakness of looking at Average Engagement Rate per Day is that someone who posts A LOT can accumulate lots of small engagement numbers that roll up into a big number. Inked Magazine, Petflow, and several news organizations are examples of the phenomenon of posting 40+ times per day vs. Ziglars 4.5 per day. My own takeaway here is to evaluate sorting by engagement per post vs. per day.

Observation #10: Promoted posts can work.

A frequent customer request we get at Rival IQ is whether we can tell if a post (by a competitor) has been promoted on FB or Twitter. While you can see organic vs. promoted activity for your own Facebook page via Insights, neither Facebook or Twitter exposes publicly if a post has been promoted. But, there are public data points that strongly suggest a post has been promoted, such as a very high ratio of post likes compared to overall page likes.

I was surprised when this popped up on the list and a little more digging shows that there is a series of videos on retirement and saving that Principal Financial Group rolled out recently. These were likely promoted.

Another one… was this post from USPS promoted? This is a little less clear but looks like it.

Once you look a bit closer at all the top posts by USPS and how much engagement they have gotten it looks like this is a coordinated campaign.

Observation #11: Sometimes just plain cool works.

There are certainly examples where it is clear something went viral due to being just plain cool! Look at this image/post from the New Zealand Herald:

Observation #12: Quotes are working very well right now.

Several B2B FB pages like Kim Garst’s Boom! Social leverage inspirational quotes very effectively.

Facebook Success List

I could keep going for quite a while but I’m going to stop for now. If you are interested in digging into this collection of top companies like I did you can make a copy of this landscape in Rival IQ for free right here. I would love to hear what insights you find so please share with me in the comments.

The list of observations I collected includes:

  • When ranked by Engagement rate per day News sites seem to be the dominant category with strong showing by Retail, Outdoor Recreation/Sports and Fashion

  • Almost all top FB pages are targeting consumers or are B2C

  • Few really big brands made it onto the top list

  • Contests can drive significant engagement on Facebook

  • People respond to positive quotes, nostalgia and human interest stories.  You can use this even a business environment like Kim Garst does at Boom! Social

  • Images work – only 8% of the top 50 posts did NOT contain a photo

  • Ask for the share/like sometimes, be careful to not over do it but it doesn’t hurt to ask.

  • While not published it does appear you can identify posts that were likely to have been promoted.

Facebook Comments


  1. says

    Great study, John! Totally fits with what I found in Contagious Content (Six categories of things people share).

    And I’ve applied those lessons a lot lately to some of our clients- PrideStaff ( and Careers at Rx relief ( I think if they used Rival IQ they would have done well in your study.

    We find that the quotes + amazing images combo works well. But we also look at what themes each audience responds to and doesn’t in terms of engagement rate (likes/impressions). Every audience responds a little or a lot differently to the same content. For example, the pharamacists and pharmacy techs at Rx relief respond well to the topic of learning, but the PrideStaff folks don’t.

    We combine those types of posts with post promotion ads. And then more factors come into play- when you advertise, sometimes a post just gets a lot of photo view clicks. Each company will have to decide if that action is valuable given their campaign goals or not. In some cases, it’s not! And by the way, some of the images that do best in non-paid situations have too much text to get promoted- so that’s a trade off.

    • says

      Brian, thanks for the insights and great follow on reading. I absolutely agree that engagement varies by audience. To some degree that is part of what we are trying to help people with at Rival IQ, see how similar audiences (e.g. audiences of companies similar to me) respond to various content and what they engage with.

      Best, John

  2. says

    Brian, thanks for the insights and great follow on reading. I absolutely agree that engagement varies by audience. To some degree that is part of what we are trying to help people with at Rival IQ, see how similar audiences (e.g. audiences of companies similar to me) respond to various content and what they engage with.

    Best, John

  3. Mark Rowan says

    I would concur that Facebook B2B is more challenging to get right. But ingenuity goes a long way. Good info all around. Thanks!

  4. AdScientist says

    John, the idea is right and no doubt that engagement of total fans tells a story, but the foundation of the data is not as accurate as it appears based on the Facebook news feed algorithm.

    The best engagement rate formula you can use is engagement as a percentage of those reached (and if you can show non-fan only and fan only, this is even more ideal). There are many constraints within the Facebook news feed algorithm that prevent users from seeing the content organically. Sure, engagement triggers an increase in reach and thus more engagement, but that is over time and with time there is an increase in decay and decrease in content potential.

    Also as Brian said, page clicks is considered engagement, so even if users clicked to enlarge the image in the light box view (maybe to read fine print) or it was a gallery and meant for clicks, you will see spikes in the percentage who were “engaged” including the eventual like, comment or share.

    But I do appreciate the data we can learn from Rival IQ, Social Bakers and Simply Measured. Most of the time as a single company, you dont have access to client data to do the most appropriate of calculations, so what can be measured (that is useful), should be. Even as an agency team, you typically only have single brands in each industry across your portfolio to gather insights from.

    • says

      There is no doubt you can get a lot more data on an FB account where you are authenticated. A perfect analysis would be using that data including impressions, clicks, etc. across an entire analysis set. Alas we don’t live in a perfect world ;-). So you are right we do the best we can with the data that we can gather and as you indicated insights, albeit not perfect, can be gained from that data.

      Thanks so much for the comment and clarification. Best, John

  5. cnet says

    Great study. Can you help me with the formula though? Using the numbers you show in your article, I can’t replicate any of the avg engagement rate per day numbers you list.

    • says

      Hey there, the only way you will be able to dig into the average engagement per day numbers is by digging into the Rival IQ landscape I shared as part of this post. At a high level 1) Engagement rate = Total Engagement / Total Likes * 1000 2) Average Engagement Rate per Day is sum of engagement rate of all post during time period divided by number of day.

      You wouldn’t be able to back into the per day average number based on what is in just in the post or in the Gdoc spreadsheets I shared. If you were going to try to calculate it by hand you would have to sum up all of the posts during the time period analyzed (in this case April 15-May 14).

      Hope that helps. Best, John

Leave a Reply

Your email address will not be published. Required fields are marked *