Does Using a 3rd Party API Decrease Your Engagement Per Post?

[Updated: 12/01/11 11:25AM ET] // We did a very preliminary study now that Facebook has changed the news feed and has been made aware of a particularly important bug. We outlined our findings here: Did Facebook Really Fix The 3rd Party API Penalty?. At a preliminary glance, it appears that from an Impressions perspective, 3rd Party APIs may have regained equal value. //

[Updated: 11/10/11 6:45PM ET] // Facebook has reportedly adjusted EdgeRank to reduce the negative effect of posting as a 3rd Party API. In effect, this validates our findings while hopefully removing the negative impact that 3rd Party platforms were experiencing. We will be following up on this information to test if Pages are actually experiencing this change. //

It’s been a general consensus in the EdgeRank community that 3rd party APIs (this simple list provides examples of potential 3rd Party APIs) are punished in Facebook’s EdgeRank algorithm. We ran a study across our entire data set to provide evidence that our hypothesis is true. The result of our study was quite eye opening. Using a 3rd party API to update your Facebook Page decreases your likelihood of engagement per fan (on average) by about 80%.

The Details

We ran our analysis on 1,000,000+ Updates on 50,000+ Pages that influence over 1,000,000,000+ Fans. We took each individual post and analyzed the engagement (comment & likes) along with how many fans the Page had at the moment of updating. We then analyzed Facebook vs The Top 10 Other APIs (based on our sample size). We did a weighted average for the 3rd Party APIs, which resulted in 0.22% engagement. The result is a percentage that represents engagement per fan per post (on weighted average, of our Top 10 3rd Party API). Our sample size ranged per API, although we cut it off at the Top 10 APIs outside of Facebook.

Why are 3rd Party APIs so much lower?

We have a few theories explaining the results:

  • Facebook Penalizes 3rd Party API’s EdgeRank
  • Facebook Collapses 3rd Party API Updates
  • High Chance of Being Scheduled or Automated
  • Content is not Optimized for Facebook

Facebook Penalizes 3rd Party API’s EdgeRank

When an object is created in Facebook, it is assigned a weight. We believe that Facebook strategically reduced the weight of objects created through the API. The reason behind this strategy would be to encourage more content creation within the Facebook Platform. This ultimately increases the value of their platform while increasing ad impressions.

Facebook Collapses 3rd Party API Updates

When the same 3rd Party Platform has multiple updates within your feed (regardless of the Page or People who created the object), Facebook will collapse the objects and only display a single object. This can potentially kill visibility for objects that are caught in this collapse.

 

High Chance of Being Scheduled or Automated

Posts that are scheduled typically struggle to have high engagement. This is most likely due to the nature of a scheduled update. It’s difficult to create unique engaging content several hours or days in advance. Any negative impacts of scheduled posts are most likely correlations with poorly developed content.

Automated content typically performs horrendously, due to lacking human touch and craftsmanship. These types of updates struggle for engagement as is (even if they were manually posted onto Facebook). We highly advise against any automated content on Facebook.

Content is not Optimized for Facebook

One of the conveniences of a 3rd Party Platform is that you can simultaneously update Facebook, Twitter, LinkedIn, your blog, and hopefully not MySpace :P. Unfortunately, this typically requires the social marketer to optimize their content for all the social networks at once. Twitter has character limitations at 140, Facebook allows for many more characters. Some networks allow photos to be attached, while other networks do not. This distracts the social marketer from specifically optimizing their update for Facebook.

Should I Stop Using a 3rd Party API?

Most 3rd Party APIs are still a great tool that we recommend while consulting social media marketers. In our opinion, 3rd Party APIs are to be used in moderation and not in all situations. Scheduled updates work great when there is no possible way for you to update the account on time. It’s better to have a scheduled post go out, then no post at all. 3rd Party APIs still work great on other social networks.

What About the 3rd Party APIs Outside of the Top 10?

We decided to focus on the Top 10 3rd Party APIs. The Top 10 was based on sample size of our data. A few Platforms that didn’t crack the Top 10 (due to a small sample size) performed better than the Top 10 average. However, we hypothesize that these 3 Platforms’ engagement ratios would decrease with a larger sample size.

How Do I Take Full Advantage of this Study?

In order to maximize your Facebook marketing efforts, we advise that you post directly on Facebook.com or use an Official Facebook App (iPhone, Blackberry, etc). Posting directly will give your posts the greatest chance of engagement (which drives EdgeRank). Don’t immediately abandon 3rd Party APIs, they still have their place in social marketing. We recommend not relying on them day-to-day for Facebook updates.

98 Comments on “Does Using a 3rd Party API Decrease Your Engagement Per Post?

  1. Thank you so much for sharing this data – it’s good to see the research. Your summary in terms of posting the Facebook directly and then using an application (ideally an official Facebook App) to post for example when mobile is great advice.

    I for one still value 3rd Party APIs especially for their Twitter application and it can still prove helpful for scheduling content to Facebook if absolutely necessary.

  2. I find it hard to believe that Facebook should have developed a system for penalizing updates from 3rd party tools. As you mention in your conclusion, it’s probably more likely that scheduled posts are not as relevant as other posts.

    • Chad Wittman

      It’s hard to validate our hypothesis that it is in fact being punished. However, it’s quite a drastic decrease for all 3rd Party APIs. We’ve been asked “what about custom publishers?”, I think that may be our next study. If custom publishers “bridge the gap” sort to speak, I think it would lend that 3rd Parties may not be punished. But until then, I think the theory is relatively sound.

    • Why is it hard to believe?

      Facebook wants eyeballs on its site for ADVERTISING. If you are using an app to post content, then you are not on the site, and hence, you are not viewing the ads. Why, then, is it hard to believe that they would devise a system to reward you for using their site directly as opposed to populating their site with content using offsite apps.

      and it’s not just scheduled posts that they collapse.

      • Danny

        That’s not true. The platform through which posts are published has no effect on where the posts are seen. Posts are seen in the news feed, or the brand Page, regardless.

      • Patti Majeski

        I agree 100%.

  3. Kaitlyn

    Thank you for this great case study, I have always wondered about this because I know other 3rd party apps do not rank well on Facebook. I will share this with my clients, thanks again!

  4. Ben Teoh

    Intersting post.

    I still prefer to do most of my Facebook work directly in Facebook. If there are cases where I need to schedule something I will. But, with all things social, scheduling needs to be well balanced by live posting.

    Thanks for doing the research. Great work.

    • You make the mistake of confusing correlation with causality here. Facebook says it only shrinks objects that it deems as feeds, which makes sense, given the experience of users like myself and the others posting in these comments. I post 8-10 updates on a page through a 3rd Party API daily, and I check my work regularly through an account that doesn’t admin the page. I’ve never seen posts from my page shrunk. If I ran an RSS feed into the Facebook, or if I posted content that was so unoriginal it looked like a feed, I’d get shrunk. The way to avoid this is to post interesting, engaging content, as you note. That’s really the bottom line with everything when it comes to Facebook.

      • Chad Wittman

        If the data wasn’t so drastic, I’d possibly agree. However, there’s a substantial drop off from using Facebook.com directly to using any 3rd Party API. This drastic decrease is too much for me to write off as being solely content correlated.

      • In my experience, “Interesting, engaging content” is still collapsed, regardless of whether it is a feed (real or perceived), when it is posted/shared /updated on Facebook using “via (anything)”.

        I have seen regular people’s stellar content get collapsed because they use an app to share to Facebook rather than using the FB site directly.

        I have also seen Feeds get collapsed.

        What I have seen survive being collapsed is if it is posted/shared/updated on Facebook using the Facebook site itself rather than any app. That (using the FB site directly) is the ONLY way that I have seen to avoid anyone’s content from being collapsed.

  5. Do you know if it punishes you as bad if you are using a 3rd Party API for only its scheduling features? Meaning you are optimizing the post for Facebook, except in the 3rd Party API?

    • Chad Wittman

      That’s a great question Isaac, and something we internally discussed after going over the results. It’s really hard to tell from the data, but I would venture to say that you risk collapsing along with a potential API penalty. So in effect, even those updates can be punished, however I’d argue the negative impact would be less intense.

      • Interesting, personally I’ve not seen 3rd Party APIs collapse recently, but I suppose it is a potential. Personally all our clients pages that we use 3rd Party APIs on maintain an above average or above excellent Edgerank score (guess we are doing something right) with the exception of one (and that one has other issues).

        We get a lot of interaction on our post, and given the amount of clients we manage it would be difficult to do them all organically. Guess if it starts being a bigger problem we’ll have to look at other options. Wish Facebook had it’s own scheduler like WordPress.

    • Anything that is posted “via something” is a third party app and gets collapsed. It doesn’t matter if it is scheduled or not – it is HOW the content is populated. If it comes in “via” an application, then it is less favorable content compared to that which is populated from the Facebook site itself.

      The only way to avoid it is to use the site directly.

      See above reply (in different thread) for Facebook’s reasoning why.

      Conclusion: Take the extra time and post your Facebook content to your pages by hand rather than using a 3rd Party API.

  6. ryan

    Thank you for sharing this information. I wasn’t aware that API were getting hammered by Facebook. I’ll be sure to share this with my clients.

  7. Larna

    Thank you for the post and the research… It will be a valuable resource for supporting my recommendations.

  8. Daan

    Hi Chad,
    great article. One explanation for this result may also be that fan pages that use 3rd Party APIs are professionally managed fan pages with a large number of fans that have lower engagement by definition (e.g. Coca Cola vs. my local coffee shop).
    What I love is that you guys are doing some rigorous testing! Keep it up!

  9. This is something we’ve been aware of for quite sometime at hubze. On our main Facebook page we never post through a 3rd party and do it all manually. Sure it’s more time consuming but we never want one of our posts to be in the “See 40 more posts from Networkedblogs” category..

    And even if you haven’t seen it doesn’t mean it doesn’t happen. Just means you aren’t following alot of pages that are using it.

    From a social manager standpoint I assume everyone will see this collapse so it’s always best to post. My theory is use the 3rd party platforms to interact, but only use Facebook to post.

  10. Thanks for an enlightening post. I wasn’t aware of this potential problem. I love 3rd Party APIs for vacation time – this summer I didn’t always have access to internet yet still published and posted critical information by scheduling.

    However, now I know not to rely on it for my daily posts.

  11. To my knowledge, collapsing has nothing to do with which tool was used and everything to do with volume and automation. If you post 10 times a day from FB’s website, they will still get collapsed for some users. If you post 10 times a day with similar content, collapsed for more users still.

    The proper benchmark here would be posting the same content to two different pages with a similar fan base, at the same frequency from various tools. Otherwise the 70% less engagement claim seems very gray.

    This would also suggest that photos posted from 3rd party tools such as Instagram, or game updates would be equally poorly weighted, which just isn’t the case.

    • Chad Wittman

      If a user is a fan of multiple Pages and each uses the same API, the user may experience these updates collapsed.

      I like your idea of the proper benchmark, and our data set maybe able to accomplish this. It will be a much more time intensive task, but we will consider studying it.

      Bear with me on this, but Instagram does not post (as Facebook defines) “photos”. Instagram actually posts links with a thumbnail to the image. Even if Instagram posted photos within the feed, Facebook would still register this as “via Instagram”. Which could in effect decrease EdgeRank.

    • If it had nothing to do with what tool was used then it would say “See 31 more updates from Justyn” rather than “See 31 more updates from NetworkedBlogs” or “See 22 more updates from Foursquare”. When you do click on those to expand them, you find SEVERAL users’ content, not just one person. For instance, right now I have “See 109 more updates from NetworkedBlogs” on my feed, and when I click that, I get a whole screen full of a 100+ different page’s blog posts, not just one person that has posted 109 different things recently.

      It has everything to do with the “via” and (almost) nothing to do with the user. Yes, if an individual user/page populates a lot of content within a short amount of time, it may get collapsed (rare), but anything shared “via” an application will (often) be collapsed.

      Facebook wants you to be using their site directly, not via apps, and has devised a system to encourage you to do so (because they have a financial incentive to do so for the advertising).

    • Justyn, glad you stepped into the discussion. Upon reading the report, I immediately thought of you and the sprout team. Hope you’re doing well.

  12. Kemp

    THE REAL QUESTION:
    Are posts from 3rd party tools responsible or are those who use automation tools more likely to be absentees on Facebook and less engaging?

    I haven’t seen 3rd Party API’s posts collapsed in at least 8 months and would love to see any examples. As I understand there are preferred Facebook developers that no longer have their posts collapsed.

    • Chad Wittman

      That’s a great question, that there isn’t a clear answer for. It’s not confirmed that preferred developers don’t get collapsed. Would be interested to see when that went into effect. Either way, if it’s been 8 months, a majority of our data would have came out of the past 8 months.

    • Chad Wittman

      As I said in the post, the general consensus in the industry was that something was happening, but this is one of the first studies addressing it.

  13. Julie

    You guys really got to check Post Planner. We’re a third party app but we aim at making FB posting easy, and we really mean it: no collapsing in the NewsFeed for those users who subscribe to our White Label version… Only saying :-)

    • Chad Wittman

      Your company showed up in our data (sample size was smaller than the Top 10), but your platform definitely had positive trends. We should connect, I think you would like to see the data on your platform (I’ll send an email).

      • Julie

        Hey Chad, just thought I needed to make something clear in here first: posts published through a White Label app do collapse in the NewsFeed, and we see this as an advantage. However, posts published via two or 20 different White Label apps will NEVER collapse together in the NewsFeed…
        Will reply to your email now, thanks Chad.

  14. The fact that Facebook are now grouping posts together is true, but this is only happening if a user filters their news feed by “Most Recent”. Having said that, we have not seen a drop-off of engagement with Sendible users. In fact, our users seem to be getting better engagement ratios when posting using Sendible under their own branding, than on Facebook directly.

    One of our users is the largest gaming company in the world with over 50 million Facebook fans and they are posting using Sendible, but under their own branding. (This is one of the features we offer – see http://bit.ly/sendiblerebranding.) They have seen higher levels of engagement than they would’ve seen had they been posting manually on Facebook.

    • Chad Wittman

      Thanks for chiming in Gavin. I suspect that white-labeled publishing may perform at a higher than level than it’s 3rd Party cohorts. We’ll have to analyze it, will take some time to dig through all the different custom publishers.

      I want to mention that while releasing this study, I had conflicting emotions. I value and admire platforms such as yours and believe that they drive innovation in our industry. What’s unfortunate is that Facebook is definitely doing something against these platforms. My hopes are that Facebook begins to reward platforms as opposed to penalize them.

    • I use “Most Recent” for my news feed always and nothing is collapsed in my feed… I chose to change the default setting to “All friends and pages” and that’s what makes the difference…problem is, not everyone is going to do that and many don’t know they can!

    • Gavin, I have been using Sendible and everything was fine until about 8 days ago. Now the Sendible posts aren’t even being collapsed. Now they don’t show up at all. I maintain a number of client accounts so I can easily check when a post shows up. None, not any, of the sendible posts are showing in any user feeds. I can post something via sendible and nothing..then wait and post directly and then immediate interaction.

      I’ve mentioned this to sendible support and they have said they are certainly aware of this…but said it might be “weeks” until, or if, it is resolved. It pretty much renders the sendible platform useless.

      I have to say I did see great results with sendible until FB made changes but now it’s a dud.

      I see that Gavin recommends the branded version but one would be crazy to pony up additional cash for a branded version until one knew for sure that at minimum the sendible platform would be viable.

      However, sendible works for twitter…but the fee makes it kind of expensive to use simply for twitter.

  15. Heidi

    Hi there Chad. I was wondering if you could talk a bit about the research you did because my findings don’t show the disparity that yours do. We have looked the engagement on our clients pages and have found very little difference between those posts that come from our platform (Sendible) and those that come direct from Facebook. Granted, impressions changed but the actual engagement did not-which in my humble opinion is the important thing here. I’m not trying to come off defensive but since I am writing on behalf of a 3rd party platform I understand I may sound a bit biased. I am genuinely curious about this and would love to hear more about your research.

    Thanks!
    Heidi

    • Chad Wittman

      Absolutely, that’s what this data is for. We have a large data set of anonymized data for 50k+ Pages. We simply took over 1M Facebook Posts and measured how many engagements the post received, how many fans at the time of the post did the Page have, and where did the update come from. We took the total engagement / number of fans and averaged it out per post (for each of the APIs).

      This was simply the result of the study. I outlined why I think the data maybe this way in the blog post. Also, consider that news feed optimization experts have had this general consensus for sometime now, this study only lends evidence to the claim.

  16. I strongly suspect there’s a bias in this data producing results that are more simply explained by the content types 3rd party clients are able to post.

    Almost all clients only post text-type updates, which have one of the lowest starting EdgeRank “weights”. Links included in these updates will transform the update into a link-type update, but there’s no ability within clients to edit the link title & description, or choose the picture associated with it — key elements that affect engagement with link-type updates.

    A better analysis would be to analyze only text-type updates, and see if there’s still as big of a difference in engagement.

    • Chad Wittman

      I agree that there could be interesting data hiding in our study in regards to post type & 3rd Party APIs.

      I’m interested in possibly testing each post type against each API.

    • Chad Wittman

      Jay, I decided to try out your suggestion. We parsed out the data to remove photos, videos, etc. The difference between a Facebook Status & a 3rd Party API status was still close to our weighted average.

    • Lots of 3rd party apps have the ability to post photos and change the link info, etc. I use Hootsuite and it does all that. Maybe you need to find a new app :)

  17. Isn’t it true then that content only gets collapsed if there are several posts within a short time frame? Also, I have a few clients who only post on FB via a third party API and quite frankly, the engagement for them is adequate. It seems to me that at the end of the day, engagement is driven by the relevance of the content to the reader no matter how it gets posted.

    If I am representative of other FB users, I follow the pages/ppl I find of interest and will indeed click through to read content if the subject line grabs my attention. I hardly see the benefit in FB punishing anyone for using the 3rd party APIs since it does not keep others who are intent on being on the site from being there. Anyway – I have stopped trying to figure out what FB does the things it does – after all it is still for the most part a free tool – instead, I try to truly make my experience as social as possible and take advantage of any business opps that might surface as a result.

    • Chad Wittman

      Not necessarily, we’ve seen cases of updates over 24 hours apart still being collapsed. As for your clients, I could expect some of them to break the average and hover closer to Facebook’s percentage. But if there’s a risk of being penalized, why take it? You won’t know your potential success unless you also test posting directly from Facebook.com.

      There’s a variety of reasons that Facebook would penalize APIs. It might not even have to do with a intentional punishment. Hypothetically (very hypothetically), their algorithm could decrease the weights of APIs that receive high (hidden on the news feed) notifications. Only APIs can be automated, therefore they would be the only option to be punished by an automatic sliding scale built into the algorithm.

      As I said in the post, it has been a general consensus for sometime now that 3rd Parties are penalized somehow. What we’re revealing isn’t new, we’re just providing evidence to that claim.

  18. Enrico

    thanks for the tips for a better time

  19. Paul Cowan

    @Chad

    Very interesting study. Of the 50K source pages, what was the split between personal and corporate pages? I’ve found engagement rates to be much, much lower on corporate pages. In fact, the average engagement (like+comments) rate on Fortune 500 brands is .00079%. This study could lead brand community managers to inappropriate conclusions if it includes (and skews toward) personal, niche communities and large brand communities.

    • Chad Wittman

      We only have data on Facebook Pages, meaning we don’t have any data on Personal Profiles. The Facebook Pages range from Fortune 500 companies to Mom & Pop type presences.

  20. Thank you for posting. I had always wondered about this. Great info to know for all social media gurus out there. :)

  21. Very grateful to have this research, thanks. A game changer for social media marketers, and it will certainly change our timetable in the office. Must be one of the few companies where we will be actively encouraging employees to now spend more time on Facebook. :)

  22. Dan

    What about using CoTweet?

  23. Dan

    Oh, nevermind….that’s not applicable.

  24. Was there an attempt to account for other potential factors and adjust the data for things like time of day, frequency of posts and length of posts?

    • Chad Wittman

      With 1,000,000 posts analyzed, those types of variables most likely averaged out. We’re discussing a few alternative studies to help dive into this deeper.

      • Sharon Machlis

        Thanks for the response. However, I’m not sure I’m comfortable with saying, in essence, “Since there were a lot of posts, by definition that means we’re confident there are no other important differences between posts from third-party tools and posts on Facebook.com except the platform used.”

        I don’t think those two groups of data can be assumed to be randomly assigned and thus equivalent in every other meaningful way, any more than I’d make that kind of assumption about people posting by desktop computer vs. mobile device.

        • Chad Wittman

          This finding isn’t really new, it’s just been finally done on a large enough sample size. Before we released this data, we knew of another independent social media analytics company that found similar results. Our data lined up relatively comparable, so we felt even more confident in our data. Since we’ve released this, several other companies have come out of the wood work to acknowledge similar results as well.

          This isn’t an isolated study, completely unverified. Nor is it the final study to solve this question. It’s simply more evidence to support an already widely accepted theory. We will continue to dig deeper to create a more well rounded case.

  25. Chad,

    I would love to find out if my Application appeared at all in your data. theSociable.me I there a way that I can find out how my Posts are Trending?

    • Chad Wittman

      I took a quick scan, looks like none of our data used your API to post.

  26. Hey, Chad. Thanks for this data. I feel like my own experiences on my largest client’s page (~115K fans) support your data, but my experiences with smaller pages does not, necessarily. This is not empirical yet, just my gut.

    I wanted to ask you something else, though. My own experiments on that large page tend to make me think that posting using the official FB Android app results in more impressions than posting from a desktop browser. I’m convinced enough that I regularly transpose text-only posts and submit using my phone, even if I’m sitting at my desk. Have you considered doing a study similar to this one that would examine this possibility?

    • Chad Wittman

      Thanks for the question Joe, I can’t comment on any specific 3rd Party APIs. My apologies.

    • Chad Wittman

      Yes, we had some minor edits that were made.

  27. Heya Chad, I posted a top-level response here: http://www.stagirainc.com/2011/09/facebook-edgerank-calculation-missing-link/

    In response to the Lazerow AdAge article “Posting to Facebook: The Truth about Third Party Applications” response to your report, there is one point I want to emphasize: Facebook is and always will be a product in progress, and in this, the facebook platform team cannot comprehensively monitor, measure, optimize and rework the UX (or any other aspect) of the platform when users (regardless that they may be using facebook on behalf of a brand) are, more and more, accessing the platform, off platform via API. It’s widely held by just about every developer on the planet: Facebook’s API’s suck. But they suck by design and for a reason.

    This is why I believe Google is no hurry to open API’s in and out of G+, one of the chief lessons Google learned with Buzz.

    Identity and especially Proximity need to be added to the EdgeRank Calculator equation.

    two cents spent, Jason

    • Chad Wittman

      Completely agree about Proximity! That would be an excellent addition to the EdgeRank algorithm (if it isn’t already in there).

  28. Interesting article… I think some of the commenters are confusing “shrinking/shortening” with “collapsing”. Networked blogs is the worst for this in my feed. When I have something I really want people to see, I wait a couple of days and then post again from within FB, just to be sure.

    However, I think if you have your feed settings set right, it won’t collapse anything? I seem to remember changing something on my feed setting a while back and I don’t get anything collapsed in my feed now. BUT, the question is, did your customers make that change? If not, you’re in trouble, and the chances are they did not.

    I just checked my impressions on my own page and I’d say it’s about even between those published through Hootsuite and those published from FB. Some are really high from each side and some are really low. So hopefully that’s a good thing? and FB isn’t punishing me too much for using HS.

    Looking forward to future installments on this study.
    Anita
    Head Twirp, Twirp Communications
    http://www.twirp.ca

  29. joerg

    One way around the punishment of 3rd party applications would be to simulate a user who does his posting. If facebook punishes usage of the API, so let’s not use it then.

  30. Thanks for this research. Good food for thought when considering social media marketing/engagement/publishing strategies for clients. A Hybrid social posting/engagement solution made up of using a favorite 3rd party app (hoot, sendible, sprout, argyle, postling, etc) for scheduled posts AND posting directly from within Facebook for quality, thoughtful, engaging posts (using trackable shortened urls) might be a healthy mix for companies with the resources and willingness to do it.

  31. (Perhaps ironically, I’m cross-posting this comment from another blog that linked to this one – so I know one or two others above have made this same argument.)

    This need not necessarily be anything Facebook is doing. Since the stats were measured from how many people interact with or Like the link, it could simply be that (on average) the type of content posted via apps is less likely to be of interest to others than content posted directly.

    This would certainly be in line with my experience – when I see an update that is clearly cross-posted from Twitter or from some app, I am much less likely to find it interesting than one entered by hand. Direct updates are more likely to be things my friend has done or is talking about rather than some business contact or marketing update.

    Even if Facebook doesn’t do anything different, I’d be less likely to click on this content. And if Facebook does apply any automatic algorithm to monitor what we click on, they’d probably pick this up.

    What we would need to resolve this is a controlled experiment with the *same* content posted using two different platforms, and then observations across many accounts to see first if the update shows, then how many people click on it.

  32. Brian

    Hi Chad!

    Thanks for bringing the truth to the light in this study. I think this is something everyone has been thinking but not willing to discuss recently.

    Were any of your studies associated with using Facebooks “secret” email to post to a fanpage? Many of our clients use this option to post while they’re on the road OR even in the middle of a hectic day.

    Thoughts?
    Brian

  33. Chad – any thoughts as to mobile vs. desktop submissions of posts?

    • Chad Wittman

      Unofficially, it looked as though posting mobile was more valuable. Momentus Media did a study looking at it more in depth here: http://www.allfacebook.com/facebook-posting-mistake-2011-09 I’m unsure if the reason behind the increase was due to the quality of the post (for example a behind the scenes photo, live event video, etc).

  34. jda

    This is utter conjecture and you need to test your results to see where you’ve messed up… yes it would be logical for facebook to want people on the site to view ads, but not companies or merchants! Take mrporter as an example, they use conversocial, instagram and others, shows up in news feed everyday… 100′s of likes, shares and comments.

    It’s hardly seen as spam if you’ve liked the brand/company and want to view their updates daily!

    • Chad Wittman

      The data is correct, the only room for speculation is our interpretation of the results. I definitely some Pages have been able to provide exceptional content that seems to reduce any negative effects.

  35. Kris

    We currently have a fanpage with over 420k fans and we have been using a 3rd Party API (hootsuite) and hadn’t experienced any issues until about two days ago when our fan engagement and traffic completely fell off the map. At first we thought maybe it had to do something with the latest facebook updates or maybe the quality of our recent post. Then we had something feed threw we knew for sure would be an engagement and traffic driver and it went flat (only 3 “likes”) so we decided to wait an hour remove the post and re-post it manually straight threw facebook and within minutes of doing so it had dozen of comments and over 100+ “likes” so there is something defiantly going on with using 3rd Party API at the moment at least for us anyways.

    • Chad Wittman

      Really appreciate your input! We’ve been testing this new feed extensively, hope to have more information soon.

      • DeidraG

        This is strictly from a user standpoint, but I’ve been doing a lot of testing on my own page since the last update because I noticed a SUBSTANTIAL decrease in 3rd party posts immediately after the update. I have a lot of friends who post via Twitter, Foursquare, etc., but I rarely see any of those posts. I’ve tried changing all Subscriber settings to see All Updates, played around with Lists … nothing seems to work.

        One theory I’m working on, is that some of these 3rd party apps are getting lumped in with “Comments and likes,” which a lot of users are turning off to get rid of the “Friend X likes a photo on a public page” stories that are now cluttering their News Feeds. (Again, this is just based on my own, limited testing, and because of FB’s additional filtering, it’s hard to tell exactly why something does or doesn’t post. But with “Comments and likes” disabled, I never see a 3rd party post. With it enabled, I might see one occasionally. So it’s something to look into.)

        Also, there used to be a News Feed option that would allow you to turn off FB’s additional filtering, as you know. You could choose between seeing the updates of all your friends and pages or just the ones that you interact with most. Prior to the update, I had opted to see everything. Now that option is gone. So, users have lot a lot of control over what they see.

        Basically what I’m saying is that what I see in my News Feed has changed dramatically since the last update, ESPECIALLY with 3rd party apps. I rarely rarely see them. These are things I’ve opted to see, some of them I even interact with heavily, but I can’t get them to post to my News Feed or lists, no matter what I do. So I’m not surprised at all if a business has noticed a decline in interaction since the last update.

        • Chad Wittman

          Thanks for your insight into the topic, appreciate it. One thing to keep in mind, is that the average user doesn’t touch any custom settings. Which would imply the average user isn’t even customizing things, they’re probably experiencing things as Facebook presents it.

  36. John

    Interesting post. Thanks for doing the research.
    I have been wondering about this exact question for a while. Your research confirms my personal experience using Hootsuite. I rarely get a comment when I use the 3rd party format, just a few ‘likes’. Most of these 3rd party posts are less personal which is probably a factor, but I have noticed when I post business thoughts opr quotes more personally on Fb that I do get more engagement from my friends.
    I guess this means I go back to taking time to post in person, and only use the 3rd party platform when I am very busy and cannot post at all.

  37. Worob

    Does scheduling tweets on a 3rd Party API also decrease engagement per post on Twitter?

    @Worob
    PR at Sunrise

    • Chad Wittman

      We can not determine if the updates are scheduled or not, but I’m willing to bet that this is also a factor. To answer question, yes scheduling updates (according to our data) probably also deals with the same limitations.

  38. Melissa

    Super helpful article. I’ve had this exact hypothesis for months now and watched my impressions shrink drastically from HootSuite posts (especially in the last few weeks). I mean… like 5% of what I’d expect. Pretty severe. I noticed that everything I posted directly to the fan page was getting fantastic impressions. I’m looking into alternative solutions (Post Planner app, anyone?) and would love a follow-up article about which apps are not penalized in edgerank.

  39. Anon

    Just chiming in here. My Facebook page has over 1,000,000 likes. When we post through a 3rd party API, we get between 20,000-50,000 impressions. When we post through Facebook directly, we get 400,000-600,000 impressions. It’s a stark difference. I’m not using 3rd party AP for scheduling, merely to try to get analytics. Obtaining the data definitely isn’t as valuable as the engagement, so I’ll likely be switching back to exclusively Facebook.

    I noticed a similar trend when using a more well-known 3rd Party API with my 500,000 liked page. We’d get about 150,000 impressions with that program, vs. 250,000-300,000 when posting directly to Facebook.

    It’s too bad, our current 3rd Party API is very user friendly and does exactly what we need it to do…except get our content seen.

    • Chad Wittman

      I’ve heard this as the reasoning, although I’m not convinced this is the only reason. This bug applies to the new hybrid news feed, not the old one.

      • Well, they have “fixed” the bug, with a comment that includes some interesting language about aggregation being the issue. I look forward to your follow-up study to see if 3rd Party API updates are still being affected.

        • Chad Wittman

          Thanks for the heads up. Definitely an interesting response. I have a feeling we’ll dive into the changes soon!

  40. John

    How can your analysis influence over 1 billion+ unique fans? There’s only 500 million people on Facebook total.

    • Chad Wittman

      You can’t actually count unique fans across multiple Pages. Of all the Pages we’ve helped, their total fan base is 1B+. This really is the only number we can provide to help illustrate the quality of these Pages. I would love to know the unique fan count!

  41. Max

    Just following up on this thread — should I still be worried about posting from Hootsuite to Facebook? Any updates are appreciated……

    • Chad Wittman

      Always be careful when using anything automated, ever. It appears there is no direct punishment on using the tool. However, this does not mean that there isn’t a correlated decrease with engagement using the tool (which most likely is attributed to using it incorrectly).

  42. jsncruz

    This would make sense. Why would Facebook give third-parties business which they can gain for themselves? Not fair, maybe, but I suppose they have to look out for number one…

  43. I think hootsuite is a great tool and if you make it engaging and relevant your content will float to the top.

  44. Aaron

    Is engaging content like cute cat pictures, aggravating political commentary, etc? All that I seem to see coming through now is memes…

    Some good brand pages have resorted to posting memes constantly because it’s the only way they can be seen.

    It’s sad.

  45. Chris Germano

    Would a third party app posting open graph article actions (ie sharing not social readers) on users profiles positively impact the edgeRank for the domain itself? In other words, are domains of articles associated to objects influenced by edge rank if a third party open graph api is used to post?

    Example:
    100 users click the like button on an article on xyz.com
    10 users use a sharing application (third party with open graph api to post open graph action) on an article on xyz.com

    Does the edge rank for the majority of posts via the like button for articles on that domain have any edge rank boost because other objects are being posted through the open graph api on the same domain?

Leave a Reply

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


six − 5 =

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>