September 5, 2013 by Benjamin Rey
On the relationship between Facebook organic posts and promoted posts
Promoted posts, it started with a suspicion
Since May 2012, Facebook lets users promote posts. At the time there was a lot of suspicion about Facebook reducing brands’ reach to force them into promoting their posts. A few studies, including ours pointed out that even if some brands had indeed been penalized, it was in most cases for the greater good. Users see less spammy posts. Now that things have settled down on the organic front, there are debates about what happens when a post is promoted:
Do promoted posts cannibalize organic results? Just like you would fear on Google. In other words, if you pay for search ads, will you still get as many clicks on you organic (free) results? Google says you will
How does a promoted post impact future posts? Some suspect, for instance, that all the extra negative feedback has a really bad impact on the page.
From our data’s perspective
Since we optimize brands’ social performance through both organic and paid (ask for a Nitro beta invite at http://www.wisemetrics.com), we had to look into these questions in detail. Just like in our previous post on Facebook posts’ lifetime, we thought we would look at this from a data perspective. We randomly selected 5K promoted posts from 1.5K pages. It is a sample, as possibly random as we can deliver. It will be missing some special cases, but it does give a good general idea of Facebook pages behavior regarding to promoted posts.
No it won’t cannibalize your organic reach
Facebook provides us with detailed statistics for promoted posts, differentiating paid impressions from organic ones: if you pay for 10,000 impressions, your post might get 11,327 impressions, 1,327 of them being organic. Are these 1,327 in line with the usual impressions you get for an un-promoted post? This is what we looked at. From our 5K sample, for each promoted post, we compared the organic part of its reach to the organic reach from the 10 un-promoted previous posts. On the graph below, we present organic impressions for promoted vs organic impressions for usual un-promoted posts, with an area covering 90% of the 5K data points (posts).
The result is clear: promoting a post does not cannibalize organic reach at all.
Indeed, median relationship between promoted and un-promoted is really close to the y=x line, meaning that for 1,000 organic impressions when not paying, we get 1,000 organic impressions when paying. If anything, small promoted posts (a few hundred impressions) are getting a better share of organic impressions than they would, would they not have been promoted.
And it has no impact your Edgerank either
EdgeRank is dead, long live EdgeRank! Even if it’s not named EdgeRank anymore, the 100K knobs model has still been built by Facebook engineers, and at some point, they did decide to filter out effects from promoted posts… or not.
Neither for the best
This time for each promoted post, we compared impressions from posts immediately before to those immediately after. And here again impressions before and after promoted posts are at the same level. All the extra buzz one gets from a promoted post doesn’t translate into improved visibility for future posts.
Nor for the worst
We also looked at the impact of negative feedback from a too wide promotion: does accumulated negative feedback impacts future posts?
Here is the plot of impressions before and after the 5% spammiest promoted posts (posts with the highest negative feedback ratio).
It looks like the negative feedback from promoted posts doesn’t impair future unpromoted posts.
Ok, maybe it’s not so great for your brand to be spammy, and it is more than advisable to closely look at negative feedback metric (check out our Wisemetrics Stream tool wisemetrics.com/stream for this). But at least it doesn’t directly impact your Edgerank.
It turns out Facebook does filter out all the extra (promoted) impressions, clicks, engagement and negative feedback from “Edgerank” modeling.
Brands shouldn’t be afraid of promoting posts, and apparently, there’s no need to wait for a post to wear down. Promoting it early won’t cannibalize organic results. If promoting a post won’t help you increase your future audience, at least it doesn’t hurt.
What’s your feeling about these findings?
If you have any questions or suggestions for things we should dig into, drop us a line, we’ll be happy to take a look and share the results.
bitly for this post: http://bit.ly/1eawM8F