Can you trust Facebook stats?

I'm not sure if this is a stats problem, an issue with how Facebook shows shared posts, or whether I'm reading this all wrong to begin with. Any thoughts would be appreciated!

We recently had a very popular post on Facebook about smoke damage , and as you can see below, it was shared 36 times. What is confusing/frustrating me, however, is that the "viral" reach was only 481, which, if I'm interpreting this correctly, indicates that each share only reached an average of just over 13 people. Now, I know there are a lot of very inefficient fan pages, but to average just 13 reach per post seems awfully low.



Thoughts anybody?

Some more validation from a different post : In this example, the Engaged users, which is supposed to show unique users who clicked on the post, is somehow greater than the reach, which is unique people who saw the post. In other words, according to these stats, some people managed to click the post without ever actually seeing it



Now, for this particular post, we saw dramatic increase in reach when looking at the "Weekly Total Reach", which means that the weekly total reach probably is calculating reach correctly by including reach from shared posts, while stats for individual posts are not showing that reach properly.

Unless I am waaaaaay off base here, it is amazing that Facebook is so inconsistent when reporting these same numbers in different areas.

Engaged users are those that have clicked anywhere on your post. Now, assuming that all 36 shares had someone click on the post once it was posted, then you have roughly 1.5 more clicks to your post per share. There's your variance of 53 people. (524 Engaged Less Reach of 471) - Assuming that NONE of those clicks came from people that like your page and initially saw the post from the pages actual feed, that is in fact how your engaged # is larger than your reach. Your post counts as having reached someone when it is loaded and shown in their news feed - So I am going to assume here that if I shared it and my friends uniquely saw it because of my share, it wouldn't count towards the calculations on your page because I was the one that introduced it to their feed - however if they clicked it, it would count towards your engaged number, because as we know, clicking on a shared photo takes you back to the original source. (side note, one of my fan pages shared the couch photo - here are the insights on it - It brought you a whopping 4 clicks)


Virality is the number of people who have created a story from your post as a percentage of the number of people who have seen it. According to your organic number (which btw is the number of unique people, fans or non-fans, who saw this post in their news feed, ticker or on your Page) 861 people initially SAW the post, of those 36 shared. Facebook is figuring your viral number of 481 from the number of unique people who saw this post from a story published by a friend. You are correct in that there is a reach of roughly 13 people per share. BUT what Facebook doesn't include here (and can greatly affect this number) is the type of audience it was shared to -

If you have other fan pages sharing the image, perhaps they don't have a high level of engagement. We all also know that fan pages that aren't extremely consistent in posting, that don't have a large following often get filtered out of news feeds. Facebook is only sharing with fans who repeatedly return to the page, post on the page, comment on the page, or otherwise engage the page.

So while the average may be 13, there may be some shares that didn't get viewed at all, while others were seen 20-30 times, if not more. Most of my (in my opinion) under performing FB pages average about 25 views per post. I have one page that gets NO views - (and that's another discussion post entirely). I have another that averages 300-400 views per post. I think that this factor could have more weight on your reach than you realize. Also, pages with less than 30 likes aren't privy to insights. That could be weighing your reach down too.

Before stressing yourself entirely out, be sure to take that into account too.

Also, for a personal profile share - I heavily use lists and control who can see my posts because of the nature of my job. I have work lists, close friends lists, school lists, a standard friends list, etc. Typically when I post something via a list, it doesn't get a lot of traction in anyone's news feed until someone has liked or commented on the post itself. (Someone is the first to comment on something I posted 3 days ago and bam I am flooded with notifications from others within an hour about said post) Furthermore, I am noticing that based on my usage of the lists, I miss a lot of shares/updates from friends until I actually go to the designated news feed for that list. And there are certain people who's status updates I have set to "only important" or "most updates", particularly if they are guilty of regularly sharing things that have the "eww" factor. Otherwise, my standard news feed has about half of the updates from those people and pages I engage regularly.

So, my dear Brent, while I certainly love insights and numbers, and find a lot of value in them, sometimes I know when to take a "win" and run with it. You've had a great couple of posts that got a good number of shares, comments and views. I do expect the smoke image it to get a few more over the next day or two btw - especially now that it's a discussion topic. SO, let's chalk this up to a quick lesson in insights but a bigger lesson on the importance of consistency and quality posting.

Thanks for the great reply, Kristi! I think was definitely underestimating the possibility that the fan pages could have an average of 13 reach for a post. Just seemed too low to be credible.

As for the other issue about reach versus engaged user, I see what you are saying, but in my mind, that way of accounting by Facebook doesn't make any sense. If you are correct, then Facebook only wants to count reach one time. So since it was seen through a shared post from another Page, it would be counted on that Page but not on mine. That's not a problem EXCEPT that's not how it is treating Engaged Users. If Facebook wanted to be consistent, it would have to treat that number in the same way, but it clearly doesn't. So what we are left with is a wonky situation where Engaged Users is greater than Reach, which logically should never happen.

(I'm glad you are here to have this purely geeky conversation with me!)

I should really stop looking for these. Here is an example of a Facebook Question, where the Engaged Users should be at least the number of people who answered the question. So in this example it should be at least 88, but it is only showing 50.

My biggest concern out of all of this is not the data itself, but rather Facebook is using this data to sell advertising. If you can't have faith in how they are determining these numbers, how can I buy advertising based upon any of them?

I agree that a lot of it seems completely illogical and is a baffling mess of "what does it all really mean?"

I am with you on the asked question post - how is it that more people answered the question than engaged it? I took a look at one of my pages that asks questions regularly too and see the same exact thing with there being more votes than engaged users on every single question they've asked. The only logical explanation I can come to is that those that are answering are counted in some other metric (click the number of people engaged, then click "other clicks" to see this definition) -I have no idea what that metric is, other than it seems as though maybe FB is decreasing some numbers on the engaged end to make the viral number larger? Which makes me say .. "but why?" and opens up a long line of questions about the metrics themselves and how much stock we should even place in them.

And I am 100% opposed to paying for advertising from Facebook presently. For one, I think it's going to take some time to decide if it does or doesn't work. Why not let the bigger corporations spend their money to test the waters first? For two, the insights (as our conversation is proving) are very difficult to decipher sometimes. And for three - there are brands that I engage with regularly because I believe in the brand itself, the service it provides or I find their content relevant and interesting. I personally (social media marketing aside) am much more hesitant to engage a "promoted" story. I am leery of why someone would pay to promote the message that they are; And I think that a lot of other users may be this way too. Privacy has always been a big concern and I think that when "promoted" appears under a story it is currently the equivalent to a red flag. PS - It also irritates me a little to see promoted trends on Twitter too.

I think it comes down to either the stats are simply wrong, or the layout of the stats is so convoluted that you would have to be the designer of the system him/herself just to accurately decipher it.

And with that, I'm done analyzing this stuff - I have enough evidence to realize that insights are probably good to get a general feel for how the page is doing, but any in-depth analysis is pretty worthless.