Understanding Facebook Insights Terminology Redux

By on in with 9 Comments

When the latest Facebook Insights was released, I quickly put up a post that both tried to explain the new metrics that became available and proposed some probable KPIs.

Well, a few months have passed, Facebook has quietly rolled out some changes to Facebook Insights, and we’ve gotten a chance to actually dive into some of these metrics. This post and the next two are the result of some digging that Mike Amer and I have done on behalf of Resource Interactive and our clients.

Note: This is minimally a post about the web-based Facebook Insights interface. Rather, it is focused on the slightly deeper data that is available behind that interface, which is available by exporting page-level and post-level data or through the Facebook API.

Understanding the Basics – Reach, Talking About, Engaged Users, Consumers

I get a little depressed when I think about the number of times I have read and re-read the same one-line Facebook Insights definitions for various metrics, which have the illusion of being crystal clear on an initial reading, and then get increasingly confusing with each subsequent cycle of trying to actually interpret the data.

I continue to think that the best way to understand the main new metrics is via a Venn diagram. But, the page-level Venn diagram has evolved a bit since my initial post, as Facebook quietly added a page-level Engaged Users metric, which is the union of People Talking About and Consumers. I also think that Facebook changed the definition of Consumers to include clicks that generated a story, but I haven’t tracked down old printouts to fully confirm.

Below is an updated Venn diagram for page-level Facebook metrics.

And, below is an (unchanged) Venn diagram for post-level metrics:

What About Paid/Organic/Viral (especially Viral!)?

At both the page level and the post level, Facebook breaks down a number of metrics by “paid,” “organic,” and “viral” Here’s how I’ve been describing these when it comes to page-level reach:

  • Organic – unique people who visited the page or saw an item published by the page in their news feed or ticker
  • Viral – unique people who were exposed to content as a result of another user generating a story (“talking about” the page – liking the page, sharing a post, etc.)
  • Paid – unique people who saw a Sponsored Story or Ad pointing to the page

A single user can be reached by multiple ways in a given time period (e.g., they saw a post from the page that they’re a fan of in their news feed – organic – and then saw that a friend of theirs responded to a question on the page in their ticker – viral – and then was exposed to a Facebook Ad – paid), so, when it comes to reach, the sum of organic reach plus viral reach plus paid reach is greater than the total reach. Reach measures are always de-duped to be a count of unique users.

When it comes to impressions, though, there is no de-duping, so the sum of the different types of impressions equals the total impressions.

In my next post, I’ll dig into “virality” a little deeper (it turns out to be a bit of a bugaboo metric, but it’s also one that turns out to reveal some sneaky little unpleasantries about Facebook’s EdgeRank algorithm).

9 Comments


  1. Pingback Anonymous

  2. I have had a couple of questions about FB Insights that I have searched for weeks online and have not found the answers. Even talking with FB reps didnt help (go figure). They all have to do with the Page Level stats and I even included the worksheet names for your convenience.

    1) Key Metrics tab:
    a. Why is there a count delta between my Daily Organic Reach and Daily Reach of Page Posts? What is not included in Daily Organic Reach this included in Daily Reach of Page Posts?
    b. Can you reach the same person Organically and Virally? Why does Organic Reach + Viral Reach equal to more than Total Reach (and Paid Reach is 0)?

    2) ‘Daily Likes Sources’ tab:
    a. What are the full definitions of each of the sources: profile_connect, mobile, api, recommended_pages, page_suggestions, timeline, external_connect, page_profile, hovercard, search, ticker, like_story
    b. Are mobile likes independent of the others?
    c. Why would the Daily New Likes column in the Key Metrics tab not equal the summation of all the columns for the same day in the Daily Likes Sources tab?

    3) ‘Daily Viral Reach by Story Type’ tab:
    a. What are the full definitions of each of the story types: fan, page post, user post, mention?
    b. If we normally get 1-5 viral uniques from user post, and then one day get 1.5k, what is the likely source of this?

    4) ‘Daily Page Consumers by Consumption Type’ tab:
    a. What are the full definitions of the consumption types: other click, link click
    b. Are photo views and video views included in other clicks?

  3. Good questions, Matt. And, my answers are based on what I’ve been able to observe and infer (with similar lack of luck when it comes to Facebook reps’ ability to answer questions or dig up any more than the base documentation). Here we go:

    1a) If someone simply searches for and visits your page, they’ve been reached organically without being reached by a page post, I suspect. It seems like there should be other “non-post” ways that users can be reached organically…but none are coming to mind. Or…wait…are you seeing Organic Reach of Page Posts being *higher* than Organic Reach of Page? That seems wrong.

    1b) Absolutely the same person can be reached both organically and virally. See the example immediately under the bulleted list towards the end of this post.

    2a) I don’t have full definitions. Most of them I can guess at, but documentation is nonexistent. I do think it seems that “profile_connect” and “page_profile” are the same thing (somebody visits the page and uses the Like button on the page)…but one was used before a date a month or so ago and one was after. Let me know if there’s a specific one or two that you’re wondering about (maybe I’ll do a post where I postulate full definitions of each).

    2b) I don’t know. I would think so (and wonder if those are limited to Likes from within one of Facebook’s mobile apps, maybe?)

    2c) Are you seeing Daily New Likes as being greater than the sum of the parts? If so, I’d guess that the discrepancy is people who like and unlike in the same day and, thus, don’t show up in the Like Sources but do show up as a Daily New Like (and also a Daily Unlike) for the same day.

    3a) I puzzled over that one, and what I *think* is:

    – fan: the number of people reached because someone Liked the page

    – page post: the number of people who were reached because of a story generated when a user reacted to a post (commented, shared, responded to a question, etc.)

    – user post: the number of people who were reached as the result of someone simply posting on the page’s wall or responding to a user’s post on the wall

    – mention: the number of people reached because someone tagged the page in a photo, video, or status update.

    3b) Did you have a celebrity of some sort post on your page? That seems like quite the anomaly and would take some digging.

    4a) “Other clicks” are clicks that generated a story, such as liking a post, liking a page (from the page only — not liking it from an Ad, for instance; super ambiguity here, as consumptions are limited to “page content,” which is ill-defined), etc. The other ones seem fairly self-explanatory…except when it comes to video views.

    4b) Photo views and video views should not be included in other clicks. However, what’s stumped us on this (although we haven’t worked too hard to reverse-engineer it using YouTube Analytics data), is whether a YouTube video that gets posted, when clicked, gets counted as a link click or a video view (in other words, are “video views” limited to “videos that are uploaded directly into Facebook”).

    I hope that answers at least some of your questions. It would be fantastic if Facebook moderated a forum where these sorts of things could be asked and worked out.

  4. Tim,

    Thanks for the quick response. I found the info to be incredibly useful and appreciate the effort you put into it. I too look forward to the day that Facebook moderates an analytics forum, not just a dev forum for tech questions. 🙂

    Regards,
    Matt

  5. >> It would be fantastic if Facebook moderated a forum where these sorts of things could be asked and worked out.

    The two of you are being incredibly forgiving of Facebook, far more than I am feeling!

    Facebook shouldn’t be getting these questions, not a single one of them. Facebook should have already published detailed, explicit and clear definitions of every single metric. Just like the real analytics companies do!

    The bogus pseudo-definitions and generalities that Facebook has supplied are disgraceful. So is the utterly incompetent drivel that comes from Facebook’s so-called sales support people.

    #best-rant-youll-hear-today 🙂
    #expertly-crafted-to-get-on-FB’s-social-sentiment-monitor

  6. hoping these comments become the forum where we all discuss it (assuming FB doesn’t do the right thing and simply clarify).

  7. Hi Gilligan,

    thank you for this lovely post. do you happen to know if there is now any documentation, forum or any kind of online help on facebook insights?

    thank you
    artemis

  8. Pingback Facebook EdgeRank – How to increase Affinity « British Schnitzel Reloaded

  9. Unfortunately, the only guidance I’ve gotten from Facebook is to post specific questions on their (closed) Professional Marketing Developers Facebook group, and the times that I’ve resorted to that have been pretty unsatisfactory. If someone starts one up, I’ll promote it and participate! If I find some free time, maybe I’ll think through options for what that could be.

Leave your Comment


« »