KPI explanation – Facebook

Data provided by Facebook and Fanpage Karma.

Growth Rate

The growth rate tells your the average weekly growth of a Facebook page. It does this by answering the question “How much had the page to grow in a week on average to get from value 1 to value 2?”.

That makes sense, because it gives you a more “stable” impression about your growth without being too vulnerable to sudden spikes or extreme differences.

Example:

Page A has 2000 fans, when two weeks ago it only had 500. Just the difference is 1500 fans or 300% of the starting value of 500. The calculation of the average weekly growth rate “only” equals to 100%:

• The page grew by 100% in the first week, which is 500 fans: 500 fans + 500 fans = 1000 fans.
• For the second week it’s another 100%: 1000 fans + 1000 fans = 2000 fans.

Engagement

The Engagement shows an average amount of how often a fan interacts with the posts of a page. It is calculated by dividing the daily amount of likes, comments and shares by the number of fans. If you are looking at a longer period of time it takes the average of the daily Engagement rates.

Example:

Page A yields the following interaction values:

• Monday: 20 Interactions and 3500 Fans
• Tuesday: 0 Interactions and 3590 Fans
• Wednesday: 37 Interactions and 3700 Fans
• Thursday: 100 Interactions and 3750 Fans
• Friday: 0 Interactions and 3755 Fans
• Saturday: 0 Interactions and 3780 Fans
• Sunday: 2 Interactions and 3783 Fans

That leads to the following daily Engagement Rates:

• Monday: 0.0057 (20/3500) or 0.57%
• Tuesday: 0 (0/3590)
• Wednesday: 0.01 (37/3700) or 1%
• Thursday: 0.027 (100/3750) or 2.7%
• Friday: 0
• Saturday: 0
• Sunday: 0.00053 (2/3783) or 0.053%

For the whole week, we get the average Engagement value of:

(0.0057+0+0.01+0.027+0+0+0.00053)/7 = 0.0062 or 0.62%

That means, in this week each fan interacted 0.0062 times with the posts of the page.

Post Interaction

The post interaction shows how active the fans engage with a post of a page. It shows the average amount of all interactions (likes, shares, comments) for each fan per post. As opposed to the Engagement Rate, the Post Interaction ignores days without any posts.

Example:

Page A yields the following interaction values:

• Monday: 20 Interactions and 3500 Fans and 2 Posts
• Tuesday: 0 Interactions and 3590 Fans and 0 Posts
• Wednesday: 37 Interactions and 3700 Fans and 1 Post
• Thursday: 100 Interactions and 3750 Fans and 3 Posts
• Friday: 0 Interactions and 3755 Fans and 1 Post
• Saturday: 0 Interactions and 3780 Fans and 0 Posts
• Sunday: 2 Interactions and 3783 Fans and 4 Posts

Now we sum up all interactions by fans for every post and divide it by the amount of posts.

(  (20/3500) +  (37/3700) + (100/3750) +  (2/3783)  ) / (2+1+3+1+4)
= 4.29% Total Interaction Rate / 10 Posts
= 0.429%

That means, in this week each fan interacted 0.00429  times with each post of the page.

Service Level

The service level shows how many userposts get a reaction of the page in form of a like, a comment or deletion. Reactions below “normal posts” (published by the page itself) do not count.

Example:

Page A got 20 userposts within the chosen period of time. 4 of them got a like, 5 a comment and 2 were deleted; 11 userposts got a reaction, which equals 55% (20*55/100=11).

Response Time

The response time describes how long a page needs to respond to userposts. For this value, only taken into account when a page makes a comment to a userpost. This is mainly because Facebook does not provide time values for likes and shares.

We first calculate the response time for each userpost and then take the median of all posts for the selected period of time. The median is the value, that is exactly in the middle of all values, when you sort them.

Example:

Page A from the previous example reacts within the following timeframes:

• Userpost 1: commented by page  after 10 minutes (0.17h)
• Userpost 2: commented by page after 45 minutes (0.75h)
• Userpost 3: commented by page after 2 hours (2h)
• Userpost 4: commented by page after 20 hours (20h)
• Userpost 5: commented by page after 53 minutes (0.88h)

If you sort all durations you get the following list:

• 0.17h
• 0.75h
• 0.88h
• 2.00h
• 20.00h

The value in the middle is 0.88h, so the Response Time is 0.88 hours.

Why don’t we take the average of all values? The average has two problems: First of all, the average is very prone to overevaluate outliers. In the example, one post was responded to only after 20 hours. If you calculated the average of the values, you would get to 4.76 hours. This is much higher than you would think fair if you looked at the other posts. Secondly, these 4.76 hours are a completely fictive value, whereas the 0.88 hours of the median can be clearly and correctly explained to any interested party (your boss or client) as: “Half of all userposts are responded to in less than 0.88 hours.”

Ad Value

The ad value is calculated by taking post reach and an average price for online advertisement into account (cpm). It shows how much you had to spend if you would have wanted to reach as many people with common online ads as you did on Facebook. An estimated cpm of $8.50 is used and multiply it with your reach.

PPI

The Page Performance Index (PPI) is a combination of engagement and growth. It combines both figures to provide an estimate value for a pages success and is based on the average growth and engagement values of all pages in our index. The PPI will be set to 100%, if a page is part of the top 10%, so 90% of the pages which are monitored have less engagement and growth. Afterwards the engagement and growth are multiplied, the square root is extracted and the values scaled to 100 to present the top end.

Example:

Page A has an average Growth rate of 3% and an Engagement rate of 0.2%. Both values are better than 90% of all pages in the index, which produces “temporary” values of 100% Growth and 100% Engagement. These temporary stand-in values multiplied equals 10,000 (100*100), the square root of 10,000 again is 100, which results in 100% PPI.