Are you a social media/community manager who struggles to determine what a “good” Facebook page metric is?
It seems that the only number people are really looking at is “likes” – as if that means something really important in this day and age of Facebook algorithms, reach and “people talking about this”. It’s essentially like looking at Twitter follower numbers, but not how many replies or RTs an account gets… And we all know likes can be bought.
I’d noticed a trend for a while – that despite Facebook page growing like numbers, the “talking about this” number didn’t really climb, and the percentage of fans talking about the brand dropped!
So I decided to investigate. Here’s my very unscientific method:
I picked a range of pages, mostly from NZ and Australia and a few from the US that:
- had between 1,000 and 500,000 fans
- had posted in the last week
- were ‘official’ pages, not fan pages
and I cross-examined ‘likes’ with ‘people talking about this’. Here’s what I found:
Even as the number of fans grew, it was rare to see the “talking about this” figure go over 15,000. Here’s what the same data looks like when you change the scale:
It’s not that you’d expect a page of 500,000 fans to have 500,000 people talking about them – but why is it so low? Maybe if I had a bigger sample of pages sized 300-500k, I would have had more outliers.
Lets look at ‘talking about this’ as a percentage of fan numbers. I’d expect this to stay reasonably constant, but…
I know it’s difficult to infer from a small sample size, but maybe this is a trend – that when a page gets bigger than about 140,000 fans, engagement slips below 10%. I thought it might have something to do with people not wanting to have their say when a certain number of others are – the old “what’s the point of me saying something when 1,000 other people are”.
Is there are critical mass for Facebook conversations?
- Pages may have been running promotions and ad campaigns which skew the data
- Also note, I looked at many more pages that had <200,000 fans – simply because there is more of them.
- I am not a statistician and have not run any of this via a researcher. That doesn’t mean my data is bad: Just that I’m putting this out there as a thought starter.