Facebook Analytics: PART 2

In the last post I made about the Facebook Insights I laid down a framework of this interesting tool. This post will be about the progress we made with the usage of this tool and the experiments and insights we’ve gotten from it so far. We started our Facebook page in May 2015. Since August 2016 we started posting on a regular basis. Between the 5th and the 26th of September we did an experiment where we used two kinds of parameters to assess our progress. The amount of postings we put out and the type of content these contained. Especially the statistics and progress of these last couple of months will be the object of inquiry, because this is the period we’ve been actively busy with the building of our crowd and her involvement.

In this post we will go over:

  • Our starting point
  • The experimentations fase
  • Latest progress
  • Our best messages
  • Best frequency of posts
  • Who is our audience?
  • When is our audience online?

Within this post you’ll get insight into the steady progress we’ve made so far with looking for the best ways to communicate and build our crowd. We start with the very beginning of our work. As I said we started in August of 2016.

Our starting point

Our starting point with Facebook was around a few hundred people a week who viewed the messages on our Facebook page. We almost got no visitors on our site before we started posting regularly. This was not very strange since we didn’t put out any messages. The page only contained a few things about DELITELABS.

2-8aug385views

In the overview above you see the week 2 till 8 Augustus. You see that the reach of our page (shown in the below left graph) is around 385 and had an increase of 6% that week in comparison to the previous week. This was the week before we started posting more frequently than we had before. You can see an enormous change the week after when we began the week after this one, like you see in the graph below.

8-14aug7-200views

So only posting a couple of post that week made a huge change to the amount of views and involvement we got from our audience. We continued this for a couple of weeks until we wanted to conduct an experiment with this.

Experimentation fase

What we did. We had a couple of weeks where we actually tested with the amount of messages that we posted. To be very precise we did 3 weeks of experimenting with this, from the 5th till the 26th of September. The first two weeks we scheduled two post per day. The main subjects of the post we put out were: Our upcoming course, posts about entrepreneurial subjects and famous entrepreneurs like Ramit Sethi, Gary Vaynerchuck and Simon Sinek. The views we got per message were actually pretty poor, but went way up if it was about applying for the course or pictures of the students that had started on the 19th of September. You see in the next paragraph “Best messages” that two of the most successful messages that we have put out so far were on this very day.

Reach facebook 5 sept - 26 sept

The last week we put out a few messages in a week. These messages were mostly about our course because we had seen that these were the messages that got the most reach and the most involved people behind it. I was curious why there were such deep lows in this graph. Reason: The 16th and 17th nothing was posted and on the 23th, 24th and 25th of September, same story. So one of the things to keep in mind, every day needs at least one post.

We wanted to know what made the most impact. So here is an overview of the things we have been checking to improve our Facebook output.

Latest progress

26sept-23oktfboverview

This is the progress we have made in the last month. From 25 till 22 of October. We see that we have an increase of the pageviews, the involvement and videoviews. Only in the amount of likes we see a small decrease of 17%. According to Jordan, digital marketing expert, involvement is what it all comes down to, so in that respect we are on our way.

Best messages

To know what the best messages were I checked a couple of factors. The parameters I took in account were what kind of post it was(photo, link or video), the time it was posted, the amount of people saw the post, the clicks on the messages, the reactions or shares that these messages had and last but not least the kind of content was provided.

To give you an overview of our best messages in the last period (Sept – Okt)

 

Type of post Date Time Reach Clicks Reaction/share Tags (y/n) Type of content
Link 8 Oct 11:10 1,4k 337 5 n Getting to know our participants with pictures of 4 of them
Photo 11 Oct 12:17 1,3 222 72 n Usual hashtags, @utwente
Photo 5 Oct 13:57 1,1k 190 25 y, 12 students picture of students, hashtags
Video 17 Oct 10:23 869 165 42 Video and hashtags
Picture and link 7 Oct 0:01 1,3k 133 20 First CFpost and hastags
Picture 19 Sep 14:28 1,1k 118 55 y, Violeta and guestlecturer Hashtags
Picture 13 Oct 18:55 879 103 3 n hashtag and pitching @utwente
Picture and link 10 Sep 12:20 1,9k 100 72 n Apply course
Video 19 Sep 12:15 958 90 30 n hashtags
Picture and link 7 Oct 0:01 1,3k 90 2 n post CF 25%

Analysis done 2016

What’s interesting in this table is that most of the messages are pictures of video. But almost all the messages have in common is that our students are in the picture. Not our company, not posts of famous entrepreneurs we also posted, but our own students. The mission we have to help these people. The advice that for example Jordan gave us in an interview I did, was very accurate to the findings we have here. The advice was that we should make our posts as personal as could be and we did. Another advice we only used a couple of times paid off as well, the tagging of other people in the post we made. You can see this in the photo we posted in the 5th of October, third best post on our Facebook page so far.

Best frequency of posts

We have began frequently posting from August 2016 onward. But we waited till the beginning of September to devise a more structured system to it and set up an experiment. In the three weeks from September 5th till the 26th of September we did our experiment. In the first 2 weeks of the experiment we scheduled Facebook messages twice a day. From the messages that we put out in the first two weeks we could conclude that 2 a day was a bit over the top, especially because in the third week, when we put out 5 messages in stead of 14 we got more reach than the previous two weeks.

We had seen that the amount of posting was incredibly important, too many posts was a turn off for our audience. What also very much mattered were the kind of posts we put out. Especially the third week showed a big difference in the amount of viewers.

Who is our audience?

With these statistics we can start to describe our target user. For example the following image will show what percentage of a certain age group has the most interaction with our Facebook page. Additionally the image shows what percentage of age groups are most involved.

screen-shot-2016-10-24-at-00-28-11

The highest amount of people that visit or interact with our Facebook page are people between the age of 25 and 34. The younger or the older people in comparison to this group, the less likely it will be that they will interact with our page. The ratio between men and women is pretty similar in the clicks they give. You see that the percentage of involved people (people who react or share a message) is higher with the men.

Another interesting statistic that you can see in Facebook Insights is the geographical information of our target user. We see, if we look at the places where most people live that this is actually where we were stationed. Amsterdam and Valencia.

targetusergeographics

 

What time is our target group online?

Facebook Insights has a very handy tool to find an answer to this question. When you click on the button messages when you are in the statistics screen of your Facebook page you can have an overview of the amount of people per hour are online.

OktoberOnlineMensen op facebook

This is the graph I was talking about. The thing that immediately draws the attention is the large gap between midnight and 8am. After 10pm the amount of fans that are online drops. Furthermore you can see that on the rest of the graph that its very slightly rising. You can also choose a day to see if the pattern changes during the days. We have found slight changes in behavior. For example, On Mondays and Sundays our fans approximately are an hour later online in the morning than the rest of the days. This difference is shown in the graphs below.

Monday

screen-shot-2016-10-24-at-00-09-53

Tuesday

screen-shot-2016-10-24-at-00-13-36

 

Conclusion

A couple of conclusions I can draw out of this. First is that, because of our effort we set up a Facebook page that has a couple of thousand views, and growing, in stead of a couple of hundred.

Second, from the experiment we did we can now say that people get oversaturated when you post more than two or more times a day

Thirdly, personal messages, messages about Delitelabs and their students get a much higher involvement and view rate than other postings. This is what you can see in the “best messages” paragraph. Another, rather obvious finding was that video and pictures work better than just plain text

Fourthly, our audience consists of almost a fifty/fifty division of male and female, although we can see that men are a little more involved than the women. Most of them are between the age of 25 and 34. The main languages are English, Spanish and Arabic and most of the people that visit our site are from Amsterdam and Valencia. There is also an interesting pattern found in the time that our target group is online. Roughly sketched they are online from 9:00 to 22:00, with a small variety on Mondays and Sundays.

 

Leave a Reply

Your email address will not be published. Required fields are marked *