Audience Retention??

The Bassment

The Bassment

Posting Mad!
How is the % of overall audience retention calculated?

If I had two videos, one 10 mins long, the other 2 mins long, and average audience retention of 1 min, would my average audience retention be based on a simple average video length of 6 mins and therefore be only 17%? If this is the case, what if virtually all the views were on the 2 min video, meaning that audience retention is actually more around the 50% mark?

The only workaround I can see for this is to calculate an average video length using the weighted average length based on views, but do YouTube go to that much trouble?

Sorry if this is completely incomprehensible, many thanks to anyone taking the time to work out what the hell I'm on about ;)
 
I think you have to find the percentage per video, and then average the percentages. Although it's tedious, that will get the best results. It can also average the Total video time, and the viewer time watched and find that percent.

So it needs to be : Viewer Time Watched / Total video x 100% = percent

So lets say for one video is 10 minutes and they watch for 5 minutes
and one is 2 minutes and they watch for 1 minute it is easy to see you SHOULD have a 50% retention percentage.

5/10 = .5 x 100 = 50%
1/2 = .5 x 100 = 50%

Average = 50 + 50 / 2 = 50% average

So the second method:

10 + 2 = 12 / 2 = 6 minutes total

1 + 5 = 6 / 2 = 3 viewer watched avg.

3 / 6 = .5 x 100 = 50%


Now, as per your example, you are right about the 17%. Why? Think about it, you have MORE video, but LESS is being watched, therefore you have low. Now let's say 17% of your video is being watched on average, dropping all of your videos down to 2 minutes will not help this due to the fact that your average will drop because there is overall LESS video.

10 x .17 = 1.7 minutes being watched.
2 x .17 = .34 minutes being watched. Average them to get about 2 (rounding errors)

Why is this significant? In your ten minute video, you had more video, therefore meaning that more was watched to bring the average higher, but in the two minute video; only .34 minutes were being watched, so if you had all two minute videos

.34 +.34 /2 = .34 average viewer time / average time = 2 + 2 /2 = 2

.34 / 2 x 100% = 17% of your video being watched, which is kept constant.

In conclusion:

Length manipulation will not combat your retention percentages, only a change in content will. There has to be at least some reason why people decide to leave, so focus on that, rather than wanting to get a stat at a pretty number. If this was purely for knowledge, and you didn't want to manipulate length for this number; then nevermind.

If you have any further questions about what I did or why I did it, don't hesitate to ask :D
 
Low audience retention has got to be one of those things I dread to read on my channel.
 
JWB said:
I think you have to find the percentage per video, and then average the percentages. Although it's tedious, that will get the best results. It can also average the Total video time, and the viewer time watched and find that percent.

So it needs to be : Viewer Time Watched / Total video x 100% = percent

So lets say for one video is 10 minutes and they watch for 5 minutes
and one is 2 minutes and they watch for 1 minute it is easy to see you SHOULD have a 50% retention percentage.

5/10 = .5 x 100 = 50%
1/2 = .5 x 100 = 50%

Average = 50 + 50 / 2 = 50% average

So the second method:

10 + 2 = 12 / 2 = 6 minutes total

1 + 5 = 6 / 2 = 3 viewer watched avg.

3 / 6 = .5 x 100 = 50%


Now, as per your example, you are right about the 17%. Why? Think about it, you have MORE video, but LESS is being watched, therefore you have low. Now let's say 17% of your video is being watched on average, dropping all of your videos down to 2 minutes will not help this due to the fact that your average will drop because there is overall LESS video.

10 x .17 = 1.7 minutes being watched.
2 x .17 = .34 minutes being watched. Average them to get about 2 (rounding errors)

Why is this significant? In your ten minute video, you had more video, therefore meaning that more was watched to bring the average higher, but in the two minute video; only .34 minutes were being watched, so if you had all two minute videos

.34 +.34 /2 = .34 average viewer time / average time = 2 + 2 /2 = 2

.34 / 2 x 100% = 17% of your video being watched, which is kept constant.

In conclusion:

Length manipulation will not combat your retention percentages, only a change in content will. There has to be at least some reason why people decide to leave, so focus on that, rather than wanting to get a stat at a pretty number. If this was purely for knowledge, and you didn't want to manipulate length for this number; then nevermind.

If you have any further questions about what I did or why I did it, don't hesitate to ask :D
Click to expand...

Thanks for the detailed response. I'm not trying to manipulate the number, purely to understand the percentage calculated in my analytics.

Using the example above, my issue is that if I had a average view time of 1 min, and an average percentage of 17%, I have no idea whether I should be worried because the percentage is low, or whether it's just an issue with the calculation method. In that actually all my views are on the shorter 2 minute video and really retention is more like 50%, it just looks lower purely because I have uploaded a ten minute video, even though it hasn't actually got any views, which increases the average video length upon which the retention is calculated.

To illustrate, say at day 1 I have a single 2 minute video with 100 views and an average view time of 1 minute:

1min/2min = 50% retention

Now on day 2 I add a second video that is ten mins long, but it gets no views at all. So I still only have 100 views of one minute average length on my first video. Will my analytics now read:

(2min+10min)/2 = Average video length 6 min

1 min/6min = 17% retention

Or is YouTube cleverer than that and does a weighted average calculation like:

(2min x 100 views +10 min x 0 views) / 100 = average viewed video length 2 mins

1min/2min = 50%[DOUBLEPOST=1370775509][/DOUBLEPOST]Aha, worry not, just remembered I have an old channel with only 3 videos so I did a quick check and it is a weighted average that is used. This makes the retention percentage far more accurate on channels like mine with wildly different video lengths so I am very pleased :D (also my old channel has 103% retention, 9 second clips ftw!)
 
You must log in or register to reply here.
Back
Top