|to view this playlist - see below|
This is Part Two of the UK Police on Social Media Survey for Aug 2014. The focus of this post will be video at YouTube and video editing.
Method Public data has been collected from fifty UK Police official force YouTube Channels.
I've used a second site, vidstatsx for collating the list of most recent videos. This site was useful for delivering the average watch time and view count.
From public data at YouTube, I used a Most Popular view that sorted the videos, for an example see West Mids Police/Most Popular.
As usual, there is a public data set which contains one spreadsheet. When open, the sheet will give
|data sheet A-M|
Note, a Police Force maybe quite unique on YouTube in that it uploads a video, often having thousands of views in a short time - and when circumstances change, that video is often made private.
For example, a public appeal video following a violent assault had 35,000 views in twenty fours hours. The suspect saw himself on YouTube and handed himself in. The video, was made public for legal reasons.
|data sheet M-Z|
With that in mind, all data at any Police Youtube channel is observable, and not what that channel holds privately.
An example maybe a video that has a million views, if that if now private, then I or you have no way to detect that publicly - unless the channel owner says so.
What did we find out? In Part One, I said there were 27,469 youtube subscribers with 25,693,859 total views (cells B56 and C56).
Vidstatsx tells us the average most recent upload is 2:53 - this figure came from visiting each force data set (from 30-50 videos each) and copying that number onto my sheet. The shortest average was 00:33 with the longest being 18:18.
Average Views were 3,219, with the max being +Essex Police with 23,356. This is a very subjective figure for all forces as it' so dependent on how many videos are made and shared.
|video 4 on playlist|
Column F looks at the most viewed video for each force.
It's worth saying that police videos are not a popularity contest, they are there for a good reason - whether that be an urgent appeal, general public awareness or a host of other reasons.
I've recorded the most popular publicly visible video for each force. And then I've made a playlist for the top twelve.
The slide on the right shows daily views - notice how this video from Avon Somerset Police started off slowly, gathering modest views from Aug 2008 to March 2010 when suddenly views started to spike, and then a second spike Feb 2012.
Contrast this with the +hertspolice upload about the 999 Hoax call - show the classic power law curve.
This playlist was created 11 Aug 2014 - this is not the top 12 videos all all time, it's the top 12 across all forces, where each force gets one video vote. For example, Essex Police have four videos above 341k - but they get only one position. And West Midlands have thirteen videos above 112k views.
The number one most watched video is by Essex Police for the Murder of Nahid Almanea, Colchester (17 June 2014) which has been viewed 721,995 times (The population of Colchester is about 104,390 - just shows the wide appeal of one single video).
Here is a playlist of those twelve videos - you will see they vary in length from just 18 seconds to nearly ten minutes.
So what's the ideal Police video look like? I would say about 60 seconds to two minutes and contain many slide changes (or clips).
As an example, take a look at this video from +Greater Manchester Police which is reporting on the Emergency Services Open Day 2014 (held on 7 Aug 14). It's an ideal length of 1:53. What particular good, and why I've chosen it, is the edit and production.
You will see an opening sequence, then an interview to camera (lasting 6 secs), then video of the show (still with voice over). Interview two starts at 00:31 (with the Chief).
There are 113 seconds divided by 31 video events which equals a change on screen every 3.6 seconds (which is good - if you are bored watching TV, count the scene changes on Eastenders etc).
The visuals need to change every 2-3 seconds to keep an audience interested. Feature films however, can change scenes between 10-40 seconds, very long - but when dialogue happens, guess what, it's back to 2-3 secs again.
There are multiple audio tracks on the piece that shows evidence that some NLE has been used. In other words, it was not just a point, shoot and upload.
For me, analysing what happens with the police videos at YouTube will be both challenging and rewarding (if we identify some patterns). I'll do a lot more work on this, this post is far too long - and I've not even started to explain my ideas so far, will wrap it up for today.
With a ton factors in the mix for making videos, like the choice of a pro camera versus a mobile or varying the length at upload will determine the audience retention.And where do those views come from anyway? Could be mainstream press, viral social media or a celebrity tweet.
Thankfully there is room for the unusual, like this one from +West Midlands Police the night the Meteor fireball was spotted.
click through to Part One, UK Police Social Media Aug 2014