I'm going to give this an introduction so it all makes sense, before proving popularity scores. So please excuse the length of the opening.

I've been running an experiment this year using a social media analytics platform which can gauge the popularity of brands, it's a platform which keeps checks on Facebook, Twitter, G+, Vimeo (irrelevant here), YouTube and Instagram and tracks performance of a brands high level and granular metrics like all analytics platforms and has an insight tool that tracks realtime interaction metrics independently of the overall social profile metrics so you can see performance without historical pollution.

A kind of clever metric it provides is a popularity metric, which is a day by day score that measures user interaction (discussion, likes, retweets, follows etc.) and weighs it through a gamification engine, giving higher precedence to consistent interaction from an audience that is invested than a flippant interaction. It provides this because a follower can be pretty meaningless on it's own as following someone is often done on a whim without any emotional investment from the person doing it. It's kind of like the Klout score but for popularity.

As an experiment I plugged the XF contestants in to see whether there is any relationship between social media popularity and the result.

So far the results from the platform have been shared privately and in an AP thread. After a few requests and a few wins I've decided to bring it to the main forum each weekend (possibly at my own peril).

Using the real time interaction and popularity scores, weeks 1 and 2 were pretty much written off as the lower end of the results were so close together. Week 3 onwards a spread started to really show up so that's where I am counting from.

Week 3 it got Hannah
Week 4 it got Tamera vs Kingsland Road
Week 5 it got Hannah vs Abi

It's also shown quite a lot of interesting things.

Nic, Sam C and Sam B leading at the top over weekends (Sam C and Nic strongest through the week... Nic by a ridiculous amount),

Seen Luke gain popularity rapidly since last weekend

Seen Abi race up the leaderboard on her good week (4th from top) before bombing out last week

Seen Tamera and Hannah appear to pendulum, Hannah's B2 saw Tamera lose a huge chunk of her audience the following weekend, to Hannah.

Shown through it's influencer identification platform that Sam B and Nic have a broad audience who include younger folk so their audiences aren't just older people.


So I thought I'd share it all and share the scores periodically. I am not going to say that it will get the bottom 2 right every week, nor that the results it gives will match voting results, the point is to find out whether there is any relationship between social media popularity and results and how frequently or accurately it can highlight contestant positions.


There appears to be a pattern each contestant has, they have a baseline which is where they sit during the week, a peak over the weekend, their biggest growth is on the Sunday and they move around a lot and settle into place on the Sunday afternoon.

Also Luke appears to get the biggest boost in score on a Sunday afternoon, moreso than any of the others.


Anyhow, here are the current scores from the real time insights:

Hannah: 6,415
Tamera: 13,520
Rough Copy: 20,399
Sam B: 40,557
Sam C: 92,814
Luke: 163,923
Nic: 189,708


These will boost a lot today and will reach much higher peaks tomorrow, well unless anyone crashes and burns. They will also most likely change but so far the change pre show to post show has given some idea of what they may look like the following day and the pre show scores give their current buzz (which is essentially what this score tells us).

Hi i'm glad you posted a thread, looking forward to the scores/prediction throughout the weekend

very interesting. thx!!

really interesting,I shall be watching to see how it goes.

Fascinating, Captain.

It seems the boys have the highest scores, does that tell us something about the demographics of those tweeting?

We'll see how it goes. I think tbh, whether it gets it right or not this weekend will probably be dwarfed by the other thing it shows.

* grumbles... Nic stupid score *

Oh and if it's wrong I am holding you personally responsible as it was your PM that nudged me to posting these up in here.

As an aside, whether these popularity and interaction scores reflect voting or not they are accurate indicators of social media buzz and social media interaction.

So if it all goes tits up and gets it wrong from here on in, you can still get some stuff out of it. I think it'll be right most of the time though.

Does it take into account how active the subjects are on social media themselves? i.e. Nicholas is quite active on twitter, so that could skew the results in his favour prompting his core fan base to retweet/favourite a lot etc?

I find stuff like this really interesting, and even though it won't be accurate 100% of the time due to the demographics that use social media the most it gives us a general idea of popularity. Thanks for posting Looking forward to seeing how things go over the weekend.

Sort of, it has an influencer identification mechanism which includes consumer influencer identification so it does give an indication.

Just looking at the influencers, Nic has a majority female the age groups look quite broad.

Sam C is mostly girls as you'd expect.

Sam B is surprisingly varied, mostly younger people on Twitter and older on Facebook.




Well yeah but then that would also have a knock on effect that his fans will influence others.

It does break down unique discussion from sharing though which is where you can see flaws in Sam's score.

I certainly think Sam and Nic's scores are inflated but whether or not enough to make a difference we'll see.

Thanks for sharing this- seems interesting

Just a question if I may... Does it differentiate between positive and negative online mentions or is it based on all mentions/ views etc? For example if ' Contestant X is shite' is retweeted 100 times, does it weigh the same as ' contestant x is brilliant' with the same amount of retweets ?

If it can see an interaction is definitely positive it gets a positive score, negative then a negative score and if it can't it gets a neutral score.

Yes that is a flaw in the platform

Could Sam and Nicholas' inflated score mean that their fans are spam voting them moreso than the other contestants' fans? It could make up for them having a bad unique discussion/sharing ratio.

Could do, I don't know.

An aside due to something someone said elsewhere, it's an analytics and influencer campaign platform, not a tracking and profiling platform... just in case people don't know the difference, it's no different in what it does to say Google Analytics and everything else that's in common use.

It doesn't track and store stuff, just gets metrics and highlights influencers (advocates).

Thanks. I suppose as Cowell used to say ' All publicity is good publicity' so increased mentions probably lead to more votes.
Looking forward to seeing how your system does this weekend.

The thing with figuring out whether something is positive or negative is that it just does a keyphrase check on mentions to see if certain phrases appear (a bit like a profanity filter). It works but isn't 100% reliable, far from it.

So most stuff could be neutral and also it doesn't account for the fact that someone who attracts a negative reaction could get a positive reaction in more votes because they want to one up the show.

Current (pre show):

Hannah: 6,937
Tamera: 15,428
Rough Copy: 23,184
Sam B: 50,544
Sam C: 122,286
Luke: 183,659
Nic: 215,658

Next will be posted after the show.

Aside: Last weekend Nic peaked at a staggering 1,200,000 on the Sunday which is multiples of Sam C who was behind him. Nic's metrics are just silly.

Looking like it could be Hannah vs Tamera then unless either pull something out of the bag.

I'm expecting Hannah v either Tamera, RC but it can change drastically.

Let's hope so, I'm thinking if Hannah is in the bottom 2 for the 3rd time she's toast against either of them....

If it's Hannah v tamera the vote leaks are correct

Well the metrics look as though Hannah and Tamera are vote sharing as one goes up the other down. I'm not sure whether Hannah has enough traction to avoid the bottom 2, especially after tonight's performance.

Tamera and RC seem close in popularity too.


If this order is right then it could be good for Tamera to be in the b2 this will against Hannah as RC could be next and she would probably pick up their fans as they are of the same type of genres as her.

OK, so post show, given a bit of time to get some scores in and it's quite exciting as it looks nothing like it's looked before.

Hannah: 16,003
Sam B: 75,653
Sam C: 158,379
Rough Copy: 167,605 (big kick in for RC)
Tamera: 179,383 (an all time high)
Luke: 212,191
Nic: 326,869


ETA: Just checking Tamera out by searching her on Twitter it appears that the majority of the tweets are supportive so she may be getting a god almighty sympathy vote.

Hannah, no traction.

I think Tamera will be safe and we could potentially end up with our shocker this weekend, depending on how it goes over tomorrow.

Think Tamera's will be a bunch of "Tamera forgot her lyrics!!!"

Just checked that out, yes there is a lot of that but it's supportive, in the "omg she needs our help now" and "it doesn't matter" kind of way.

From the looks of it, we will have Hannah and Sam C.