How IRN-BRU made its first million

The short story

This is what an IRN-BRU film being shared across Twitter looks like under the data visualisation equivalent of an electron microscope.

More precisely, this is a data snapshot four weeks after a YouTube link to a new IRN-BRU ad was first tweeted by a single account with just over 300 followers.

Each red dot, or node, represents a Twitter account that mentioned our YouTube link. And the black lines, or edges, represent the number of reactions (mainly retweets) generated by each mention.

This is a snapshot of an unusual launch strategy that worked.

Standard practice when launching a new ad is to chuck the social media and PR equivalent of the kitchen sink at it.

This was different.

We gave the YouTube link exclusively to a single fan of the brand. We let her release the film into the wild and we used the Topsy API to watch what happened next.

We watched the film rack up over 1 million views in less than a month.

We watched the link being shared across Twitter.

And we watched three very different network dynamics behind that sharing.

1) A few high-influence accounts

This Twitter dynamic coincided with 100,000 YouTube views within 24 hours of launch.

2) Lots of small, connected groups

This was the dominant Twitter dynamic as YouTube views increased from 100,000 to over 650,000 in the next three weeks.

3) Lots of isolated mentions

Three weeks after the initial tweet, the ad was aired in a few high profile television spots during the European Football Championship.

This generated an additional 300,000 YouTube views in 48 hours. And it triggered hundreds of additional, isolated mentions of the link on Twitter.

The data snapshot above encapsulates all three dynamics.

The large number of unconnected nodes at the centre represents the wide reach delivered by television.

The nodes round the edges, each responsible for many reactions, are the influential accounts that shared the link in the first 24 hours.

And in between are the large numbers of small, connected groups that represent three weeks worth of Twitter word of mouth.

This view of the data is a visual summary of how great content, a bold social media strategy, and broadcast media worked together to make IRN-BRU’s “first million”.

Read on for the full story, including the film itself, the subtleties of the launch strategy and a detailed explanation of the data visualisation, which was created by our clever friend Francesco at sister agency Face.

Awesome content

At the time of writing this IRN-BRU film has had over 1.5 million YouTube views. It is the brainchild of another sister agency, Leith.

It’s hard to predict the word of mouth performance of branded content in advance, but we *knew* from the moment that we saw the script that this was going to be special.

IRN-BRU and “Fanny” is as close to a sure social media thing as it gets. That belief and confidence led us to be more adventurous with our digital seeding and amplification strategy than usual.

Our belief in the awesomeness of the content gave us strategic options.

Rather than going in with all social media and PR guns blazing, we decided to give the new ad to a single, randomly selected fan and see what happened.

Meet Rachel…

Rachel from Motherwell

One IRN-BRU fan would be given exclusive access to the YouTube link for the new ad. And they would enjoy the 15 minutes of internet fame that came with sharing that link with the world.

To be selected for this honour you had to enter a simple lottery competition.

And we launched the lottery with a one-word tweet.

(At this point we understood the Fanny reference but no-one else did). was a landing page on the IRN-BRU website. It announced that one lucky fan would launch the new IRN-BRU ad and invited people to enter their details.

545 eager Fannys beavers did so.

Rachel Orr, (@larachie as she is known to her Twitter friends) was the chosen one.

She is a 23 year-old student from Motherwell, who had 153 Twitter followers when we met her.

Before she launched the ad we did everything we could to boost her profile.

We helped to double her following to 329 before she tweeted the link into an unsuspecting Twittersphere.

The most influential tweeters in Scotland

As soon as Rachel was chosen as our “vector” we did all we could, as publicly as we could, to draw attention to her through IRN-BRU’s various social profiles.

But we were also working behind the scenes to pique the interest of some of Scotland’s most influential tweeters.

We asked them to follow Rachel and retweet her when she first shared the link.

But, no matter how warm the relationships, asking for social media favours can be a bit crass.

So we introduced an added incentive, which took the form of a friendly challenge. Because we were collecting Twitter data from the outset, we could assess the “influence” of each person mentioning the link as measured by the number of reactions generated by their tweets.

Yes it’s a crude measure of influence but it allowed us to write this blog post.

The post announced our collaboration with Face to collect and visualise the campaign data.

And it issued an invitation for people to demonstrate and test their influence. “Retweet our YouTube link and we’ll scientifically gauge the level of reaction you generate”, or words to that effect.

The blog post link was shared with various people via Twitter direct message, after which it gained some public momentum of its own through LinkedIn, Google+ and Twitter.

So the data visualisation aspect of this story was not just an exercise in insight and learning. It wasn’t just about telling us what had happened after the event. It also played a valuable role during the event, helping us to recruit influencers to the early seeding effort.

Let’s look at the progression of the YouTube link over time to gauge the impact of our seeding strategy and the role of influencers within it.

Rachel tweeted the link on May 20th.

Within 24 hours the ad had been viewed 100,000 times.

And this is how the Twitter data looked a day into the campaign.

The red nodes are individual Twitter accounts.

The black edges represent reactions (retweets) to the sharing of the YouTube link by each individual.

The further from the centre the greater the influence of the node, as measured by the number of reactions generated.

It is clear that these influential early tweeters played an important, if not vital, role in generating initial momentum behind this piece of content.

The role and identity of the individuals concerned can be seen more clearly in this alternative view of the data.

This cut of the data includes all mentions of the YouTube link, including those generated by the television spots.

In the centre is a dense mass of individuals who mentioned the link, but who did not generate any secondary reactions. This is the “wide reach” effect (in social media terms) of television.

As we move away from the centre, the size of the red nodes increases in proportion to the number of reactions generated.

The further from the centre the greater the level of reaction generated. Rachel, the original vector, is the most influential tweeter associated with this campaign. But there are several dozen other “influencers”, all of whom played an important role in generating word of mouth momentum.

You know who you are. Thank you.

The power of many, small, connected groups

Our influencers played a significant role in propelling the link to 100k views in the first 24 hours after launch.

But we observed a very different dynamic as the number of YouTube views grew from 100k to 650k over the next 21 days.

Here is another view of the data.

The previous visualisations illustrate the reactions to tweets.

This view shows the connections of the people actually doing the tweeting.

These people (grey nodes) all tweeted the YouTube link. And what we can see here is the follow connections (red edges) between people who have that YouTube link in common.

The size of the nodes in this image is not related to the number of reactions generated. It represents the number of connections to other profiles in the “network of networks”.

The data cut for this visualisation was taken on 10th June, before the ad was shown on TV, so it is a graphic representation of the dense network of mini networks through which the link spread under its own steam.

It’s worth saying that again: the dense network of mini networks.

We assume, but we can’t prove, that a similar dynamic was happening on Facebook. However, anyone familiar with the work of Paul Adams, based on his access to “big” Facebook data, will probably side with us in thinking that this is a fair assumption.

Everyone wants word of mouth. Everyone wants content to go viral.

And it’s intellectually comforting to believe that this can be achieved by targeting relatively few influential individuals who will put your content in front of large numbers of eager and attentive consumers.

But social media life isn’t like that.

As we’ve seen, influencers can play a role. But that role is limited.

To achieve significant reach through word of mouth in a social environment you need content that is relevant to many small networks of people. The biggest influencer in this case study was the content itself.

Humour unifies.

Originality and creative risk taking unifies.

And, in Scotland, IRN-BRU itself unifies.

This big idea relied on lots of small groups of people, with the above factors in common, to generate organic word of mouth at any kind of scale.

The impact of television

Watching this campaign move up through the gears by way of data analysis was an education.

1st Gear – the YouTube link seeded by a single, solitary fan of the brand.
2nd Gear – 100,000 views in 24 hours thanks to some co-ordinated influencer activity.
3rd Gear – 650,000 views over 3 weeks as a result of relevant content working its way through a large number of small micro-networks.

4th Gear was an unfashionable offline medium called television.

The Fanny film was aired in just a few high profile spots during the European Football Championships.

But those few spots took the scale of this campaign to another level, as can be seen in the diagram below.


The inclusion of television in the mix was directly responsible for an additional 300,000 views in 48 hours, and for propelling the YouTube version of the film to 1.5 million views over the following fortnight.

There is a lot of misguided “either/or” talk when it comes to social media and broadcast channels.

As this case study hopefully shows, an AND approach is far more effective than an OR approach.

The two channels play entirely complementary roles.

Social – significant reach over a period of weeks with the added credibility and relevance that comes from personal endorsement. (DEEP REACH).

Television – mass scale reach in a couple of days. (WIDE REACH).

These complementary roles can best be seen in this visualisation of the Twitter data, which combines both word of mouth and TV driven mentions of the YouTube link.

You can clearly see the combination of deeper, networked mentions and more shallow unconnected mentions.

The visualisation, and the campaign it represents, is all the richer for the integrated online/offline approach to this campaign.


There is always a degree of post-rationalisation to any case study.

The posthumous telling of the story always adds a comforting linearity to the narrative that wasn’t necessarily apparent at the time.

In this instance Twitter and Television weren’t the only things that happened.

We chose to launch the link on Twitter because the channel is a good control laboratory for investigating social dynamics. What happens on Twitter is a (large) niche version of what happens elsewhere. On Twitter we can control the variables and the resulting data is entirely trackable.

So Twitter was our seeding channel of choice.

But, once the link was in the wild, it will have been taken from Twitter and shared via Facebook, email and other channels. We have no way of accurately allocating the credit, as measured by YouTube views, for the overall word of mouth effect.

This project was about how a link spread across Twitter, not about how many views Twitter generated.

Nonetheless, the principle of being relevant to many small networks of people definitely stands.

Once the word of mouth effect developed its initial momentum, other nice things happened.

Within 3 days of Rachel posting her tweet, the Fanny film was featured as The Poke’s “viral of the day”.


And the film trended on YouTube’s own dashboard.


Events like these obviously boost profile and reach.

There was also a strand of paid for activity on YouTube itself. And this activity obviously made a significant contribution to the number of views that the film received. In fact YouTube analytics tell us that 28% of total views can be attributed to YouTube advertising.

It is worth stating again that this case study is about how the link was shared across Twitter rather than about how many views can be attributed to Twitter.

The main mechanism by which the link was shared across Twitter, namely lots of small connected groups, was unaffected by the YouTube advertising. It took the power of television to alter that dynamic.

In a nutshell

Awesome content is the key to interesting social media strategy.

Influencer theory is useful but only up to a point.

Many small networks are more important than a few highly followed individuals.

Think AND not OR when it comes to the role of social and broadcast channels.

Film content by Leith

Social media strategy and execution by Blonde

Twitter data analysis and visualisation by Face Group