Dumb Ways To Die and social media bullshit

Karalee Evans, Mumbrella

Karalee Evans is sceptical about a comment made at last week’s Meet The Marketers evening when a Metro Trains executive said the organisation had seen a 20% drop in risky behaviour since its video went viral.

You know, up until Friday I was a big fan of the brilliant little pocket of content gold, Dumb Ways to Die.

I mean, that little animated content piece that could, with accompanying catchy tune gleaned over 39 million views on YouTube and still sits as Australia’s most viral video of all time. That’s a lot of eyeballs on a cartoon with an annoying song, demonstrating the boring topic of naughty and dangerous behavior on Melbourne’s train network.

At the time, it was reported that uniquely, the video went viral thanks in large to mobile viewers. McCann and Metro had seemingly found the formula to viral; mobile. And it didn’t hurt that industry blogs, news outlets and the general internet picked up on the video as it began to snowball, creating a Streisand effect and pushing it into viral wonderland.

But the kicker? Most of these stories, and the links and subsequent viewers, were from outside of Australia let alone the target, Melbourne.

Now, that’s not to say that the content itself wasn’t engaging or a welcome differential in what is a notoriously difficult area of communications. Public transport safety, indeed all forms of road and transport safety social marketing, is difficult. People are difficult. Behaviour is difficult. Generating ongoing change is difficult.

Take for instance level crossing safety; would you believe after 20 odd years of implementing hard-hitting, multi-faceted campaigns to change behavior and get Victorian drivers to stop at crossings, the incident rate is still largely the same?

Indeed, Victorian and Federal Governments have ploughed significant monies into complex programs of facility upgrades, improving trains and train drivers and researching the motivators of drivers to understand why they don’t give way to a 41.5 tonne train that takes at least 400 meters to stop. And then turning that into integrated social marketing (behavioural change) programs.

So that’s why Metro Trains’ seemingly unchallenged claim that a viral video that was watched by a majority of eyes not on Australian soil let alone Melbourne soil and was a short-burst awareness message leading up to Christmas holidays, “led to a 20 percent reduction in risky behavior” is social media bullshit.

Why it wasn’t challenged is a mystery. Perhaps B.J Mendeleson is right and we as an industry have truly digressed into “the asshole based economy” with “cyber utopians” driving inflated metrics that have little to no substance but support the social media marketing hyperbole that a viral video will indeed change the world [hello, Kony].

Indeed, Mendeleson asserts “packaging and selling of bullshit is currently fueling an economic bubble that’ll have disastrous effects for everybody when it pops.”

It was later clarified that this 20 percent reduction in risky behaviour was compared to annual safety figures. Seriously, that’s the clarification.

The claim is that in the two months following the video, risky behaviour at Melbourne’s train stations, including drivers at boom gates, reduced by 20 percent. Risky behaviour can be classified as driving around activated boom gates, forcing train doors open, standing over the yellow line on platforms, etc. A 20 percent reduction is extraordinary. We’re talking about hundreds of incidents on one of Australia’s biggest metropolitan rail networks.

My point? If Metro and indeed McCann really are claiming that Dumb Ways to Die led to a 20 percent reduction in risky behaviour; e.g. a significant behavioural change in a period of two months, then we really need to see the empirical evidence.

Was the campaign (read; video) evaluated on a random sample of Christmas holidays commuters to ascertain prompted and unprompted message awareness at comparable stations benchmarked for incidence of risky behaviour over the same time period? Was the messaging measured against motivators and detractors of said behavioural change outcomes? Were environmental factors such as Christmas holidays, station and facility upgrades, weather et al included?

They’re just the easy questions.

Don’t get me wrong; I loved the video and the use of fun and frivolous delivery to raise awareness of behaviours that need to be changed. The execution was brilliant. McCann’s CD and team should be sitting back still revelling in creating such a viral piece of content they’ll get prospects coming to them with a simple brief of ‘make it go viral like Dumb Ways to Die’, all while they’re polishing their creative awards for a content piece that worked to cut through. And they deserve it.

But if we’re going to be taken seriously and avoid becoming an industry that already smells like an asshole based economy; we need to be a bit more transparent and apply more critical thinking over a longer period of time before we agree that the campaign was indeed effective in what it sort to achieve; a behaviour change.

And if Metro put the figures where their claims are and disclose that they have indeed evaluated the campaign messaging against the annual safety figures in a way that can categorically prove Dumb Ways to Die wasn’t dumb, but did the impossible, yes of course, I’ll apologise. Publicly.

Until then, let’s just write the awards articles for Dumb Ways to Die, revelling in the social media viral video that changed behaviour and saved lives. In two months. Without disclosure of evaluation methods. Or a bit more time to cook to you know, actually change behaviour. On a sustained, and therefore effective level.

Because who doesn’t want to be a cyber utopian in a bullshit economy of social media hyperbole?


  1. Citi Zen A
    11 Feb 13
    12:01 pm

  2. “we really need to see the imperial evidence.”


    Unless you’re looking for Darth Vader and an Excel Spreadsheet.

  3. Keith De La Rue
    11 Feb 13
    12:06 pm

  4. Fair point. We already have enough sloppy thinking without that. And yes, I loved the video too!

    One thing you may like to fix in the article – I believe you mean ” empirical evidence”, not “imperial evidence”.

  5. Andrew
    11 Feb 13
    12:10 pm

  6. I appreciate that the 20% statistic probably doesn’t hold water. But that title is absolute linkbait.

    Why not just “Dumb Ways to Die Statistics Don’t Hold Water”?

  7. mumbrella
    11 Feb 13
    12:11 pm

  8. Hi Citi Zen A and KDLR, Empirical/ Imperial typo now fixed…


    Tim – Mumbrella

  9. YaThinkN
    11 Feb 13
    12:53 pm

  10. Great article. True, playing silly buggers with Social Media or Internet stats drives me insane and makes the whole industry look dodgy :(

  11. nextbrett
    11 Feb 13
    1:28 pm

  12. Was my first thought exactly.

    Would love to see the statistical analysis to back this up with a good few years of data collected.

    Really the success of this campaign is in the follow up, with a brilliant platform to converse with consumers set, now the real job of changing behavior can begin.

  13. Citi Zen A
    11 Feb 13
    1:41 pm

  14. @Tim Apologies for being _that_ guy. It just amused me…

    Really interesting article, though. It’s a bold statistic to claim, hope they respond to clarify.

  15. Karalee
    11 Feb 13
    2:00 pm

  16. Apologies for the awful typos, guys (and gals). Lesson; don’t write an op-ed in the pub on a Sunday after more than 2 beers and rely on computer-enabled spell check 😉

    @Brett – totally agree. The disappointing thing is exactly as you’ve pointed out; Dumb Ways to Die phase one (video) was creating interest in a public so used to hard-hitting, shock campaigns. The key is to progress the conversation to ‘correct behaviours’ and have a long tail program of change. Then the 20 percent reduction stat would be believable. And awesome

  17. nell schofield
    11 Feb 13
    2:02 pm

  18. yes, i agree that social media is largely bullshit and in a coupla years time everyone will look back on it and say “what the hell was that really all about?”

  19. Grammatical correction needed
    11 Feb 13
    2:13 pm

  20. Current phrase:
    “it sort to achieve; a behaviour change.”

    it should be:
    “it sought to achieve; a behaviour change.”

    You could also change ‘behaviour’ to ‘behavioural’ but it’s a coin toss.

  21. Offal Spokesperson
    11 Feb 13
    2:23 pm

  22. As i said in previous comments about DWTD, likes, clicks and talk is not what means an advert is successful.

    The viral/online/digital world rarely translates to (real) results.

  23. Bob
    11 Feb 13
    2:24 pm

  24. It wouldn’t surprise me if people were humming the tune as they wandered across the tracks in front of an on-coming train.

  25. Nicole Stirling
    11 Feb 13
    2:29 pm

  26. Excuse my ignorance if this has already been addressed but exactly who is this video meant to target? Are the railway rule breakers domestic or international? After all, Bondi Rescue is a great program for educating people on surf safety but it’s aired in Australia and have you noticed how the majority of rescues involve international tourists? Wrong market! If the Vic government is targeting foreigners before they hit our soil then perhaps the 20% figure is valid (to some extent – comparing to annual safety figures isn’t an accurate marker) since they reached their target audience. I’d be keen to see if they release any other “interesting” analytical insights now.

  27. Snooze
    11 Feb 13
    2:33 pm

  28. Sorry, sorry, sorry, but I think you meant “sought to achieve” not “sort to achieve”. Yes, I am that annoying person.

  29. Client
    11 Feb 13
    2:39 pm

  30. Great article.
    I just can’t help but feel that much of the quantum claimed success in social media is a bit like the early internet start ups…Views = a success
    The bubble will only burst when more people ask rigorous questions rather than buying into the hype !

  31. Janelle
    11 Feb 13
    2:42 pm

  32. Brilliant and honest piece Karalee.

    Recently signed a new NFP client who had been advised by two social media ‘gurus’ (cowboys) that they could run a social media campaign for them, that would make 10 million dollars in 10 days. All it would cost the NFP is $60K…

    These kinds of people and stats like the above cause issues for social media as a communications tool and create distrust in PR/Marketing professionals.

  33. AdGrunt
    11 Feb 13
    2:45 pm

  34. Welcome to every award case study since 1853.

    83% of stats are made up, etc.

  35. Kelly
    11 Feb 13
    3:15 pm

  36. Sorry what? I was distracted by all the semi-colons. http://theoatmeal.com/comics/semicolon

  37. Dude
    11 Feb 13
    3:40 pm

  38. How did McCann crack viral with mobile? That makes no sense.

  39. NS
    11 Feb 13
    3:52 pm

  40. @adgrunt – that might be true, but these are more likely to be the losing submissions 😉

  41. Old School
    11 Feb 13
    4:42 pm

  42. I agree with all above that the concept and execution of DWTD is well worthy of recognition, but I’ve been dissapointed with how this has translated to the physical livery on platforms – I would’ve expected to see more than the odd A3 poster. All very well to (admirably) harness the power of social media and build awareness, but where’s the follow-through? e.g. traditional non-sexy communications tools like point-of-decision prompts have a much stronger and longer track record in actually influencing behaviour change.

  43. Angie T
    11 Feb 13
    4:55 pm

  44. Thanks Karalee for calling the ’20 percent’ to account. I couldn’t believe the claim when I read the article and would also like to see the stats and definition of risky behaviour. On the other hand, happy to be proven wrong by Metro if it means people stop throwing themselves between the doors of a train when they are snapping shut …

  45. blubber
    11 Feb 13
    5:01 pm

  46. Isn’t 4 X “bullshit” a bit of an overkill?

    Haven’t seen this language used in a while either “… smells like an asshole based economy..”

  47. The Fat Controller
    11 Feb 13
    10:27 pm

  48. I didn’t know they made ‘meters’ to stop trains.

    And it takes at least 400 of them to stop a 41.5 tonne train, so does this mean that each ‘meter’ can stop a roughly 100kg object. Cool. I’m off to SuperCheapAuto to try and find one!

  49. Doug Chapman
    12 Feb 13
    10:13 am

  50. Hi Karalee, good call out of the BS factor. One small point, methinks a train weighs more like 415 tonnes not 41.5 tonnes. Or am I exaggerating!

  51. Karalee
    12 Feb 13
    11:02 am

  52. Thanks, Doug.

    To your point (and others) about weight and stopping distance, it’s interesting – the weight of a typical train varies across Metro’s rolling stock, and how many couplings it has, etc. Weight cited is a single carriage Siemens (source: http://en.wikipedia.org/wiki/Siemens_Nexas) – so yes, if it were a peak 6-car Siemens, you’re looking at 6 x 41.5 = not going to stop in time for you or your car trying to beat the boom gates…

    It’s also now that I admit my secret pass time as a gunzel 😉

  53. Leethal
    12 Feb 13
    12:56 pm

  54. I was overseas when it first went out, and was pretty amazed at the timing, especially after a tragic rail accident(?) on the previous weekend. My question is, what do groups like Beyond Blue think of such campaigns. These groups spend all this time and money educating the public, yet a stupid catchy tune that mocks methods of death is actually sanctioned by a government department. It’s good to see that it has worked so effectively… NOT!!! 39 Million hits, yet most of them were from everywhere but the target market, Melbourne.

  55. Blind Freddy
    12 Feb 13
    2:31 pm

  56. Try these stats on for size:

    90% think DWTD was an engaging piece of creative.

    50% of comments focussed on tone and grammar trivia — not the substance of Karalee’s opinion piece.

    85% of the media ran the “20% reduction in risky behaviour” without so much as beg-yours.

    0% of anybody has seen a shred of evidence to substantiate Metro’s claim.


  57. Mike
    13 Feb 13
    10:49 am

  58. I’m with Karalee, excellent article. Notwithstanding the silliness of the Metro claim, it IS the sort of thing that might well be cited by an agency next time it pitches. I’ll write this slowly – Having a lot of hits – and this one certainly has – is not the only measure of success.
    As Karalee points out, it really is a hit for the agency, a big one, and for Metro which (eventually!) is seen to be the client it’s helped raise its profile in a positive way. The reality is that people hurting or killing themselves by train is a massive community cost in terms of delays to me as a passenger, metro as a business which is penalised by the government for late trains even if it’s nothing to do with them.
    The trouble with social media is that it is often portrayed as THE solution, usually by acolytes on the bandwagon. This ‘campaign covered all manner of ‘dumb ways to die’ not just falling/jumping in front of trains, driving around boom gates, tagging or riding on the outside of trains. Most of the offerings in the catchy song were in fact, dumb. Like Karalee I’d like to see the figures for other deaths, serious injuries and attempts at self-harm before and after this campaign. Perhaps Mumbrella could have a chat to beyondblue, Vic Pol or the ambulance service which actually deal with them before we decide that this campaign is really a success at the human level.

  59. Yes, Yes, Yes
    13 Feb 13
    12:15 pm

  60. Thank-you for saying what so many of us can’t out of fear of backlash from our peers for daring to be a ‘non-believer’.

  61. Karalee
    14 Feb 13
    2:08 pm

  62. Looks like Metro are standing by their claim, in fact, increasing it to 30% reduction including cars at boom gates. No evidence cited. Again. http://www.theage.com.au/victo.....2eelt.html

    From today’s Age:

    “Metro’s quirky Dumb Ways to Die campaign – which has amassed 40 million hits on YouTube after going viral – has cut the number of ‘near-miss’ accidents by more than 30 per cent… for November to January 2011/12 there were 13.29 near-misses per million kilometres travelled by Metro trains on Melbourne’s suburban train network.
    For November to January 2012/13 – after the campaign started – there were 9.17 near-misses per million kilometres.”


  63. Happy
    16 Feb 13
    7:46 pm

  64. Karalee, Im stoked these figures are coming out now and that a good effie campaign will be coming out. Your attempts to belittle it (and thats what you’ve done) is in itself unsubstantiated.

  65. Karalee
    17 Feb 13
    3:45 pm

  66. @Happy Would be stoked if those stats are proven to be from
    effect, not correlation, too. But nothing belittling about asking
    for substantiation. We should be doing that more, not

  67. Thirty five
    18 Feb 13
    2:11 am

  68. @34: It is, however both belittling and less unprofessional to slander without prior confirmation of deception.

    Poor understanding of communications sensitivity in a small community comes to mind as well.

    Until something that appears questionable is proven to actually BE that way, how can one go ahead and commit an even deeper trespass by jumping on the peak hour express to potty-mouth central?

    Making sure to not write an article after “more than two beers” is good advice. And while we’re on the subject of behaviour change…

    Analysis of the above Comments show beyond doubt that several have been deceptively planted.

    Perhaps “good” blogging PR; the real ‘intestinal end’ of online behaviour that could more than do with a clean up.

    There are multiple Comment characters above showing the precise same writing pattern. Name hyperlinks cross-checked with networks for equally obvious signs of good old bs also came up trumps.

    Moving on…

    The number of Melbourne-based youth required to view the YouTube video for a significantly measured effect on behaviour would be about 60,000.

    An actual pertinent question is how many locals have viewed it. Or, no, wait, you must know that right.

    Nevertheless I would expect serious egg-on-face, not just invisible egg such as but not limited to reduced new business as a direct result from the conduct in the article, but from the obviously massive effect that DWTD is having on both the reduction of high-risk behaviour, and on reduced incidences of self-harm acts involving trains in Melbourne over the coming years.

    Whether by accident or on purpose, DWTD manipulates multiple behavioural (and antecedent) change contributory constructs including mortality salience, attention and recall, empathy vs perspective-taking, cognitive structuring and heuristics, and false consensus effect (def check that one out on wiki) in such an effective way that the only mystery is when its behavioural effects will begin to wane.

    To wit the question, do you know anything about behavioural analysis?

  69. Thirty five
    18 Feb 13
    2:12 am

  70. *less than professional*

  71. Karalee
    18 Feb 13
    5:39 pm

  72. @35 (anonymous behavoural expert that can reference TEDx) Would be good to clarify how asking for proof of claims is slander. Also, would be good for you to clarify if you are inferring I, or indeed others, have astroturfed this post.

    Because while asking critical questions of metrics is perhaps uncomfortable, it is not unprofessional nor slanderous. And goodness me, if people can’t distinguish quotes from a Book (the colorful language you call out), between unquoted passage, it is a worry.

    And are you seriously asserting my questions here, all while praising the campaign, will negatively impact New Business?

  73. eaon
    21 Feb 13
    9:08 am

  74. started out as a comment but it’s got too long so I made it into a separate post.

  75. @eaon
    21 Feb 13
    10:16 am

  76. thank you eaon for taking the time to write something balanced, intelligent and non-hysterical.

    the fact that you probably didn’t spend the afternoon at the pub prior to hitting the keyboard helps too.

  77. The Accountant
    21 Feb 13
    11:47 am

  78. @Hard to argue with these numbers

    It is easy to argue with the numbers. Comparing one year in a period to another means nothings. Comparing one year to another in a period over 5 years means something. Also, state authorities have a habit of moving the goalposts; a near miss may have been redefined, much like the rail authority always redefines on time running according to their own agenda.

    S it is valid to say that year on year there was a decrease. It is utterly invalid to say it is a trend.

  79. Client
    21 Feb 13
    2:07 pm

  80. Eaon……I read you blog……”The 20% number acts as ‘social proof’ that the behaviour has begun to change.

    Because in any conditions of uncertainty (ie all the bloody time) we take cues on how to behave from what we see or percieve to be the norm (what others are doing) then the mere description of a shift in behaviour acts as a spur to the continuation of that behaviour”

    So……I tell my boss I am doing a world class job, I am growing sales by 20%……and even though I am not…..it has the effect! Awesome! Sounds more like “bull shit factor” buddy….

  81. Number Cruncher
    21 Feb 13
    2:54 pm

  82. @ The Accountant

    Quite right too. I have been in so many meetings and argued with colleagues, when they only take the benefits for the business out of intelligence v taking the actual true facts:

    – The benefits are taken out and are spun to the masses for short term gain. To customers and staff alike (most staff can see through them, because they live and breath the product or service…)

    – What should occur is the actual facts are taken out, addressed and acted upon for the long term gain of the business, customers and staff.

    Sadly in most larger companies, it is everyman for him / herself and everyone is trying to impress the boss, get a pay rise or a promotion and long term the business suffers. For these corporate whores it is all about “me, me, me” and short term wins. Greed.

    I have witnessed apparent (in the eyes of their bosses and the yes men and women below them) great leaders, fudging numbers to suit their agenda’s. In reality they are con artists and deserve to fail, however often are able to stay camouflaged in their bureaucratic worlds. They rarely have any real friends and live to work, act a bit like David Brent – you know who I am talking about…

    Stats – certainly always take them with a pinch of salt.

  83. eaon
    21 Feb 13
    4:46 pm

  84. @client very droll. but 8 out of ten cats prefer my explanation.

  85. dude
    24 Feb 13
    7:17 pm

  86. how many Australia views and how many victorian views did it receive? Like anything that goes viral far more views are from other countries but that doesn’t mean it did not get many in views Vic. So what is the number? If it got 500,000 AU with no promotion that is a big success – realise this is not the point of the story but interested to know

  87. paul
    7 Mar 13
    3:59 pm

  88. Re, Streisand effect… You are using that word. I do not think it means what you think it means.