There’s this wonderful little piece whizzing around the Internet lately that showcases all the highly-improbable things that are more likely to happen to you than your consciously clicking a banner ad.

I get it.  It’s healthy for our industry to have a frank conversation about click-through rate (CTR), click fraud and response rates in general.  The strategist in me wants to ignore this cute article, since anybody who launches a broad-based, untargeted banner campaign in the hopes of achieving high CTR has probably already figured out that there are more efficient things to do in digital media if your goal is ad clicks.

But the math geek in me won’t let this go.  The figures circulating with this story are misleading at best, and downright wrong at worst.  So let’s set the record straight.

First, let’s talk about surviving a plane crash.

It surprised me to find out that 95.7% of people involved in plane crashes survive, according to the NTSB.  (Warning: PDF, see page 7.)  Granted, the data is somewhat old (cumulatively from 1983 to 2000), but bear with me.

To survive a plane crash, first you have to get into one.  Scouring the web for a citation-worthy risk of getting into a plane crash, I didn’t find much.  But I did find a number representing your annualized chances of dying in a plane crash.  That number is 1 in 11 million.  It’s a U.S. Department of Transportation number that I’ve seen cited all over the web.  And it’s probably something you’re going to want to repeat over and over to yourself when you experience turbulence on your next trip to Disney with the family.

Now, if your chance of dying from a plane crash is 1 in 11 million (0.00000909%), and only 4.3% of people involved in plane crashes die, then your chance of being in a plane crash at all is somewhere in the neighborhood of 0.00021%.  Multiplying that number by the 95.7% gives you a chance of being in a plane crash and surviving of around 0.000202%.  Just to use round numbers, let’s call that two ten-thousandths of a percent annual risk.

And then there’s the annualized chance of clicking on a banner ad.  Here’s a neat page with some of the stats we’ll need.  If comScore says 5.3 trillion ads were served to U.S. users last year, and 31 percent of them were non-viewable, that’s 3.657 trillion that were viewable.  (I know, I know.  I’m giving benefit of the doubt to a company that’s got a dog in the viewability hunt.  Bear with me.)

The average web population in the U.S. (3 month average ending December 2012) was 220,582,000 according to comScore.  My calculator tells me that means your average American who went on the web via their PC last year was served a little more than 16.5 banner ads last year.  (Yes, I know I’m discounting mobile, and I’m failing to account for non-web users.  Again, bear with me.)  Assuming each of those banner ads had a 0.1% chance of being clicked on, the average American had a 1.65% chance of clicking on a banner ad last year.

Even if we assume that half of those clicks were fraudulent or accidental, we’re still in the neighborhood of 0.825%

So, does 0.0002% beat 0.825%?  Nope.

Sorry to geek out on you like that.

Realistically, the conversation should be about whether it’s a good idea at all to run huge, untargeted banner campaigns in order to garner clicks.  If you’ve spent more than six months in digital media, you already know the answer to that one.  And it’s good to have conversations about fraud, viewability and whether CTR is even a valid metric for a campaign’s success.

But we all know that unless somebody debunks this, somebody in a marketing meeting is going to trot out the “You’re more likely to survive a plane crash than click on a banner ad” argument against spending in digital.  That type of sound bite can be damaging.  Not only does it assume that the entire value of a banner ad is represented by the clicks it can deliver (it’s not), but it also dismisses the ROI that well-targeted digital ads can bring to the table.

For instance, our agency has a long history of being able to show that the ROI for banners can often come from the effect they have on brands.  If you measure awareness, purchase intent and brand preference, for example, you may find that your ROI comes not from people who immediately click on your banner and visit your website, but from the preference they develop for your brand and the subsequent lift in off-the-shelf sales.  We’ve also got a rich history of being able to illuminate the other tactics to invest in if you absolutely, positively must increase traffic to your website.  (Hint: It’s not massive, untargeted banner campaigns.)

  1. Ari Jacoby March 27, 2013

    Now all you have to do is account for small/personal/non-commercial grade aircraft

  2. Tom Hespos March 27, 2013

    Somehow, I don’t think that will get you there. Even if you triple the number, you’re still orders of magnitude off.

  3. Robert FOrmentin March 27, 2013

    I was following your comments this morning Tom and had the same reaction you did. A quick search revealed the same data you uncovered. At that point I knew that you were making a valid point even if I didn’t complete the math process you did. I didn’t need to. Reminds me of another story I read; mobile banner CTR 10 times higher than online. Of course anyone with two brain cells will know, through experience, that many of those “clicks” are mistakes. Happens to me all the time. I don’t know how many, but the story never even mentioned it.

    Your very important statement “that type of sound bite can be damaging” should not be overlooked. I see this all the time; publications making wild statements about our industry using ridiculous math. Being a math geek like you, I also start running numbers when I see them. What I’ve found in many cases is that these types of stories lack substance because the math they use in making their point is flawed. My advice: unless you’re willing to invest the time to do real research, resist the urge to make these claims for the sake of posting a story.

  4. John Cappiello March 27, 2013

    When statistics like survivability of an incident are cited, they’re often assuming the incident in the first place…

  5. Tom Hespos March 27, 2013

    I see what you’re saying, John, but somehow, “being in a plane crash and surviving it” wouldn’t have the same impact as “surviving a plane crash.” I think that’s part of the appeal of using this in a PR effort – people tend to lump the improbability of getting into the plane crash in the first place with the possibility of surviving it.

  6. Chris March 27, 2013

    Great job, Tom. Don’t believe the hype…

  7. Gonzo March 29, 2013

    One point to start off with – the blip they actually show doesn’t say anything about annualizing the plane crash side of the equation. That inherently skews the meaningfulness of the claim, but also gives them another 78.2x in their favor. By your math, bringing it to .0105 vs .0002 without changing any other parts. Of course, they need it to be .000042 rather than .0002 to fit your calculations with their 475.28x claim. A difference of 250x. The quickest way to get there is to adjust the least accurate number in this calculation – the accidental and fraudulent click through rate. If we change it from 50% to 0.2%, then it works.

    I have personally never once clicked a banner ad for the intended purpose of the ad. I have, however, clicked a friend’s banner ad to help him see if it was counting clicks or if the agency was cheating him. I’ve also accidentally clicked many times. To me, a 0.2% intentional rate seems realistic, but I have no data, and the only data that could be had on the matter would be survey data, which would be limited to people who bothered to take it – and it’s hard not to believe a willingness to participate would be directly linked to a willingness to be advertised to.

    I don’t believe their blurb is straightforward truth, but I don’t think it’s the bald-faced lie your calculations imply either. It seems to be careful wording to make a point.

    • Gonzo March 29, 2013

      Actually, this is simpler than we’re making it.

      You have a 95.7% chance of surviving a plane crash (it doesn’t say “get into and”, just “survive a plane crash”). You assume a .1% chance of clicking a banner ad at all, intentionally or not (it doesn’t specify intentionally). 95.7/.1 = 957. Given their 475.28x claim, they are assuming closer to a .2% chance.

      So, true, but not as relevant as it sounds.

      • Tom Hespos March 29, 2013

        I know it doesn’t say “get into,” but that’s what most people would think when assessing whether or not they’ll be surviving a plane crash in the near future. One necessitates the other.

        By way of example, if a fortune teller were to tell you that you’ll survive a shark attack sometime in the next year, your thoughts wouldn’t gravitate toward the percentage of people who survive shark attacks. You’d probably call bullshit on the fortune teller because of how farfetched it would be to be actually be one of the small handful of people who get attacked by a shark each year.

    • Tom Hespos March 29, 2013

      I probably should have mentioned that I got the accidental click and fraud data from Ari Jacoby of Solve Media in a Facebook thread. We were talking about this very subject, since Solve has been publishing these things, and he said his data were showing roughly 25% fraud and I should probably use 25% for accidental clicks as well. Instead of challenging it, I thought I’d spot him the 50%, since it makes little difference to the comparative numbers, which in my mind are still orders of magnitude off from one another.

  8. Stuart February 14, 2014

    The article above suggests that the ods of clicking a banner ad are 1.65% per year?

    So the author is saying the average American clicks on a banner ad approximately once in 60 years? This is clearly not the case.

    Actually I think you will find that 3.657 trillion banner ads served to 220,582,000 Americans is 16,578 banner ads per year (1000 times more than suggested in the article).

    If we assume that the CTR is 0.1% then the average American clicks on 16.5 banner adds PER YEAR.

    The chances of clicking on a banner ad over the course of a year is 1,657% (you will click on 16 ads).

    To summarize, the average american will click on a banner ad accidentally or deliberately approximately every 22 days.

    And to put the original issue to bed. You won’t be walking away from plane crashes with that kind of frequency.

3 Trackbacks

  1. […] mere fact that marketers are debating whether you are actually more likely to survive a plane crash than click on a banner ad is as telling as any statistic out […]

Comments

Post a comment

Your email address will not be published. Required fields are marked *

*

Underscore Marketing LLC on INC5000
Copyright © Underscore Marketing. All rights reserved.