For a time during World War II, the chances of a member of US bomber crews actually making it back from any given mission were on the side of slim.
The nature of the work meant that bombers were out for a long time; they were massive cumbersome planes visible from a long way away, and their ability to do serious damage if successful meant they were the number one targets of both the guns on the ground and in the air.
For the bomber crews, each subsequent mission piled up the odds against them making it back this time.
And the Army Air Force couldn't make planes quick enough to replace the ones that went down.
The situation was unsustainable.
In the hope of a solution the military engineers examined the bombers that made it back from their missions.
Patterns started to emerge. They saw the damage tended to accumulate in the same places.
They observed clusters of bullet holes along both wings, down the center of the bomber's body and around the tail gunner area.
The answer was clear. The bombers needed more armour.
However they couldn't just reinforce the entire plane - the weight would prevent them from even taking off.
So, based on the data they had, the obvious solution was to put thicker protection where they see the most damage, and ramp up reinforcement in the areas where the bullet holes clustered.
Just to be sure they were doing the right thing, the engineers called in a statistician, one Abraham Wald.
Wald was a member of the military's 'Applied Mathematics Panel' - a secret boffin unit working out of Columbia University applying the science of probability and statistics to the war effort.
And good job they did, as Wald saw immediately that they were about to make exactly the wrong decision.
Because the common patterns of bullet holes actually showed where the planes were strongest.
The holes showed where a bomber could be hit repeatedly and still make it back.
The planes that didn't make it home were being hit in different places.
Until Wald's intervention the military were overly focused on the planes that made it home and almost made a potentially catastrophic decision by ignoring the planes that got shot down.
That's a long-winded way round to pointing out that the same survivorship bias is prevalent in marketing departments and agencies every day.
Just like our Air Force engineers, it's easy for marketers and agencies to get distracted by the high response rates and dramatic ROI that appears to fall out of marketing discounts or offers to a particular segment of heavy customers.
On the surface it appears logical. But these are people who are likely to buy anyway.
These customers are the clusters of bullet holes that registered on the wings of the planes that made it home.
And the bigger the plane the bigger those clusters will naturally be. This is because the bigger brands in any given category tend to have slightly higher rates of bullet hole frequency (and loyalty) than their smaller competitors.
For just about any brand, attracting the mass of category buyers who are light and non-buyers of the particular brand - just like the bullet holes that didn't show up, or barely registered on the bombers that made it home - holds the key as to whether the mission is going to be successful or not.
Survivorship bias in marketing is your tendency to focus on heavy buyers instead of light or non-buyers and on activities that look like winners in the short term, which turn out to be losers in the long game.