World War II ended in 1945. Hiroo Onoda, an intelligence officer for the Imperial Japanese Army, didn’t get the memo. Holed up in the mountains of a remote Philippine island, Onoda and two comrades continued to fight. For the next few decades, they remained a dangerous menace to island locals. The sole survivor of his group after nearly 30 years, Onoda finally surrendered in 1974.
It turns out that the Japanese government had repeatedly tried to inform Onoda that the war had ended. This culminated in 1952 when:
…letters and family pictures were dropped from an aircraft urging them to surrender, but the three soldiers concluded that this was a trick.
They had spent 7 years on the island, waiting. Should they have realized the war was over? Wouldn’t an enemy combat aircraft try to kill them, not drop papers on them?
The line between persistence and stubbornness is blurry.
For great founders, persistence is a core value. But can it go too far?
Sometimes, a startup’s target market isn’t developing. Does that mean it won’t materialize? Perhaps, although sometimes markets don’t develop until, all-of-a-sudden, they do. Persistence pays, right?
Not always. Sometimes, the startup charges forward, but the war is over. The market opportunity isn’t there and never develops.
How can founders avoid waiting on a market that won’t materialize? I think the best advice is to be vigilant against confirmation bias. When I’m hoping for something, it’s very tempting to interpret any evidence as confirmation for my hopes.
Every startup has a theory of their market opportunity. If they can build a certain product, they think they’ll see strong customer demand. In science, a theory must be supported by evidence. It also must be falsifiable.
So a good process is to ask: what will confirming evidence for the theory look like? The key early signal, I think, is desperation in potential customers. What does dis-confirming evidence look like? Often, the most potent dis-confirming evidence is lack of desperation. What constitutes a signal of desperation? What falls short? To avoid bias, it’s critical to specify this before a theory is tested. Writing down specific predictions ahead of time is a strong antidote to confirmation bias.
In SaaS, it shouldn’t take too long to gauge if there is desperation in a market. 2-4 quarters should provide a sufficient period to hone an idea, build wire frames or a prototype, and then collect evidence.
If I could go back to the late 1940s and give Onoda advice, I would ask: if the war is still going, shouldn’t there be more evidence of continued fighting? No bombs were being dropped. No battleships were firing ammo in the distance. The most potent evidence was in the lack of evidence.
Confirmation bias can turn just about anything into seemingly convincing evidence. I suspect Onoda wasn’t well-versed in decision-making theory, and thus succumbed to confirmation bias.
Fortunately for Onoda, after his surrender he went on to enjoy a prosperous life. He wrote a successful autobiography, which is now on my Kindle. Although my text search for “confirmation bias” yielded no results, I’ll keep on reading.