Loss Aversion
Loss aversion is the idea that losses feel 2-3x more painful than equivalent gains feel good. Kahneman and Tversky showed that people reject a 50/50 bet to win or lose $100, even though the EV is neutral. Thaler’s mug experiment showed that sellers demanded twice as much to part with a mug as buyers would pay, demonstrating that ownership inflates perceived value. This principle is a strong explanation for why consumers resist change.
The 9x Effect
Gourville calls this mismatch the 9x effect. Consumers overvalue existing products by a factor of 3, while companies overvalue their innovations by another factor of 3, which creates a 9x perception gap. Bridging that gap requires leveraging human psychology.
Strategies for PMs
To overcome this resistance, leverage four strategies to drive adoption:
- Make products 10x better. Incremental improvements aren’t enough. Grove’s 10x rule: only when a product is dramatically better (ie MRI scans over X-rays) will users tolerate change.
- Eliminate the old. Remove the incumbent. Canada succeeded in switching from paper dollars to coins only after withdrawing the bills. Otherwise, inertia wins.
- Seek out the unendowed. Target consumers not yet attached to the old system. Burton thrived by appealing to non-skiers.
- Make behaviorally compatible products. Reduce perceived loss by designing products that fit existing habits. Toyota’s Prius succeeded because it drove like a normal car.
I think electric vehicles are a classic illustration of these principles. Though invented in 1910, they failed for decades because the market wasn’t ready. Only when modern EVs felt familiar (fast, reliable, easy to charge) did consumers finally perceive the gains as outweighing the losses.
Self-driving cars example from my Waymo ride the other day
Autonomous vehicle adoption follows a similar playbook as the four strategies above, especially leaning on the last one. I took a couple Waymos the other day (it never gets old!) and I noticed that their UX simulates habits that ridehailers are already familiar with– internal configuration of the Jaguar mirrors a normal taxi even though in theory you could totally re-imagine it from scratch considering you don’t need a driver’s seat anymore (the seats could face each other! It could be a ring! etc.). Almost no one gets in the driver’s seat of the Waymo even though you can, probably out of sheer habit, because that’s where the Uber driver sits.
Example from WindBorne
At WindBorne (the startup where I work), our constellation of stratospheric balloons is the largest in the world and offers coverage of Earth that no one else provides – we have observations over the 85% of the atmosphere that otherwise goes un-observed. That means that we’re essentially doing points 1 through 3 – (1) we’re definitely >10x better than the alternative of 0 data; (2) there’s no precedent; and (3) people aren’t attached to the old system– rather, they’re already desperately seeking solutions. Finally, on point 4, we’ve taken care to make our AI forecast outputs compatible with processes from our customers (USAF, Navy, NOAA, traders) for minimal friction.
