[mwexler]

While I admit the form reset is annoying, I don't see where the author shows any evidence that this usage pattern happens enough to make a difference.

Sure, when N gets large, every little bit helps. But if that behavior pattern only happens rarely, then changing it is unlikely to make much impact.

For all we know, their testing might have shown that folks who are planning trips between 2 options (flight to Vegas, no hotel needed, vs. flight to LA, hotel needed) wanted the form to reset so they didn't have to keep swiping and typing over leftover text in the fields. The frequency of usage patterns, weighted by their value to the business, would be a useful guide.

I'm not saying that the recommended change is the right or wrong thing to do, but I am saying that some data would really help back up any claims over size of impact. Otherwise, it's a good finding ruined by a totally exaggerated headline.