The two most costly for me were both email bugs.<p>One defaulted people's email preferences to off when they signed up. The other was a poorly chosen subject line that caused all the wrong people to read one of the most important emails that company had.<p>I can't give out an exact dollar figure for either. But both were in the 7-8 digit range. And both took a surprisingly long time to track down because everything <i>seemed</i> fine. And so the problem sat out there. And even once we knew that there was a problem, the problem was only discovered by accident. (But having been through it, twice, I now know that if business just starts drifting downwards, double check what is happening with email. Again.)<p>By contrast the obvious disasters that I have seen have cost much less. Surprisingly so. The database crashed without a proper backup taken in the last 2 months? OK, we're out of business for a week and a half while we replay all of the logs, but then the email blast we sent saying we were back up attracted more than enough business to make up for the time lost.<p>And the toughest bugs? Surprisingly cheap. It took literally months to track down why we had to do a careful dance bringing up Apache or else nothing worked. Yet despite months of grief the actual cost in lost business was a few hours of downtime at off hours. I'd be surprised if we lost 5 figures on that. And the memorable one where you couldn't fill in a particular form if your screen was the wrong size? I'm pretty sure we lost $0 on that. (The bug, for the curious, was an invisible div that was in front of the window. If the window geometry placed it over the form, it became rather hard to fill out said form.)<p>It is a hard lesson that what developers care about and what businesses care about are only tangentially related.