He looked at what we were doing and he said 'double your price'.
I laughed at him, and said no way, people will cancel in droves, if more than half cancel it'll be a net loss.
Sales went up.
We basically doubled our income overnight.
What happened to the existing users when you doubled the price? Did your cancellations go up? Or did you freeze the price for the old users?
Most chose to lock in, and a very small number of them still have that rate even 3 years later.
We have fewer members now but that is due to other reasons.
Cancellations for the locked in accounts went down because those people thought they were getting a very good rate and knew that if they canceled and would sign up again it would be at the new rate.
At the time our retention was 3 months, so within 3 months the larger portion of the accounts were being charged at the new rate.
The really interesting thing about the whole affair I think is that our retention became much better, and that's the hardest figure to tune in any online service.
If you give me your daily signups and your charge rate I can basically tell you what your annual turnover will be when I have the first week of statistics. It's scary how predictable that is.
Say n is your number of signups per day, your retention is 'industry average' (3 months or so for most consumer stuff, 12 months for b2b) and you charge 22.95, then for a consumer product you will make roughly (say 10 signups / day average across a month) 10 * 90 * 22.95 / month after 3 months of ramping up, for a b2b product you'll make 10 * 365 * 22.95 per month after 12 months of ramping up.
The retention is very important because it determines how far you can grow, a 1 month increase in retention in the above (b2c) scheme is 33% more turnover, but retention is the hardest of all the variables to get under control.
It is related to customer satisfaction, card expiry, product life cycle and so on, the signup rate is only affected by the quality of your marketing and your presentation.
edit:
When people launch a for-pay service and they see their first stats they usually think that the income is going to grow unbounded, but every product has a life-cycle and you really need to factor that in to get any realistic amounts.
This mistake is also present in the business plans of many start-ups, you have to assume some upper limit to the life cycle of a customer to get realistic figures.
It is usually a safe bet to plug the industry averages (3, respective 12 months for b2c or b2b) in, you may be a bit better but usually not much unless you have a very addictive/compelling product.
It is also good practice to over-deliver by a bit compared to your plans, if you have investors it keeps them happy.
Changing prices is always tricky, the above story tells you that my judgment on this subject is as faulty as can be so you probably should take my advice with a grain of salt (or a pound).
If I were to do something like this again I would do a 7 day a/b test and see how the signup rate is affected before taking the plunge. Even that is not simple to pull off because it would mean somehow isolating users from the two groups from each other long enough that you wouldn't influence the outcome of the test.
The big lesson for me from all this is that price elasticity is a non-linear entity.