We have a lot of choices when deciding when to offer discounts as incentives for our free users to upgrade to a paid product. We have had very impressive results from our discount campaigns in the past, so we know it works well, at least in our case! The problem is that we don't REALLY know if we're leaving a lot of money on the table and just doing it completely wrong. Valid testing is hard because we have a small number of buyers paying a relative high price.
Should we send them when sales are slow? Should we send them when sales peak, reducing friction among people CLOSE to buying already? Should we just work on modeling behavior and offer them on a case-by-case basis?
Do any best practices exist around this?
> Should we send them when sales are slow? Should we send them when sales peak, reducing friction among people CLOSE to buying already? Should we just work on modeling behavior and offer them on a case-by-case basis?
Why don't you flip the question around: Everything has a value and a price. I assume in this case these customers would buy eventually. What is that 'eventually' and what price are you wiling to bring forward those sales and have that money sooner rather than later?