Be Careful with Pricing Experiments

This is a horror story of how I shredded at least ~$50k or more. Some may know me as the guy who runs Candy Japan, but this is about another project many years ago.

It was New Year, during which I always spend a week with my wife's family, and also tend to have a lot of free time to do explore something new on the computer. Usually it's some side project, but that time I decided to finally put into practice all the wonderful SaaS recurring pricing talk I had been reading online.

I had this app running with a nice monthly revenue. It had slight ongoing server fees for each customer, but I had still priced it as a single sale of $39 (not randomly, tested different price points).

But I started wondering, if I changed it to a recurring model, could I perhaps be making a lot more? So that week at my in-laws' place I did the price change (grandfathering existing customers in of course) and pushed it: $9.95 / month.

I figured if it didn't work, well, I can just roll back to the old pricing. Of course I'd have to wait a few months to see how the churn is to be able to know whether it was the right choice.

It wasn't.

I was now making much less per customer, even taking into account their lifetime spend. I think it was a combination of making fewer sales at the fixed price, and the churn being high. In any case: less money overall.

So I go back and undo the change, but I could never get the app to recover, even after putting the old single price back again.

Perhaps I upset people who had been considering buying, or I accidentally changed something critical in the process of undoing the change, or maybe just the circumstances changed during the experiment.

Sales were never the same again. I don't have the exact numbers in front of me, but I recall the new "normal" being something like a third or less of what it used to be.

The lesson? If you play with pricing, be careful with changes (maybe don't push during lull times at your in-laws' place). Make sure to be able to roll back to EXACTLY how things were, and don't change anything but the price. A/B test the new pricing if possible. Good luck.