Conflict of interest note: Windscribe is commercial open source VPN service provider based in Canada.
Why is ProtonVPN lying about running price sensitivity testing on their customers?
Their pricing page gives you one of two prices at random. This is A/B testing, but instead they're saying the price difference is due to remnants of a sale.
Their own code says otherwise.
Price sensitivity testing is when a company quotes different people different prices for the same product to see if sales hold steady at the higher price. Not to be confused with adaptive or dynamic pricing which uses your personal data to adjust prices to you.
People on ProtonVPN's subreddit noticed they were being quoted different prices for the same thing. ProtonVPN's General Manager replied that the differences were due to remnants of a recent sale that wasn't "universally refreshed".
This was a lie, and we have proof.
That proof is their own code sorting every visitor into a price variant. It's labeled right in the HTML tag:
ab-test:VpnHpAndPricingVpnPlusSensibility300726
Which in plain English is:
VPN Homepage and Pricing for the VPN Plus plan, Sensibility test tagged 30/07/26.
The price continues to flip-flop as of writing this. Expired sales don't do this.
Something like a Cloudflare-cached page doesn't add price sensitivity A/B tests to your HTML code.
Proton, if you want to run A/B tests, at least don't lie about it to your users.
I don't think it's good. To me, it reads as a red flag, an incongruity between different approaches to marketing and the service.
If there's an incongruity with pricing, I might think, perhaps there's an incongruity to the service itself versus how it's being advertised. (And in the case of Proton, this was the case).
If a company handles your unencrypted emails, trust is paramount, and breaking that trust can unravel years of goodwill.