Should a high-value user see a high price? When I asked Jacob Rushfinn, he answered, "Sometimes, but I wouldn't get caught up in the price. Price is just one variable." For some of the highest-intent segments his team has seen, a lower price doubled conversion. What the user sees around the price matters as much as the price itself.
Jacob is the CEO and co-founder of Botsi, which trains a custom model for each subscription app client to decide which offer each new user sees. He started with beta testing and private clients last year and now works with apps doing more than $100 million in ARR. He also writes retention.blog and hosts the Price Power podcast. In this conversation we cover how to move from classic price testing to personalized offers, and what you can do today without a machine learning model.
The limits of price testing. Testing a baseline price, then your biggest markets, then higher and lower variants works, but going country by country is slow. Raising prices can also lift revenue up front while the total number of subscribers drops. Jacob explains why that pushed him toward personalizing the offer.
Signals before the paywall. Device data such as country, OS, device type and language. First-party data from quiz and onboarding answers. Acquisition source, down to the ad network, campaign and Apple Search Ads keyword. Predicted LTV is layered on top so the model optimizes for long-term revenue, not just conversion.
A check you can run today. The share of users who cancel in the first seven to 14 days after subscribing is highly predictive of renewals at one year. It needs no machine learning model.
Inside Apple's rules. Botsi does not generate price points on the fly. Teams pre-select a handful of products, Apple approves them, and five to ten products can combine into a large number of offers. One example is the same price, with the monthly plan behind a see all plans page and the annual plan shown up front.
What it does to your ads. Jacob says he has not seen a real negative impact on ad performance. When a team told him their ads stopped being profitable after a price test, his read was that half of their users had been shown a bad price.
Video Chapters:
00:00 Introduction
02:38 Jacob's method for price testing
05:19 Packaging: the monthly to annual ratio
06:52 Why one price for everyone stopped making sense
09:28 The signals behind a personalized offer
12:10 Layering in predicted LTV
12:55 Price is just one variable: the whole offer
15:12 A check you can run without machine learning
17:54 When a lower price doubles conversion for high-intent users
19:23 How much data you need to personalize offers
20:59 How AI changed the approach
26:28 How personalized offers work within Apple's rules
30:02 Is personalized pricing fair to users?
34:24 How personalized offers affect media buying
41:12 Where to find Jacob and his work
Topics covered:
Why price testing alone stops working as you scale
- The signals used to personalize an offer before the paywall
- Why a lower price doubled conversion for some high-intent users
- A no-machine-learning check: cancellations in the first 7 to 14 days
- How personalized offers work within Apple's rules
- How personalization affects media buying and ad performance
Learn more:
Episode page: https://mobileuseracquisitionshow.com/episode/how-to-personalize-paywalls-for-subscription-app-price-testing-with-jacob-rushfinn/
Jacob on LinkedIn: https://www.linkedin.com/in/jacob-rushfinn/
Botsi: https://www.botsi.com/
retention.blog: https://retention.blog
Price Power podcast: https://podcasts.apple.com/ng/podcast/price-power/id1844494362
Newsletter: https://intelligentartifice.kit.com
