Per product NPS survey

requests-for-comments/2023-10-25-per-product-nps.md

@@ -0,0 +1,61 @@
+Per-product NPS
+
+We would like to run permanent per-product NPS surveys in the PostHog app, for the following paid-for products:
+- Product Analytics
+- Feature Flags
+- A/B Testing
+- Session Replay
+- Surveys
+
+## Why?
+
+- Trend of customer satisfaction, we see how customer satisfaction develops over time. If we find that user satisfaction for a certain product is low or trending down, we can react
+- By asking follow-up questions to explain the score, we can learn which areas of the product in particular need improvement
+- By asking on a per-product basis, we gain more actionable feedback than having a blanket survey for general PostHog usage
+- The customer success team has an additional data point for individual customers if they are happy or not and can react quickly
+- NPS is an industry-known benchmark and allows us to compare ourselves to other products. The scale itself allows for more granularity than the [PMF survey](https://posthog.com/founders/measure-product-market-fit#indicator-2-pmf-survey), which is similarly known, but only offers a 3-point scale
+
+## Setup
+
+- 2-step survey, NPS and open-ended question for feedback
+- Link to live survey for feature flags: https://app.posthog.com/surveys/018bcec8-6cf5-0000-c724-a51a86a4e8b1
+- Question 1:
+
+    <img width="359" alt="Screenshot 2023-11-16 at 09 04 50" src="https://github.com/PostHog/meta/assets/14750837/5640a505-5898-4632-a6bc-6cbc374190fe">
+
+- Question 2:
+
+<img width="363" alt="Screenshot 2023-11-16 at 09 05 12" src="https://github.com/PostHog/meta/assets/14750837/619a9bc6-6cb6-4f30-8de3-4149ed1fe1c9">
+
+- Thank you: Get a free PostHog sticker:
+
+<img width="366" alt="Screenshot 2023-11-16 at 09 05 36" src="https://github.com/PostHog/meta/assets/14750837/d7b88484-ae41-45c6-8231-0d69f71a9671">
+
+## Release conditions
+
+- Users on any feature flags page (app & eu). I am using the following regex: `https://(app|eu)\.posthog\.com/feature_flags`
+- Users who match the 'product usage' cohort, see definitions below, excluding PostHog team
+- Users who have signed up more than 14 days ago
+- Don’t display to users who have seen another survey in the last 60 days (set as survey setting)
+- Don't display to users who have dismissed or answered this survey (set as property, because the survey setting is not working 100%)
+- Roll out to a small number of users, e.g. 10%, so that we can track changes over time, rather than asking every user right now. (If we would roll this out to 100%, there is a risk that no user will see another survey for a while, if we keep the ‘survey wait period’ controls
+
+## Appendix: How do we define 'product usage'?
+
+In [this document](https://docs.google.com/document/d/16Z6m4kdgxui53SuSq-PeZ30lYfEsyict3u2fPm2vlsE/edit#heading=h.9kfq6mtxomjq), there is a proposal for what counts as an activated organisation for each product. I am slightly adjusting those definitions to define 'product usage' **per user**.
+
+Product Analytics
+- Completed event `insight created` AND event `insight analyzed` in the last 90 days: [Cohort](https://app.posthog.com/cohorts/48711)
+
+Session Replay
+- Completed event `recording analyzed` at least 5 times in the last 90 days: [Cohort](https://app.posthog.com/cohorts/48712)
+
+Feature Flags
+- Completed action `clicked-save-feature-flag (existing flags)` OR action `clicked-save-feature-flag (new flags)`
+ in the last 90 days: [Cohort](https://app.posthog.com/cohorts/48720)
+
+Experiments
+- Completed event `experiment created` OR event `experiment launched` in the last 90 days: [Cohort](https://app.posthog.com/cohorts/48729)
+
+Surveys
+- Completed event `survey created` OR event `survey launched` in the last 90 days [Cohort](https://app.posthog.com/cohorts/48730)