another follow-up: how can you determine what's "good" for your product or not haha it seems like one of best applications of this type of analysis might help you prioritize your investment into product areas? thinking out loud here haha because obviously strava logging runs is going to be the top used feature but is 3 days enough or 4 days, etc
love it! think the hardest thing in Amplitude or similar is finding ways to "reduce noise" so some of those charts (eng matrix or pathfinder) are useful...that's why for some of these posts, i'm tempted to just revert back to google sheets or even some sort of notebook haha
anyways, one quick question - specifically, when it comes to a freemium app..let's say Strava
how would you recommend putting this INTO action when it comes to understanding which premium features drive the most usage or growth?
I see your point regarding "reduce noise", depending on how the analytics is instrumented it may be harder than easy.
Some of the quick things to act on:
For your top features by usage, you may consider paywalls, upselling, ADs if your strategy is to double on premium growth.
Your top 2-5 features (by retention) can fuel your winback campaigns (if churn is a concern).
Use top features by DAU and DAU/MAU usage for notification reminders, this proven to perform the "safest" (low unsubscribes)
Features with the lowest volume of usage and retention can be candidates for sunsetting.
If you notice a low-usage feature is "on the way" of the user path to the strong one that has the highest retention, you need to swap them. Features with the highest retention are predictive of growth. You want them to be easy accessible.
I am using daily retention and DAU in my analysis as an example that suits Strava. WAU, WAU/MAU, weekly and monthly retention might be more appropriate for other products.
Keep in mind, some features may be expected to be used occasionally, and others more frequently. That's why I watch DAU/MAU ratio in combination with AVG usage per user. DAU/MAU ratio shows what % of users interact with this feature every day in a given month. If this ratio stands out for a particular feature it's strong signal. For example, if it is very small but AVG per user value is high, it tells you that users use the feature a lot but not frequently. It can be expected, or can also be a discovery issue (when a feature is hidden in menu or else).
Question - when doing these types of analysis. Would you recommend doing them within the toold itself (ie Amplitude/Mixpanel), or would you recommend exporting the data, in a more granular view and doing the analysis somewhere else (excel, jupyter etc). What would you say are the pros and cons of both. Thanks!
another follow-up: how can you determine what's "good" for your product or not haha it seems like one of best applications of this type of analysis might help you prioritize your investment into product areas? thinking out loud here haha because obviously strava logging runs is going to be the top used feature but is 3 days enough or 4 days, etc
love it! think the hardest thing in Amplitude or similar is finding ways to "reduce noise" so some of those charts (eng matrix or pathfinder) are useful...that's why for some of these posts, i'm tempted to just revert back to google sheets or even some sort of notebook haha
anyways, one quick question - specifically, when it comes to a freemium app..let's say Strava
how would you recommend putting this INTO action when it comes to understanding which premium features drive the most usage or growth?
Thank you!
I see your point regarding "reduce noise", depending on how the analytics is instrumented it may be harder than easy.
Some of the quick things to act on:
For your top features by usage, you may consider paywalls, upselling, ADs if your strategy is to double on premium growth.
Your top 2-5 features (by retention) can fuel your winback campaigns (if churn is a concern).
Use top features by DAU and DAU/MAU usage for notification reminders, this proven to perform the "safest" (low unsubscribes)
Features with the lowest volume of usage and retention can be candidates for sunsetting.
If you notice a low-usage feature is "on the way" of the user path to the strong one that has the highest retention, you need to swap them. Features with the highest retention are predictive of growth. You want them to be easy accessible.
I am using daily retention and DAU in my analysis as an example that suits Strava. WAU, WAU/MAU, weekly and monthly retention might be more appropriate for other products.
Keep in mind, some features may be expected to be used occasionally, and others more frequently. That's why I watch DAU/MAU ratio in combination with AVG usage per user. DAU/MAU ratio shows what % of users interact with this feature every day in a given month. If this ratio stands out for a particular feature it's strong signal. For example, if it is very small but AVG per user value is high, it tells you that users use the feature a lot but not frequently. It can be expected, or can also be a discovery issue (when a feature is hidden in menu or else).
Question - when doing these types of analysis. Would you recommend doing them within the toold itself (ie Amplitude/Mixpanel), or would you recommend exporting the data, in a more granular view and doing the analysis somewhere else (excel, jupyter etc). What would you say are the pros and cons of both. Thanks!
Often you don't have a choice:
1) If you are lucky to have all behavioral data in the tool, and data makes sense, stick to the tool. It's faster and easier.
2) If you don't have it all there, then you need to map users between multiple tools, and once they are mapped, then do analysis.
I never was lucky with (1). That being said, (2) gives you more freedom for analysis.