Newsletter: A crash course on activation metrics #10414
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
There was a problem hiding this comment.
I think this is a banger, dude. That last section seems very concrete and actionable to me. At this point you know a lot about this so your standard is high but, ime, most folks are barely starting this journey and that's already plenty of meat that leaves them better than they started.
|
|
||
| On top of this, having a clear activation rate makes it easier to have a clear picture of other metrics down the funnel. For example, a poor referral rate could just be a product of not enough users being activated and getting there. A larger sample size (more activated users) lets you do a proper analysis of the down funnel metrics. | ||
|
|
||
| ## What makes activation tricky? |
There was a problem hiding this comment.
Lots to like. Just high-level feedback for now:
-
I think we should try something less direct with the headline. Headlines around "activation" didn't poll well in the past, I suspect because it doesn't sound super interesting. I'm thinking something like "The metric product engineers should obsess over", but pithier.
-
I really like the opening line of the intro, but the second kind of deflates it. You could probably just say "Let me explain why" and leave it at that.
-
Nit: It feels weird having illustrations with different styles. Would be good if we could make those consistent.
-
Another nit: For the main H2 headings, you're flipping between statements and questions – e.g. "Why you should care about activation metrics" followed by "What makes activation tricky?". Pick one way and use the same style throughout.
-
I think the volume of images is ok, but can decide when it's in Substack.
-
The ending feels a bit abrupt. I think this is one of these cases where we should end the article by pointing people to more places to learn from / case studies, just something to flesh it out. If we could come up with one or two more examples like the Supabase one, that would be ace. Maybe someone could point you to a friendly PostHog user or two to add something here?
-
Does it matter that we don't explain what it means first? You do this in the 'what makes activation tricky?' section, obviously. Maybe we should have something in the intro section?
There was a problem hiding this comment.
Added some more suggestions.
In hindsight, I'm kind of warming to your original idea for the headline. Maybe we could pitch as a series and add a poll asking what over crash courses people want? "Crash courses" might not be the right name exactly, but we can throw around some alternatives.
| In your journey to become a world-class product engineer, there is one metric that stands out in its early relevance to both product and business decisions: **activation**. | ||
|
|
||
| Let me explain why. | ||
|
|
||
| ## The mysterious definition of activation | ||
|
|
||
| Your journey to understand activation starts with a definition, but when you look around you'll find a bunch of different advice. Some say it's your "aha" moment, others, when you show your product's value, and others, when you make your users happy, but none of these are definitive. | ||
|
|
||
| So why isn't there a simple, shared definition? | ||
|
|
||
| 1. **Activation is unique to your product.** There's no standard way to define activation (this is also why benchmarks suck).[^1] | ||
|
|
||
| 2. **It could be many things.** There are likely many things you want users to do in your product: complete onboarding, upload a file, share a link, watch a video. Activation could mean doing any of these (or a combination of them). | ||
|
|
||
| 3. **Activating may mean doing something multiple times.** Adding to the complexity is that someone might need to do something multiple times in order to get it. For example, to activate into our session replay product means analyzing not one but five replays. | ||
|
|
||
|  |
There was a problem hiding this comment.
I wonder if this should just be the intro and we lose the heading in the middle?
Also, you've got a repetition of "your journey" here.
| 1. Look at activation's impact on retention beyond 3 months. | ||
| 2. Compare revenue for activated vs non-activated users. | ||
| 3. Finding ways to convert activated users into referrers. |
There was a problem hiding this comment.
Maybe I'm missing something, but surely non-activated users would just not generate revenue? I think there are probably better suggestions to make here, like refining you ICP based on how you activated users are?
In fact, we should probably reference ICP a bit more earlier on – i.e. activation is an important part of narrowing down your ICP.
There was a problem hiding this comment.
I don't fully understand this point, I don't really know how activation can help define your ICP, it is more like vice versa (ICP -> activation), but even that I feel weakly about.
|
|
||
|  | ||
|
|
||
| If you're stuck, Lenny's post on activation metrics includes [119 different ways](https://www.lennysnewsletter.com/i/77646597/what-are-the-most-common-ways-to-increase-activation) to try to increase activation. |
There was a problem hiding this comment.
This still feels like a slightly unsatisfying ending.
|
|
||
| With our potential activation event groups, we write a query to test how they correlate with retention: | ||
|
|
||
| 1. Get companies that start with our product broken down by month. We use group analytics for this, but you could just do users too. |
There was a problem hiding this comment.
not everyone will know what group analytics means
|
|
||
| 1. Get companies that start with our product broken down by month. We use group analytics for this, but you could just do users too. | ||
|
|
||
| 2. Filter for companies who completed the activation event group in the required time. We require companies to activate within 30 days of signing up. |
There was a problem hiding this comment.
Turns out we use a mix, so adding a comment to explain
|
|
||
| On top of this, having a clear activation rate makes it easier to have a clear picture of other metrics down the funnel. For example, a poor referral rate could just be a product of not enough users being activated and getting there. A larger sample size (more activated users) lets you do a proper analysis of the down funnel metrics. | ||
|
|
||
| ## How we find activation metrics at PostHog |
There was a problem hiding this comment.
This feels like the most important part of the article, so I think it needs a bit more to make it land.
-
More examples? We could just add more examples of our activation metrics for a few more of our products. I think this would be useful for emphasising how specific and different they are – i.e. it's not just "user did this one action".
-
Add some generic examples based on common app-types? This would help people apply the logic to something they're familiar with / working on.
There was a problem hiding this comment.
Added examples of a bunch we found
I feel like adding generic examples is SOOO generic in this case and not useful. Part of the reason the Lenny article isn't a banger is because it is so generic trying to appeal to everyone.
|
@ivanagas do you know what the main image for this will be, yet? |
|
What if we use this one?
|
Changes
Feedback appreciated:
Checklist
Article checklist