Case Study: How Props Used SelkoDialog to Uncover Hidden Insights from Churned Users Case study of Props rewards app using SelkoDialog AI interview agent to uncover valuable insights from churned users

The full case study of how Props app used SelkoDialog AI user interview agent to uncover the truth behind churn:

www.selkodialog.com/blog/case-St...

We used the SelkoDialog AI agent to interview churned users of Props app.

What we learned:
– Most didn’t mean to churn, they just lost the habit
– Push notifications were off
– Small disruptions = long-term drop-off

These insights reshaped our roadmap 💪

📖 Full case study in comments 👇

SelkoDialog: AI Agent for Mass-Scale User Interviews AI agent for mass-scale user interviews for busy founders and product people!

Data tells you what is happening, not why.

That’s why we interview users.

But we can only interview so many, so we have to prioritize ruthlessly.

What if an AI agent could interview the rest?

That’s why I built SelkoDialog: an AI agent for user interviews.

selkodialog.com?utm_source=b...

Yeah, I cannot imaging how a non-developer could ever build and maintain a real production app with vibe coding

Today I dragged myself to the gym despite 0 desire 😩

And just like every other time, I never afterwards regretted going to the gym 💪

Cursor suggesting the strangest things when I am vibe coding

It's unable to figure out how to query for a specific row in the DB, so it suggests seeding the database with junk data 😂

Vibe coding is fun but a bit dangerous.

Cursor suggesting the strangest things when I am vibe coding

It's unable to figure out how to query for a specific row in the DB, so it suggests seeding the database with junk data 😂

Vibe coding is fun but a bit dangerous.

Vibe coding your infrastructure 😬

Opportunity or threat?

Vibe coding your infrastructure 😬

Opportunity or threat?

You are not really vibe coding unless you accept every code block blindly

If it does not work, just vibe code yourself out of the mess

You are not really vibe coding unless you accept every code block blindly

If it does not work, just vibe code yourself out of the mess

Is it a sign of an old-school developer to feel restless when the last line of code lacks a line break?

Is it a sign of an old-school developer to feel restless when the last line of code lacks a line break?

Vibe coding does not necessarily make you 10x more effective... but definitely 10x lazyer 😎

Vibe coding does not necessarily make you 10x more effective... but definitely 10x lazyer 😎

PMs and founders: Are you measuring your Feature Usage Rate? 📊

This helps you understand the adoption and usefulness of different features.

It is often calculated as the number of unique users who used a specific feature within some time interval divided by the total number of users

🧵👇

5/ Then use this metric when doing a feature audit. 🔎

Investigate why features are not used. Is it lack of awareness or poor understanding? Can this be improved? Or is the feature not useful at all?

Consider removing unused features, especially when they are not really useful for the users.

4/

3️⃣ Sometimes usage of a feature is only meaningful if the feature is used multiple times.

💡 You need to define your metric so that it is meaningful for your product.

3/

2️⃣ Some divide by only active users, which may be more meaningful when your product has changed a lot and you have lots of old churned users. Sometimes it is convenient to measure this only for users who logged in within the measurement time interval.

2/ There are some variants of this metric.

For example:

1️⃣ Some calculate this for an eternal time interval, in which case the metric is more about users discovering the feature

PMs and founders: Are you measuring your Feature Usage Rate? 📊

This helps you understand the adoption and usefulness of different features.

It is often calculated as the number of unique users who used a specific feature within some time interval divided by the total number of users

🧵👇

Viral Cycle Time is the average duration from a new customer's signup to when their invitees sign up.

This process typically involves multiple steps, including the customer learning to use the app, realizing its value, and then reaching the point where they start inviting others.

Viral Cycle Time is the average duration from a new customer's signup to when their invitees sign up.

This process typically involves multiple steps, including the customer learning to use the app, realizing its value, and then reaching the point where they start inviting others.

Are small SaaS founders squeezed in the middle of the sandwich?

Going niche has often been the key to success. AI brings down development costs, so all of a sudden many new niche products start making economically sense to build.

This is all good right?

Well, ...

So we face competition on both ends and get squeezed in the middle. Or do we?

How should we adapt? Where do you think the best opportunities are?

AI also lowers costs for your customers to build solutions in-house. Maybe they no longer need a SaaS product for every small problem they have?

On the other side of the sandwich, AI makes it easier for your customers to configure more complex solutions from big incumbents to their specific needs.

Are small SaaS founders squeezed in the middle of the sandwich?

Going niche has often been the key to success. AI brings down development costs, so all of a sudden many new niche products start making economically sense to build.

This is all good right?

Well, ...

What is Viral Coefficient?

Viral Coefficient measures the number of new customers an existing one successfully invites into the app.

A coefficient above 1 leads to exponential growth, as each customer brings more than one additional new customer.

This leads to a rapid growth of the customer base.

What is Viral Coefficient?

Viral Coefficient measures the number of new customers an existing one successfully invites into the app.

A coefficient above 1 leads to exponential growth, as each customer brings more than one additional new customer.

This leads to a rapid growth of the customer base.

Product metric you need to know: Time to Upgrade to Paid

In all simplicity, this is the time it takes after signup for a user to upgrade to a paid version. However, there are some practical challenges to calculating this.

The bare minimum you have to do is give an equal amount of time for all