Amplitude Analytics for Onboarding Tracking | Part 3 SaaS Analytics Stack

The third step in the Saas Stack Analytics process is onboarding. Ruben will be walking us through the onboarding metrics we should track and give us an overview of how we can see those in Amplitude Analytics.

đź”— Links:

SaaS Analytics Stacks Playlist

Amplitude Analytics:

🎓 Learn more from Measureschool:

🔀 GTM Copy Paste

🚀Looking to kick-start your data journey? Hire us:

đź“š Recommended Measure Books:

đź“· Gear we used to produce this video:


All right, let’s keep moving on to our third video in our SaaS analytics series. In the second video, we cover the signup and login so basically the acquisition steps. And in this video, we’re going to focus on the onboarding. So in step three here. Now, onboarding for a typical SaaS product typically means a series of steps after user signs up. So they create an account. And then they have to complete some kind of funnel to be able to use the product, right, this is where they might create a profile, they might upload photos, they might go to a product tour, it kind of varies. And then for companies, this might mean a completely self-designed, self-contained funnel, right. With special screens, special pages that if the user doesn’t complete, they can’t actually use the product. Or it might mean being guided through just the regular product through like tool tips or things like that, as a lot of different variation you can take. Nonetheless, in our analytics, we want to know the performance of this onboarding funnel, right.

So when we look at the events that we want to track here, we’re looking at the what we call onboarding events. So let’s imagine that you have a funnel, and the funnel has five steps, right.

And for the user to be considered onboarded they need to complete all the five steps, right, and they come right after, right. So we’ll have five events here. And under each event, we can have, you know, onboarding, step one, step two, step three, so on and we have any relevant properties under each event.

Then for the user attributes, it’s always helpful to track the onboarding completion that has the user onboarded successfully. This eventually becomes useful in things like notifications, where you can say, Okay, let me send a special kind of messages, or stop sending special kind of messages to anyone who has already onboarded, right. It’s common for SaaS products to use, let’s say email drips to complement any existing onboarding and funnel so you can stop those drips by having a just a flag, a true or false slide here.

And then the other thing that we might see here into the qualitative data will be something like user record this, right.

So if we are to watch users go to the funnel can be quite helpful. It can give you a sense of worries against stuck. Typically, you know, when we see a lot of issues and onboarding funnels it tends to come to expectations or the design that UX design of the funnel itself.

That is a user sees something and then not sure what that means, they’re not sure why they should do it, or why they should care, you know, why do you think the create profile right now, they can just do it later, things like that. So session recording should give you a sense of that because then you’ll be able to see if user is spending an unusual amount of time on a step that in your perspective as simple, just a couple of input fields, enter information, move on and the user gets stuck. So that would be actually quite helpful, right.

In this case, for data tracking this actually done on server side, and this one should be the recommendation.

So in the acquisition side, we saw a lot of things that needed develop client side just to be able to manage at this properly. In this case, we can adjust our moving server side. And that can be quite helpful especially if you have a cross-platform product that you have a product of mobile web, iOS, Android, so on.

Trying to implement everything always client side will become a little tedious, especially if the same action can happen on any platform. So this is where we can start to build some expertise around tracking events for the server side. If you can’t do that if client side is your recommended approach, where you want to do, that’s fine. Just be aware of that potential limitation.

In our tool stack is similar to what we saw before, we have segment, of course, we still want to funnel all the data through segment. And then the main tool that we use for analysis here will actually be amplitude, right, which we saw a little bit before the acquisition video.

So here’s a segment again I won’t cover this in detail cause I covered this quite a bit during the acquisition. But same thing today will flow here what is from a client side of JavaScript, or a new source, like a, like a server-side source, and it will flow down into all the different tools. And we specifically care about amplitude in this case as our behavior analysis tool.

So jumping into amplitude, really, the report here that we care about is the funnel analysis report.

And it looks like this before we show you how it’s built. So we have the series of steps. And then we can see the drop off from step to step right. Now, in this case, our steps, right, user starts the sign-up, they complete the signup, we saw this event quite a bit during the acquisition video, right, as a pretty critical event.

And then we have all the onboarding events. And you can see here, we like to prefix events with the word onboarding. And that just makes it easier to find, it organizes nicely, actually, a lot of tools will organize events alphabetically, so you can find them on one section. So this is just a minor detail that becomes quite helpful when you start working through data. So we have what the you know what the steps are. And, of course, these are relevant to the product itself, that we’re looking at this job search product.

And then the last event here is profile approved, right? So we have 3, 6, 7, 7 onboarding events in this case, right?

Something is about amplitudes that we can actually look at the events in any order or is this specific order, right. I have seen probably quite a bit of issues when you actually have to force an order, right, when it’s okay, users have to complete this one, and then that one, and then this one that one. Basically, what that means is, if for some reason, you have an optional event. Let’s imagine that you’re asking the users to enter a credit card, but optional, right, you have an event called onboarding enter credit card, but they can skip it, then all of a sudden now the steps down below might actually they will be counted, if you’re forcing it to be very specific order, right, because there it’s looking for that flow, and if any event skipped, then your numbers will be slightly off. So any orders can be helpful because I know users can technically go back and redo events, and they’ll still be counted as a whole, okay.

In the funnel itself, this is what it would look like, right. So we see a number then we see the drop-off, you know, the biggest drop of course when the user starts to sign up, but they’ll complete it. And then we go to the onboarding itself, in this case, is onboarding, this is actually quite, quite simple, quite well done. You can see there’s not a big drop offs, except maybe from the last step, or the second last step to the very last step, right, to get the profile approved, right. And, of course, just before we can click into things, and we can see, you know, the 57 users who complete it. We can also see the 26 users who didn’t complete it. And we can see some of the other things. Some of the interesting actually something like user pass.

So we can see, you know, for users who complete a profile approved, what do they do after, right, what were some of the path it take after. And that path explosion can be can start to help us understand maybe a better onboarding flow, right, which is typically in a typical one of the goals you start with, with something that you know, works, and they want to move on to something that might be a much better optimize, Of course, before here will have conversions by dates, right.

We have average time in between steps, right, and so on. So we’re able to segment things here.

That’s really it, we can now see conversion over time. So we can see the overall conversion actually of the funnel over time. So if you see if that’s, that’s getting better or getting worse over the last 30 days, 90 days, 6 months or so on. And then time to convert is typically also a helpful report, we like to see, you know, are you just taking a long time to go through this funnel, or is this something happening with a few minutes, and this is something that you’d likely want to play around a little bit of dates, you might have outliers are really bring it down your averages status. You might find 95% of all your users complete the funnel in a few minutes.

But then you have 5% of very small amount of users who take 25 days, right. That is they started and they come back 20 days later. So you want to play around here a little bit of time to get a sense of how long it’s taken to, to convert, right.

And we gonna foresee the conversion time between specific steps too.

So we have a lot flexibility here. And of course, just like before we can segment this even further, we might want to look at only users on iOS or only on Android or web or maybe from a specific country and so on. So all our segmentation options that we cover a little bit in the first video was are still available here and this is it right. This is this what onboarding is for our series and what the goal is in terms of our data.

Hey, Julian here from measureschool. If you want to follow along with the next video in this series then we have it linked in right over there and if you want to learn more about the data-driven way of digital marketing then definitely subscribe to our channel right over. Now, my name is Julian. See you on the next one.


Leave a Comment