Hey everyone! Welcome to part 3 of my Ultimate Guide to LTV where I go deep on all things LTV. In case you missed it, last week I shared how to increase LTV with 18 different tactics and this week I’ll share how to identify high value users.

*For paid subscribers only

Let’s go!

One of my favorite parts of writing my newsletter is breaking down supposedly complex problems and debunking myths. Today’s article is exactly that. It might seem like a challenging task to identify high value users but it’s much easier than you think. I’ll show you a step by step process that will help you identify and better understand your users so you can go get more of them. Let’s go!

When you hear segmentation, your body does 1 of 2 things: flight or fight.

How come “give it a hug” is never one of the options? That’s because we’ve been conditioned to hear “segmentation” and think it’s going to be a data sciency thing that’s complicated or useless. My approach to finding high value users is neither of those. In fact, it’s so easy a caveman can do it.

In part 1, I shared how to calculate LTV. Now, you take the assumptions you’ve made there (including costs and margins, etc.) and calculate that on a per user level:

Bonus: what you really want to do is do it on a cohort / user level so you can account for adjusting marketing channel mix and fluctuations in marketing spend. If you understand how to do this then please calculate at a cohort level. Otherwise, the version I have above will generally work.

Now, sort users from highest LTV to lowest. Then draw a line under the top 10% of users. Anyone below that line gets “thrown away” aka ignore that data. Take this top 10% and let’s move on to the next step…

In the step above, you could use 20%, 25%, or whatever % you want. It’s up to you. I like 10% because it’s a nice round number but using 20% (which is where the pareto rule usually comes in) is totally acceptable. Using 50% would not make sense because then it just becomes who is above and below the average.

Anyway, now you have a list of users. Great! Invite this 10% to your wedding as they’ll likely give the best gifts as they like to spend the most. Jk. What you want to do is analyze this list of users to better understand them.

We want to do that across 2 types of analyses

For each of the attributes I list below, you simply plot a histogram to see the frequency of that attribute.

I prefer histograms over pie charts because they’re easier to visualize the magnitude of differences.

Demographic data doesn’t necessarily mean demographic in the classic sense of race, age, gender etc. That’s mainly because most companies don’t have that data. Instead we’ll use demographic to talk about attributes about users that are important to know but don’t represent how they interact with your product.

As I mentioned, take your top 10% and analyze through histograms and other simple plots to understand where the majority of them look like:

Basic Demographics

Tech/Platform Info

Acquisition source

Behavioral data is 1st party data that you capture from when your users use your product. It’s the richest source of data and usually where the nuggets of gold are.

You’ll do the same thing with behavioral data which is plot histograms for each attribute below to better understand where patterns of high LTV users come from.

Usage Patterns

Transaction Behavior

Engagement Indicators

Other

As you can see the list could be endless but what you’re trying to better understand is:

Once you have all these histograms, take some time to go through all of the data and see if you can come up with a picture of your best users. Is there just one type where everyone looks the same or are there 2 ways to become a high LTV user? What patterns can you figure out.

Let’s unpack one example.

Example

You look at the histogram for iOS vs Android. Let’s say that the histogram suggest that 85% of high LTV users use iOS instead of Android. Well, that’s a pretty clear hypothesis that I should focus more on iOS app campaigns vs Android app campaigns.

But you should also take a step back and think about why IOS users are higher ltv. Is it that your iOS app is better? Or is there something about those users that’s different.

You could do something similar when analyzing campaign types (Paid vs Organic etc.). Even though attribution is not perfect, it’s still a good signal to understand where to make investments.

Each of the attributes above should lead to a new experiment hypothesis which in itself makes this exercise extremely valuable.

That’s it. That’s how you identify high LTV users and their behaviors. Not everything useful has to be complicated and this whole analysis would take no more than 2 days to put together .

Even if you think this won’t be as useful as you think, do the analysis and you’ll be surprised how many patterns you uncover. Then, how you act on them is up to you but luckily for you, I’ll share how I would use them next week 🙂

Have a great week!

Share with your colleagues or on your LinkedIn . It helps the newsletter tremendously and is much appreciated!

Here it is: Ultimate Guide to LTV Part 2: How to increase LTV