SaaS Metrics for Hackers is a 3-part article series that helps bootstrappers with technical background getting started with SaaS metrics and finances.

The first part explained the financial basics and the second part explained how SaaS metrics and the SaaS main funnel help you debug the finances.

This last part explains how to use funnels & user events to debug the SaaS metrics. And as an extra, it will explain the principles of using events proactively to increase your bottom line.

Debugging SaaS metrics

The principle of debugging SaaS metrics is simple - track user events in trouble area to find out what people actually did and get some ideas on what might cause the troubles.

Let’s assume that while you analyzed the SaaS main funnel, you found out that there are problems in getting trial users to make a purchase.

The first questions to ask are: “What tasks did the people in successful trials perform?” and “Do these events form a funnel?”

When we are talking about a trial, the key events (/actions/tasks) usually form a funnel.

In order to make a purchase, the user must:

  1. Sign up for trial
  2. Activate the trial by inputting data, installing tracking codes, etc.
  3. Get value from the trial

The actual steps depend on what type of SaaS you are running. But when you find the steps, you’ll track the events similarly than we tracked the main funnel in previous part.

When you have the data, there are many ways to visualize the funnel conversion rates. Below is the format that both Kissmetrics and Mixpanel uses.

Trial to paid conversion funnel Now you can easily see where you lose the most people.

Remember that you may not find all the troubleshooting information from the funnel events.

Sometimes, for example in cases where your product suffers from technical or usability problems, the differentiating factor may be something like “all the people who purchased contacted support at least once”.

What if the events don’t form a funnel?

If the events don’t form a funnel, you’ll need to use the event data to create profiles of your customers by their event usage and success.

And then when you do have the data, you may be able to find funnels too - you just couldn’t guess them beforehand. Some funnel analytics apps do support this and let you analyze funnels afterwards.

A good example of problems that can’t be debugged with funnels are churn problems. But when you are starting out your customer base is so small that the churn problems are often not visible/relevant. At that point it’s usually more productive to just ask people why they canceled.

When your customer base is small, all the things that you can get from user event data you can also get by talking to your customers. It’s not as fun as solving things with technology, but deep understanding of your customers is essential for succeeding.

So don’t start building event tracking systems too early. Or do track the events, but don’t put too much weight into the results.

Reactive vs proactive metrics usage

Until now, we have drilled down in your data and concentrated on problems and how to debug them. And that’s how you get started.

But debugging and fixing bugs isn’t enough. People in the TechCrunch success stories didn’t just “improve the conversions and reduce the churn”. Nope. It’s always something like “we found out that hosts with professional photography are booked 2.5 times more frequently than others”.

Proactive creative work is what makes the success stories

Building a SaaS business is proactive work. You’ll want to stay in that proactive mode also when running your business.

User events and detailed funnels give you information on what your users & customers did. That’s a goldmine of information and you don’t want to use it just for debugging. You’ll also want to use it to find opportunities for your business.

What events should I start with?

When you start out and you are not debugging a problem, track just a couple of events & funnels:

  1. Key value event 2a. The marketing funnel (check the conversion rate first) 2b. The trial funnel (check the trial-to-paid rate first)
  2. Feature usage events

After that, you can add the funnels & events as you need them.

Kissmetrics blog has a nice article about tracking the marketing funnel with Google Analytics, so I’ll be skipping it here.

Track the events that produce the most value

You can get started by tracking a single event for finding the customers that get the most value from your product.

Ask yourself: “What’s the single action/event that brings the most concrete value to your customer?”

Track that.

If you are running an app that helps freelancers to make proposals, the key event could be getting a proposal accepted. Or any event where your customer actually gets paid.

If you are running a time tracking app, the key event could be logging billable hours.

If you are running a freemium training platform, the key event could be a combination of video watching and pro subscription activation. You’ll want to know what videos people watched before they felt that they got enough value to sign up for a pro subscription.

Freemium is always trickier to analyze, because you can’t assume that people producing the biggest amount of value events would feel that they get the most value. Yeah. People are strange. People who get free stuff are even stranger.

Find your best customers

When you have the events, check out that the customers who produce the most key value events also are the customers with biggest customer life-time value.

You aren’t calculating the full CLTV yet, but you are marking down the totals of what each customer has paid you this far (total contract value, TCV). Use that data.

If they don’t match, it means you have failed to price on value. It’s always worth trying to fix that.

Reminder: If you go D-I-Y, don’t put the event data into your production DB. And always put all data collection and analysis work into background threads to keep your app responsive.

Profile your best customers

Now you have the information about customers who get the most value from your product. The next step is to find out what makes them different from your other customers.

If you have the feature usage event data too, part of this analysis can be done automatically, like I’ll be doing with Virtual Financial Officer and like Retention Factory does.

But part of it is hard to to automate.

Let’s take an example. The AirBnB feature that made some of the hosts popular was that they had nice photos of their apartment. That’s not something that you could easily find automatically.

So even when you have an automatic system at place, it’s worth to manually profile your customers every now and then. Just find 5-10 winner customers and 5-10 loser customers, open their profiles and make an analysis of what you see. Who are these customers? Can you think of anything that make winners stand out from losers? Dive into the DB if you need to.

Use the findings to advance your business

After the analysis you’ll have a profile of your best and your worst customers and you can start working on that data.

Can you offer something to losers that makes them more like the winners?

Services that help them through obstacles? Education on features that winners use? Education on the type of things that winners do?

Also think of the opposite - how can you serve the winners even better? Could you offer extra services for them? How can you acquire more of them?

This said, we’ve come to the end of the beginner metrics series. If you have any feedback, I’d love to hear it. There’s a comment box somewhere below this post.