Getting Started With Analytics

The Metrics You SHOULD NOT Use

Analytics systems make a delightful amount of data available quickly. Unfortunately, that may lead you astray and burn lots of your time as the owner of a small software business.

Here are some common issues I’ve seen among communities like Indie Hackers, and while onboarding users in my work at PlayStation, and how to handle them so they give you the insight you really need.

Don’t look at Total Visitors or Total Visits. Split them into New vs Customers Instead.

Let’s start with the biggest, highest-level, most visible, most KPI-able metric that you could see: your total visitors (or total sessions, for that matter).

Consider that your business gets some customers and starts making a little money. Your number of Total Visitors is steadily climbing. In this example, is your business healthy? 

I can tell you one thing for sure: you don’t know. 

Total Visitors could be capturing your increase in customers. Or it could also be capturing an increasing number of visitors who aren’t converting, which signals that you have problems with your landing pages and that this is bad for your business.

Put another way, the metric could be giving you good news or bad news, and you don’t know which.

That problem is worsened for Total Visits, because there’s one more variable in the equation: how frequently do your paying customers visit?

The one metric you are watching really should be separated into two

  • Your website’s new visitors
  • Visits by existing customers

Having those two numbers side by side is going to be more convenient for you to quickly get back to work on your business. Here’s why:

  • Looking at the number of new visitors will tell you at a glance how your traffic acquisition efforts are working. Is the number going up? Great! At any conversion rate above zero, that will probably equate to growth for you.
  • Looking at the number of visits by existing customers will tell you a sort of combination of how many customers you have, and how sticky your app is. Is the number going up? Great, you have growth! Is it going down? You probably have a churn issue.

The above samples explain the actual goal of analytics. It isn’t to make a dazzling amount of numbers or visualizations appear. It’s to get the information you need as quickly as possible to run your business. 

Don’t look at your conversion rate as a percentage of Total Visitors. Base it on Non-Customers Instead.

If we’re segmenting our visitors overall, we want to do the same for our conversion rate.

Too many people create a quick calculated metric that looks like:

Signups / Total Visits

But this is problematic, because if you have growth you’ll accidentally create the illusion that your conversion rate is falling.

Huh? Isn’t that crazy?

Another way to express the above metric would be:

Signups / (Visits by Non-Customers + Visits by Customers)

Note that “visits by customers” is in the denominator. When you gain more customers, overall denominator for your conversion gets bigger, and your conversion rate artificially falls.

We actually want to know how well your site is turning non-customers into customers. So let’s focus on them, and restate our conversion rate as:

Signups / Visits by Non-Customers

This is a number that you can watch over time. Even when your customer base grows over time, your conversion rate can be used to look specifically at how well your introductory content is creating new customers out of new visitors.

Don’t Monitor Page Views as a KPI

In the early days of the Internet, we counted “hits” to a page or site as a sign of its popularity. We intuitively knew that more hits was better.

Back when the Internet was almost entirely static HTML content, that was fine.

Now that we have SaaS apps, it’s hard to draw meaning from Page Views going up.

Let’s play a tiny game. Any of the following scenarios could result in Page Views increasing. As you read through the list, sort each scenario into “good news” or “bad news”:

  • You’re on the front page of Hacker News
  • Users are encountering a fatal error and they’re refreshing the page repeatedly
  • You acquired 100 customers this week
  • Your ads started running
  • You rolled out a new feature that is a multi-page process

Did that get tougher as it went on? Perhaps you answered a couple with “I don’t know” or “it depends”?

That ambiguity is exactly why you shouldn’t use Page Views in isolation.

There could be a case where looking at page views of a certain subset of your content may make sense. But the lesson here is that as a rule, you should start from the broad question of “what do you want to know?”. From there, you can back your way into a metric that specifically illustrates the case in question.

Don’t monitor Bounce Rate as a KPI

Bounce Rate is another metric that hasn’t aged well.

If you’re unfamiliar, Bounce Rate is the percentage of people who leave after a single page view.

This was fine back in Web 1.0 when pages weren’t super interactive.

Google Analytics’s bounce rate documentation explains the problem. As soon as your page sends a second call to the tracking server for any reason, the visit is no longer a bounce.

As a result, for most people Bounce Rate is perpetually deflated.

For example, any of these things could quickly send a second call, if they’re instrumented:

  • Page scroll depth plugins
  • Video playback
  • A form submission or button click
  • Moving to another section of a single-page application (SPA)

In short, the technology isn’t doing what we intuitively think it should do.

(Side note: Google Analytics isn’t the only tool that has this tricky definition.)

There is one place where it’s OK to use Bounce Rate: you might consider monitoring the bounce rate strictly from your landing pages.

Bounce Rate is, by its definition, a metric that leans toward being used at the “top of the funnel” – that is, when someone is a new potential customer and doesn’t deeply know you or your product yet. (It wouldn’t make sense to look at the bounce rate from your loyal customers, right?)

So you could use Bounce Rate when looking just at the content that’s used as your landing pages. That gives you a reliable number to watch over time and try to improve.

Look at this instead

It’s my book, Analytics for Indies.

(Real subtle plug there, right? Hey – you’ve read this far – I’m optimistic this book will be of interest to you.)

Your business deserves clear and consistent metrics that are quick and easy to check, so you can get back to running your business.

Each chapter has one event of tracking that you can install to take you all the way to tracking that’s on par with billion-dollar e-commerce sites. With each chapter you’ll get one key metric to watch in your dashboards.

The goal for this book is to take you from “not sure what to do” to “done” in a couple hours, with future-friendly analytics data capture all set up.

Sign up now to be notified when the book releases, and get a free email course with four quick wins and a freebie cheat sheet that tells you what tools to use. The only way to get the cheat sheet is to sign up!

By Blake Ellison

Blake has spent the last eight years building analytics practices at PlayStation and Rakuten, and now he wants to help independent software creators with their own projects.