Published August 21st, 2017 by Assaf Trafikant

Cracking The Google Analytics Funnel

Building funnels is one of the oldest, most fundamental practices in the analytics world, and Google Analytics in particular. Funnels let us put aside our marketing and marketing-optimization efforts and focus on the product itself: all of the core processes that make it what it is. Checkout, registration, basically any process that takes you to a certain point, has to go through the same course.
Setting the funnel is no laughing matter, and has to be done carefully and precisely. If everything goes well, our final funnel should look like this:

This funnel has three steps on the way to the “Thank-You” page. We can see that only 83.12% go from step 2 to step 3, and only 30.96% move on from step 3 to the end. Eventually, when you weight all of the steps, you can see that 25% of those who start also finish it, and that’s the number we see at the bottom. We’ll call that the Golden Ratio, the number that represents the conversion rate of the funnel – how many of those who started the funnel reached the end.

The Golden Ratio is not the regular conversion rates you see in your Google Analytics reports; those refer to the conversion rate as the ratio between the number of conversions and the number of site visits.

Now that that’s clear, let’s meet Jane.

“Hi Assaf, I have this incredible funnel, and we’re working hard on optimizing it.“

“Cool, what’s your question?“

“I have to choose a different date range to see its performance during that period. It’s exhausting, and I can’t see the bigger picture over an extended period.“

– “Excellent question! If you have five minutes to spare, I’ll show you a pro-tip. But keep it between us, ok?“

– “I just have this feeling that on certain days, the funnel behaves differently than on others, and it happens every month. Maybe it has nothing to do with the improvements I’m making? I have to get a clear visual.“

Jane is right. The percentage of visitors who move on from step 1 to step 2 might not be constant and may change over time. So she wants to see more than just the entire funnel’s conversion rate – she wants to know the conversion rate from one step to the next, and how it changes over time. To get a visual representation, she has to play around with the date range and see the numbers change. It’s a bit primitive, but it works. But what if many factors play a part in the funnel? Let’s ask some questions:

  • Does the funnel act differently on weekends and weekdays?
  • Are the chances of a visitor moving from step 1 to step 2 always the same, no matter what? Is it possible that inside the funnel itself, the steps differ by time?
  • Does the funnel act differently after a specific date each month?
  • Is the funnel behavior trend circular? Is it repetitive (going up and down at a steady pace)?

The Solution

There are several ways to tackle this issue. One is to use external tools that extract the information to an excel sheet, for example, and analyze everything there. But I want to show you a fairly simple way that doesn’t force you to exit Google Analytics and, despite its limitations, is pretty fast. Additionally, and to make things simpler, I’m going to show you a very basic way to test these claims. You can improve on it later.

Part One

Go to the admin panel and then go to funnel settings. It should look something like this:

Part Two

Copy all steps into Notepad or a similar text editor. It will easier later.

Part Three

Go to the Admin, then “View” (extreme right), click “Segments”, and create a new segment. Name it something like Step 1. Then go to Conditions and pick “Page”, which is like the URL of your first funnel page (remember those values you copied into Notepad? Paste the first one here).

Now for the second segment, we’ll use the “Sequence” type and add these two steps:

It basically means that we want to see all sessions with both steps, one after another.
And that’s how it looks for the third segment:

And for the thank-you page:

If you have a three-step funnel plus a Thank You page, you should have total of four segments, like in my example. Keep in mind that your segment setting has to be precisely the same as the funnel settings. So if the first step of the funnel use CONTAINS condition, keep it like that in the segment. If the funnel uses Regex, the segment should, too.

Part 4

Go to the main display on Google Analytics, and click Add Segment right above the graph:

Then, choose all of the segments you’ve created. There’s a four-segment limit, so if your funnel has more than four segments, select the first four you’ve created:

Click Apply, pick a specific day, and take a look at your report – by the hour, for example:

The distance between the lines represents the same percentage we saw earlier in the classic funnel diagram. the % of users who moves from step 1 to 2 etc. Theoretically speaking, we would expect this distance to be equal all along, but it’s not.

It shouldn’t matter that there are more visits in the morning than in the evening – what matters is how many moves on to the next step.
I see that at 4 pm, the lines are really close together, but they’re farther apart in the mornings. That means that at 4 pm, even though there are fewer visits, the chance of completing the funnel is higher. The question, now, is why.
Maybe because our morning campaigns aren’t that good? Perhaps the server is overloaded in the morning, which might affect the funnel’s performance? Maybe people are more willing to convert in the afternoon?
Everything’s possible. Now you can set off on your way with a beautiful piece of homework and an excellent research question.

Wait, that’s it?

It’s just the beginning. You could use this method to look at a range of dates and see if the funnel acts differently at different periods, days (weekends vs. Weekdays,) and improve the segments to include devices, campaigns, countries, or any other dimension you can think of.
In the next article, “The Golden Ratio- Google Analytics Real Conversion Rate“, I’ll show you another way to show the Golden Ratio over time – that is, the final conversion rate. It’s a crafty little trick that deserves its own post.

More Articles

All articles

My Extensions

  • Analytics UTM Builder

    50,000+ USERS

    • Star
    • Star
    • Star
    • Star
    • Star


  • Data Studio Auto Refresh

    1,000+ USERS

    • Star
    • Star
    • Star
    • Star
    • Star


  • No Stress Booking

    10,000+ USERS

    • Star
    • Star
    • Star
    • Star
    • Star