BI For Beginners

How to analyze email funnels to increase conversion

Email marketing is one of the most popular channels used today by online businesses for sales, brand awareness and growth.

Email can be deceptively simplistic from an analytics perspective which results in a lot of missed opportunity when it comes to understanding what's working and what's not.

My goal in this post is to show you how to analyze email funnels so that you can quickly and easily identify key insights for increasing conversion.

The Email Campaign Funnel

The first step in being able to analyze email funnels is understanding the concept of an email campaign funnel.

All email campaigns should have a main goal. In many cases that goal is to drive subscribers to a specific landing page where the individual can complete an action.

Some examples include:

  • Purchasing a product
  • Download an eBook
  • Upgrade to a higher tier of the service
  • Update their credit card information
  • Provide feedback or rate a product or service

For the rest of this post let's assume we are an eCommerce business that uses email marketing to drive sales of our products.

The first thing we need to do is define the steps in our funnel so let's do that.

Email Marketing Funnel

Above is a high-level view of our marketing funnel. Note that we aren't interested in total opens or total clicks but unique opens and clicks. Unique opens and clicks are defined as the number of unique email addresses which took these actions at least once.

Total clicks and total opens are vanity metrics which provide no real value.

This funnel is good but we can enhance it and add some numbers.

How to analyze email funnels

Let's break down what we see above into words.

  • We sent out 10,000 emails to our subscribers
  • 9,800 of those emails were delivered to the inboxes of our subscribers. 200 failed to get delivered. Out of these 200, 180 were hard bounced and 20 were soft bounced.
  • Out of the 9,800 that got delivered, 1,764 individuals opened our email. This translates to a 18% open rate.
  • Out of the 1,764 individuals that opened our email, 466 clicked links in our email, 441 of those clicks were on our main CTA.
  • All in all 175 out of the 10,000 individuals purchased products. 140 of them did it during the session in which they got to the site via our email, 35 others purchased products in a separate session. This translates to a 1.7% conversion rate for our campaign.

Delivery rate and bounce type

The first set of metrics from our funnel is related to delivery rate. Delivery rate is defined as the percentage of emails sent in the campaign that arrive in the inboxes of the members of the campaign. This percentage should be north of 95%.

If your delivery rate is under 90% it means that a significant percentage of the emails in your campaign either don't exist or belong to domains that are unavailable or no longer exist. The biggest danger of a low delivery rate is that it raises your chances of being flagged as a spammer in whichever service you are using to send out emails. Once flagged you will have a hard time convincing the service to allow you to send out additional campaigns.

A low delivery rate can also indicate that people don't trust your service enough to give you real emails.

To learn more about the differences between hard and soft bounces check out this useful post.

Open Rate

The next metric we want to look at is our open rate. The open rate is calculated by counting unique opens and dividing by emails sent. Since delivery rates are usually in the high 90s we typically don't look at unique opens divided by delivered but there is nothing wrong in doing this.

You typically want to aim for a open rate north of 15%. Since our list is made up of people that have purchased in the past or opted into our newsletter we expect an above open rate.

We see that our open rate is 18% which isn't bad but we should be able to get it up to around 25 - 30%.

The two most important factors in determining an open rate is the subject of the email and it's likelihood of being detected as spam.

Unique Clicks

Unique clicks is where 95% of email marketers make a critical mistake in determining the performance of their campaign. Since most services umbrella clicks on all links in the unique click rate, marketers mistakenly include clicks on non-essential links when reporting on the funnel. As you can see in my diagram it is important to separate clicks on the main CTA or CTAs, and other links. This can be easily done by counting unique clicks on specific links within your email and manually calculating a more relevant click rate. If your email service doesn't give you click counts on specific links in your email then go with what you have. It is better to have something than nothing.

In the end of the day your goal is to drive conversion so clicks on links to social media accounts, or recommended blog posts that have nothing to do with the goal of your email is just noise. Make sure to measure the funnel which leads to conversion.


The most exciting metric is conversions.

This is also the toughest part of the funnel to measure. Some email solutions allow you to add tracking pixels to conversion pages so you can tie specific utm campaigns to conversions.

Services like Google Analytics and Mixpanel can also help you understand the campaigns behind your conversions and then tie them back to specific email campaigns.

Delayed sales from email

One interesting aspect that you should consider tracking is the difference in conversions during the session and after. What I mean by this is to look at the percentage of sales from the campaign which happened immediately after the link click out of all sales which took place within say, 48 hours from the email being sent.

There is no rule of thumb for how to provide attribution for email campaigns. My recommendation is to set a time period after the email is sent and any conversion that takes place within that time period by an individual that was part of the email list AND at least opened the email, should be attributed to the campaign.

In our example we had 175 sales, 140 of them happened straight from clicking on the email while another 35 happened at a later stage.

If you notice that most of your sales from individuals that received your campaign are happening a week after they get the email, it can tell you a lot about your sales cycle. Some services require intensive due diligence and research. These businesses will struggle to make sales immediately from emails but this doesn't mean that email isn't efficient in getting the process started.

If you are sending out a lot of email and you see that it is driving more sales then spend some time analyzing the behavior of those that converted from email. You may find interesting patterns which will tell you a lot about your sales process from your customer's point of view.

Taking your email analytics to the next level

At this point you've got a very deep understanding of the performance of your email campaign. You know where in the funnel you need to focus and where the drop offs are happening.

The next big step in analyzing email funnels is to break everything down by cohort.

The diagram above showed our email campaign when putting the entire list into one bucket. The power of looking at funnels by specific cohorts is you get a much better understanding of the campaign and it's relation to the population. Let's look at an example to explain this concept.

The diagram below shows our email campaign when looking at a specific 1,000 emails from our campaign. These emails belong to individuals that previously purchased from our eCommerce store.

analyze email funnels - step by step guide

Notice how much higher the open, click and conversion rates are for this specific group of emails compared to the entire list. According to the diagram above past customers are converting at 9% compared to only the fully blended rate of 1.7%.

We can also see that 90 conversions out of the total 175 belong to the 1,000 past customers. That means that we only managed to get 85 sales from the other 9,000 emails. Such an extreme can tell us that our past customers just "get it" while our other subscribers are either a bad fit for our product or there are too many factors holding them back from buying.

Now that we have identified a core difference in our list thanks to looking at the funnel by cohort, we can test different hypotheses through smaller scale campaigns.