How to calculate email marketing ROI

The simple formula for email marketing ROI

The Data and Marketing Association consistently reports an ROI from email marketing of somewhere between $35 and $40 for every $1 spent. But how is that calculated?

Well, here’s the deal: There’s the easy calculation, which is the one pretty much every article about email marketing ROI gives. And then there’s the real one, which is considerably more complex.

In simple terms, to calculate your ROI for email marketing, you just need two numbers:

1. Expenses related to email

2. Revenue from email

Subtract expenses from revenue to get profit, and then divide that by expenses.

For example, if you generated $100,000 in revenue and had $10,000 in expenses, that would be $90,000 in profit. Divide that by the $10,000 in costs, and you get $9, or a 900% ROI.

One slightly different way of doing this is to divide your total revenue by the number of emails sent. Now, you have revenue per email. If you can also get a cost per email, then you can get the ROI per email. This method would incorporate all emails sent in a given time period.

The complexity of email marketing ROI begins

OK, so it’s that easy, right? Wrong. There are a few factors that make calculating email ROI a bit harder than that. And by a bit, we actually mean a lot.

How do you calculate your email marketing expenses?

Let’s start with this one: How can you calculate your total email expenses? Do you consider just the cost of producing an email that generated revenue? You could do that to gauge how well particular email copy and calls to action are performing. But that doesn’t give you a monthly or annual figure for ROI.

Okay, so what are the costs of sending all the emails out per month or per year?

To get this figure, you have to consider the cost of your email service provider, the salaries or hourly rates of the employees and contractors on your email marketing team, and any other support technology like email testing services and deliverability checks. It’s unlikely the DMA is including all this in their calculations.

Precisely adding up employee expenses might not be easy, because some employees may be working in multiple departments. For example, a graphic designer might create email graphics, but also design images for print materials, web pages, and digital ads.

When you accurately factor in all these expenses, it’s likely going to be much harder to reach the 3500x ROI reported by the DMA.

What defines an email conversion?

The next complicating factor concerns conversions.

Not every email is trying to make a sale. You also have emails linking to content, running polls and surveys, asking for reviews and testimonials, motivating people to share stuff on social media or download special reports and white papers, and encouraging subscribers to take many other actions that qualify as conversions.

You also have welcome emails and other automated sequences that, in theory, were already paid for when they were first created, but will continue to generate conversions.

Here are some possible conversions from email:

• Making a purchase

• Signing up for a free trial or demo

• Downloading gated content

• Clicking through to content

• Renewing a subscription

• Accessing new features in an application

• Leaving a review

• Returning to complete a purchase from an abandoned cart email

These are all worth something to you. But only three translate to immediate sales.

You also have different email segments based on a whole host of characteristics. Some of these are automated. Some may be for special VIP customers, or enterprise customers only. Some emails are about nurturing leads while others focus on retaining existing customers.

Emails to your enterprise customers inherently are worth more than emails to the masses on a per-customer basis. So is it right to lump all those together? Probably not.

Five attribution models for your email program

Consider this scenario:

A person clicks on your website while searching on Google. Here’s one very possible sequence of events that person could travel through:

1. They click on a Google Ad you paid for.

2. The landing page intrigues them, so they look at your About page.

3. From there, they go to your blog and read a couple of helpful articles.

4. On the second article, they sign up for your email newsletter to claim your free coupon offer for new subscribers.

5. They receive the welcome email with the coupon offer, but don’t use it right away.

6. They use your live chat feature to ask a couple of questions.

7. While waiting, they read some of your testimonials or case studies.

8. They view a product page, and then leave the site.

9. The next day, they get a sales email from you and they click through to the product or landing page.

10. The landing/product page succeeds in closing the sale.

Now, that all happens within a couple days. We could come up with an infinite number of scenarios like this, and others that spread out over weeks, months, and even years.

So, which of those ten steps in that example gets to take credit for the conversion? This isn’t easy to answer, and that’s exactly the problem ‘attribution’ attempts to solve.

You can make a pretty good argument that every one of those ten steps played a role in the sale. Let’s look at five attribution models that show you a variety of ways to attribute a conversion.

1. Last click attribution

As the name implies, this model gives all the credit to the last place a person clicked. In the scenario above, the landing/product page connected to the sales email would receive all the credit for the conversion.

You’ve seen this before. Marketers might claim that a particular landing page has a 40% conversion rate, for example. That number is almost surely using last-click attribution, because almost no landing page can be so good – all on its own – to generate numbers anything close to that. Something else came before it.

This model has obvious shortcomings. But, it does allow you to fine tune your focus on the asset that produces the final click, and if you can raise that number, you’ll make more money.

2. First click attribution

This is the same idea, but now we’re starting at the top of the sequence. In the above scenario, all the credit for the conversion would go to the Google Ad.

And to be fair – that is what initiated everything. If the Google Ad wasn’t there, this customer probably never would have found your site. So the Google Ad certainly seems to deserve some credit for the conversion. But should it receive all of it?

After all, the customer didn’t buy directly after seeing this ad. It took a couple days and quite a bit more marketing outreach and follow up.

3. Linear attribution

This model gives equal credit to every step in the process. So in that above scenario, all ten steps can claim credit for the conversion. Your testimonial page, the blog, the emails – all of it can claim equal shares of credit for the sale. You would probably break this up and give 10% credit to each.

So in terms of email marketing ROI, if the customer spent $200 on this purchase, the linear attribution model would assign $20 of that as ROI from the sales email, and another $20 to the welcome email. Thus, $40 in ROI would be attributed to your email marketing department’s work.

You can see why this approach works better than the first two in situations like this. But do all ten steps in the process deserve equal credit?

Even if this customer didn’t buy right away, a few more emails might have eventually done the job, even without some of these other steps. Getting them to sign up for the email list really plays a pivotal role here, right?

4. Position-based attribution

In this model, you give different weights to different steps in the process. For example, you could give 40% of the credit to the first step, 40% to the last step, and 20% spread out among all the middle steps.

With that approach, the Google Ad gets 40%, the landing page gets 40%, and both emails combined would get 5%, or $10 in the sales example above.

You could adjust those percentages if you wanted to, but since every process will be different, that would get tricky fast. Yet, the middle steps are getting shortchanged.

5. Time decay attribution

Here, the credit decreases with time. The first click gets the least amount of credit, and each step closer to the sale gets more. You would have to choose the percentages yourself.

The idea here is that the customer is being nurtured and the more they engage, the closer they get to the sale. That’s why the steps that happen closer to the sale get more credit.

Which attribution model is best?

Man, these are hard questions! And you thought email marketing ROI was easy.

The reality is, no model is perfect, and each has its pros and cons, as you can see. But the point is to try to properly assign credit for a sale to all the marketing assets that played a role in producing it.

If you have a healthy array of email segments and are regularly sending emails as your primary marketing media, then the time-decay approach might make the most sense, or perhaps even the last-click model.

If you’re sending a healthy dose of multimedia marketing, including direct mail, social media, TV, and more, then you might want the linear approach.

New customers might need a different model than VIPs, who might be different from enterprise clients, who might be different from customers who found you through social media, who might be different from your recurring membership customers.

Further complications to calculating email ROI

“Make it stop, make it stop!”

Yes, there’s more. Here are a few additional factors to consider:

Tracking your data

All of these attribution models depend on a well-functioning data tracking system. In fact, the last three depend on it quite heavily. You really have to know the steps a customer has taken to use these models effectively.

Tracking involves monitoring both customer movement through your funnels and sales processes and the actual revenue generated from each customer. Analytics services can help you track where clicks and conversions are originating.

If you don’t yet have data at this high level of quality, you’ll probably have a slightly less accurate figure for your email marketing ROI. And that’s okay for now if it’s the best you can do.

Assigning a value to conversions

Sometimes, measuring the value of a conversion is straightforward. In ecommerce, a specific cart total can often be reported back with conversion data.

But in other cases, especially when a conversion is the start of an ongoing relationship with a customer, the value of a conversion is much harder to determine.

If you know the customer lifetime value, the average amount a customer spends over the duration of their relationship with your company, you can choose to attribute this full amount upon conversion - regardless of the attribution model you choose.

While this may not result in the most accurate ROI as far as actual revenue generated during the time period being measured, it does provide decision makers with the details they need to determine long-term ROI. This can be especially helpful when making key strategy decisions about how to distribute resources to marketing programs.

So how do you calculate email marketing ROI?

Email is almost never the first click. But it is spectacular at nurturing and converting leads when used well.

Tools like Mailgun’s email ROI calculator can help you get a good estimate, but ultimately, you have to take everything you’ve read here and figure out the best approach that also works for your situation and the assets and data you have on hand.

At the end of the day, you’re always going to need two numbers: expenses and revenue. And possibly a third – the number of emails sent. So do your best to determine how much revenue you can attribute to your email marketing, and then use that figure in the simple formula.

And then, once you’ve mastered that, you can start working on getting better data and developing your use of attribution models to get a more precise number, the one that will surely get you that raise.