# How to Measure Email Marketing ROI Honestly

> A practical way to calculate email marketing ROI: count real costs, attribute revenue with UTM tags, run a holdout test, and stop trusting open rates.

Canonical: https://www.cresca.xyz/blog/measuring-email-roi
Author: Lalit Kumar Jangid
Published: 2026-09-05
Last updated: 2026-10-26

Most email marketing ROI numbers you see quoted are either borrowed from someone else's business or built on clicks that may never have happened. Neither helps you decide whether your own newsletter is worth the hours you put into it. This post walks through a way to calculate return that you can defend in front of a skeptical co-founder: what to count as cost, what to count as revenue, where attribution breaks, and how to get a cleaner answer when the stakes are high.

## What is the email marketing ROI formula?

Email marketing ROI is revenue attributed to email minus total email cost, divided by total email cost. Multiply by 100 for a percentage. The arithmetic is the easy part. The inputs cause the trouble, because teams tend to undercount cost by leaving out their own time and overcount revenue from anyone who ever opened an email.

Return on investment for email is the same arithmetic as for anything else:

```
ROI = (Revenue attributed to email - Total email cost) / Total email cost
```

Multiply by 100 if you want a percentage. A result of 2.0 means every dollar spent came back three times: the original dollar plus two more.

The arithmetic never causes trouble. The two inputs do. Teams tend to undercount cost (because their own time feels free) and overcount revenue (because anyone who ever opened an email gets tagged as an email customer). Fix both and the number becomes useful.

## Step 1: Count every cost, including your time

Write these down for the period you are measuring, usually a month or a quarter:

- **Software.** Your email platform subscription, plus any design, image or copy tools you pay for mainly because of email.
- **People.** Hours spent planning, writing, building and reviewing campaigns, multiplied by a realistic hourly rate. For a founder, use what you would pay a contractor to do the same work.
- **Incentives.** Discounts and free shipping that only exist because of an email offer. A 20% code is a cost, not a rounding error.
- **Acquisition.** Anything you spend specifically to grow the list, such as a giveaway prize or paid promotion of a signup page.

If your email marketing ROI only looks good when labour is excluded, that is worth knowing too. It usually means the program needs to get more efficient, not that it should stop.

## Step 2: Decide what "revenue from email" means before you look

Revenue is not recorded by your email tool. It lives in your store, your billing system or your CRM. Your job is to connect a sale back to an email using a rule you choose in advance, so you are not tempted to pick the most flattering one later.

### Tag every link

Add UTM parameters to the links in each campaign so your web analytics can tell email traffic apart from everything else. Google's own [guide to URL builders and UTM parameters](https://support.google.com/analytics/answer/10917952) says to always set `utm_source`, `utm_medium` and `utm_campaign`, and notes that values are case sensitive. Pick one convention and stick to it:

```
https://example.com/spring?utm_source=newsletter&utm_medium=email&utm_campaign=2026-04-spring-sale
```

If `Email` and `email` both appear in your medium values, your reports will split one channel into two.

### Pick an attribution window

Decide how long after a visit from an email a purchase still counts. Short windows (one to three days) understate email's influence on considered purchases. Long windows (30 days) let email claim sales it had little to do with. Choose the window that matches how long your customers usually take to buy, write it down, and keep it constant so month-to-month comparisons mean something.

### Watch out for codes that leak

Discount codes are a handy second signal, but codes get shared on coupon sites. If a code meant for subscribers shows up in orders from people who are not on your list, those orders are not email revenue.

## A worked example (hypothetical numbers)

The figures below are invented to show the method. They are not benchmarks and you should not compare your results against them.

Imagine a small online tea shop measuring one quarter:

| Item | Amount |
| --- | --- |
| Email platform subscription | $90 |
| Owner's time: 18 hours at $40/hour | $720 |
| Discounts redeemed on email-only offers | $310 |
| **Total cost** | **$1,120** |
| Orders from sessions tagged utm\_medium=email, within a 7-day window | $4,480 |

ROI = ($4,480 − $1,120) / $1,120 = 3.0, or 300%.

Now the honest adjustment. Some of those buyers were loyal customers who would have ordered anyway. Without a test (see below), the shop cannot know how many. So the defensible statement is "email-tagged sessions produced $4,480 of orders against $1,120 of cost," not "email made us $3,360."

## Why shouldn't opens and clicks drive the ROI calculation?

Because both can be recorded without a person reading the email. Under Apple Mail's Protect Mail Activity setting, remote content downloads privately in the background, so a tracking pixel can load unread, and some security scanners follow links automatically. Measure orders, signups and renewals in the system that records them, and treat engagement metrics as rough diagnostics.

Older ROI guides build a funnel from opens to clicks to sales. That funnel has weak links. Apple's Mail app offers a [Protect Mail Activity setting](https://support.apple.com/guide/mail/protect-email-privacy-mlhlp1205/mac) under which remote content is downloaded privately in the background when a message arrives, and the reader's IP address is hidden from senders. A tracking pixel can therefore load without any human reading the email. Security scanners at some companies also follow links automatically, which inflates clicks.

We go deeper on this in [why open rates stopped being reliable](https://www.cresca.xyz/blog/why-open-rates-stopped-being-reliable). For ROI, the takeaway is simple: measure the end of the funnel (orders, signups, renewals) in the system that actually records them, and treat engagement metrics as rough diagnostics at best.

## How does a holdout group show email's real impact?

A holdout group is a random part of your list that receives no extra campaigns for a period while the rest does. Comparing revenue per person between the two groups shows who bought because of an email, not just after one. The difference, multiplied by the size of the emailed group, estimates incremental revenue.

Attribution tells you who bought after an email. A holdout tells you who bought **because of** an email. If a decision depends on the number, such as whether to hire someone to run email, a holdout is worth the effort.

1. Export your list and randomly split it into two groups. A spreadsheet random-number column sorted ascending works fine. Keep the groups similar in size.
2. Send your campaigns for the period only to group A. Group B receives nothing extra (keep sending anything they would expect, like receipts).
3. At the end of the period, compare revenue per person in group A against group B, using your store's order data matched by email address.
4. The difference, multiplied by the size of group A, is your estimate of incremental revenue.

Keep the holdout short, a few weeks is usually enough, so you are not neglecting real subscribers. And remember that small lists produce noisy results; a difference of a few orders could be chance.

## What your email platform can and cannot tell you

Your sending tool is the source of truth for delivery, not for revenue. In Cresca, campaign reporting covers recipients, deliveries, bounces, complaints and unsubscribes. Those are the inputs that tell you whether a campaign reached people and whether it annoyed them. Revenue tracking stays in your store or analytics, joined up through the UTM tags above. Our breakdown of [the email metrics that matter](https://www.cresca.xyz/blog/email-marketing-metrics-that-matter) covers how to read those delivery numbers.

That split is deliberate. If a campaign delivered well but produced no orders, the problem is the offer or the landing page. If it bounced heavily or drew complaints, fix the list before worrying about revenue at all.

## A monthly ROI routine you can actually keep

- On the first working day of the month, export last month's orders filtered to `utm_medium=email` within your chosen window.
- Add up platform cost, logged hours and redeemed discounts.
- Calculate ROI and record it in the same sheet as previous months, along with the window and any changes to your method.
- Note complaint and unsubscribe counts next to each campaign. A campaign that earned money but drove a spike in complaints may cost more later in deliverability.
- Once a quarter, run a short holdout to sanity-check how much of the attributed revenue is truly incremental.

## When the number looks bad

A low or negative ROI is a prompt to look closer, not a verdict. Check, in order:

1. **Tagging gaps.** Are some links missing UTMs, so sales land in "direct" traffic?
2. **Delivery.** Did a meaningful share of messages bounce or get filtered? No campaign earns from an inbox it never reached.
3. **Cost structure.** Are you spending hours per campaign on tasks a template could handle?
4. **The offer.** Are you sending reasons to buy, or only updates?

Measured this way, email marketing ROI stops being a slogan and becomes a number you can improve one input at a time.
