Email Marketing ROI: Separating Claim From Evidence

Email marketing returns $42 for every $1 spent. You have read that hundreds of times.

Here is what that number actually is.

It is a 2019 figure, from the United Kingdom, denominated in pounds.

Not dollars. Not 2026. Not a like-for-like comparison with anything (Litmus, 2026).

The more current figure is $36 per $1, and it has its own problems.

Both are worth understanding, because email genuinely does work. 📧

It simply does not work the way the headline number implies.

🎁 Compare Email Marketing Trials →

🧾 主要发现概览

Claim	What it actually is	Status
$42 per $1 spent	2019 UK figure, in pounds	Widely misquoted
$36 per $1 spent	More recent, self-reported	Better, still an average
How the ratio is computed	Email-attributed revenue divided by email spend	Attribution problem
Who publishes it	Email platform companies	Commercially interested
Variation between senders	Enormous, smoothed by averaging	Rarely mentioned
Best treated as	Order of magnitude, not a measurement	Per the sources themselves

🔍 Where the $42 Came From

The number has a specific origin, and tracing it explains a lot.

It comes from a Direct Marketing Association study in the United Kingdom.

The measurement was in pounds sterling, and it dates to 2019.

How it happened

Nobody set out to mislead. Each retelling dropped one qualifier.

The country went first, then the year, then the currency.

What remains is a bare number with no context attached.

Why the currency matters

A pound and a dollar are not the same amount of money.

Quoting a pound-denominated ratio in dollars changes what it claims.

The ratio itself is currency-neutral, but the framing is not.

Why the year matters more

Email in 2019 predates significant changes to privacy and tracking.

Measurement methods changed materially afterwards.

A 2019 attribution figure describes a different measurement environment.

⚠️ The Deeper Problem: Attribution

Even the current $36 figure has a structural issue worth understanding.

The ratio is calculated by dividing email-attributed revenue by email spend.

Email-attributed revenue is not the same as revenue email created.

情况	Credited to email?	Caused by email?
Customer already intended to buy, clicks the email	是的	不
Customer had abandoned a basket, returns via email	是的	Partly
Customer discovers a product in the email	是的	是的
Customer saw an advert, then used email as a shortcut	是的	不
Existing subscriber buying on a normal schedule	是的	Probably not

Only one row is unambiguously caused by the email.

Yet all five are counted in the attributed revenue.

Why email specifically inflates

Email reaches people who already chose to hear from you.

They are your existing customers, and they were likelier to buy anyway.

Email harvests intent efficiently. That is not the same as creating it.

What the honest measure would be

Incremental revenue, meaning sales that would not have happened otherwise.

Measuring it requires holding back a control group who receive nothing.

Very few businesses do this, and the published averages do not.

🔎 Test Email Platforms Free →

✅ Why Email Still Deserves Your Time

Having taken the headline apart, here is the honest case in its favour.

It does not rest on any ROI multiple.

Advantage	Why it holds
You own the list	No platform can take it away
No algorithm between you and readers	Delivery is not rationed
Costs scale slowly	Ten times the list is not ten times the cost
Recipients opted in	They asked to hear from you
It works while you sleep	Automated sequences run themselves
Portable between providers	You can export and move

The ownership point is the strongest one and the least quantified.

Social followers are rented. Email subscribers are yours.

The contrast that matters

A social platform can change its algorithm and halve your reach overnight.

It has happened repeatedly, to businesses that built everything there.

An email list keeps working regardless of anyone else’s decisions.

Why the cost structure helps small senders

Most platforms charge by list size, in bands.

Sending to 900 people usually costs the same as sending to 400.

So the return improves as the list grows, without extra spend.

🧮 Measuring Your Own Return Honestly

Forget the industry averages. This is the calculation that applies to you.

步	例子
Annual platform cost	$240
Your time, 2 hours monthly at $40	$960
Total real cost	$1,200
Revenue attributed to email	$9,000
Naive return	7.5 to 1
Assuming half would have bought anyway	3.75 to 1

Note that the time cost is four times the platform cost.

Published ROI figures almost never include your hours.

Why halving is a reasonable default

Without a control group you cannot know the incremental share.

Halving is conservative and keeps you honest.

A return of nearly four to one is still excellent.

How to measure it properly

Hold back 10% of your list from one campaign.

Compare their purchase rate with everyone else’s.

The difference is the real effect, and it takes one campaign to learn.

📬 What Actually Drives Results

Since the averages describe nobody, here is what varies between senders.

Permission is the biggest single factor

A list built from people who asked to join behaves completely differently.

A purchased or scraped list produces almost nothing, and damages your sending reputation.

How you gathered the addresses predicts the results better than anything else.

Relevance beats polish

A plain text message about something the reader wants outperforms a designed template about something they do not.

Sending everyone the same thing is the most common mistake.

Even splitting a list into two groups improves results noticeably (Mailneo, 2026).

Frequency is a real variable

Too rare and people forget who you are.

Too frequent and they stop opening, then unsubscribe.

Monthly is a safe starting rhythm for most small businesses.

📮 The Deliverability Problem

An email that lands in spam has a return of exactly zero.

This receives far less attention than open rates, and matters more.

因素	Effect on delivery
Authentication records set up	必备
Sending from your own domain	Strongly helps
Low complaint rate	Critical
Removing inactive subscribers	Helps considerably
Sudden large sends	Hurts, if unusual for you
Purchased lists	Damages reputation quickly
Misleading subject lines	Generates complaints

Authentication is a one-off technical setup most small senders skip.

Without it, providers treat your mail with suspicion by default.

The counter-intuitive fix

Removing subscribers who never open improves your results.

A smaller engaged list delivers better than a large indifferent one.

Providers judge you partly on whether recipients engage (Eggknite, 2026).

Why complaints matter so much

One person marking you as spam counts heavily against you.

Make unsubscribing easy and obvious, because that is the better outcome.

An unsubscribe costs you one reader. A complaint costs you future delivery.

🌱 Growing a List From Nothing

The averages assume an established programme. Most people are not there.

Method	速度	品质
Asking existing customers	Immediate	最高
A form on your website	Slow, steady	高的
Something useful in exchange	Faster	Good, if relevant
In person, at the point of sale	Steady	非常高
Social media prompts	Variable	Mixed
Competitions and giveaways	快	Usually poor
Buying a list	立即的	Worthless

The fastest methods generally produce the weakest subscribers.

People who joined for a prize want the prize, not your emails.

The most overlooked source

Your existing customers already know and trust you.

Most businesses have never simply asked them to subscribe.

One message to your current contacts often builds the first hundred.

What to offer in exchange

Something genuinely useful and specific to what you do.

A guide, a checklist, or a discount on a first order.

Generic offers attract generic subscribers.

⚖️ The Rules You Must Follow

Email marketing is regulated, and the rules are not optional.

Requirement	为什么
Permission before sending	Legally required in most places
Clear sender identity	Recipients must know who you are
A working unsubscribe link	Required, and must work
Acting on unsubscribes promptly	Usually within days
A physical postal address	Required in several jurisdictions
Honest subject lines	Deception is specifically prohibited

Rules differ by country, and the strictest applicable one governs you.

If your list spans several countries, build to the strictest standard.

Why compliance also improves results

Every rule above describes what good senders do anyway.

Permission, honesty and easy unsubscribing all improve engagement.

The regulations largely codify what already works (Angarum Media, 2026).

Keep the evidence

Record when and how each subscriber joined.

Most platforms do this automatically if you use their forms.

That record is your defence if anyone ever queries it.

💵 What Email Platforms Actually Cost

Pricing in this category is banded by list size, which has an odd consequence.

What this means in practice

If you pay for a band up to 1,000 subscribers, reaching 900 costs the same as 400.

Every subscriber added inside a band is effectively free to reach.

So growth improves your return without increasing your spend.

Watch the band boundaries

Crossing into the next tier can raise the bill sharply.

This is a good reason to remove subscribers who never open anything.

Pruning saves money and improves delivery at the same time.

Free tiers are genuinely usable

Most major platforms offer a free tier covering several hundred subscribers.

That is enough to run a real programme for a small business.

我们的 审判目录 lists platforms you can test at no cost.

📈 What to Measure Instead of ROI

Four numbers tell you more than any return multiple.

措施	它揭示了什么
List growth per month	Whether the asset is building
Unsubscribe rate per send	Whether you are sending well
Replies received	Genuine engagement, hard to fake
Revenue per subscriber per year	Comparable over time

Revenue per subscriber is the most useful single figure.

It lets you compare this year with last year on your own terms.

Why replies deserve attention

Open rates became unreliable after privacy changes affected tracking.

Replies cannot be inflated by an image loading automatically.

A message that produces conversations is working (Tabular, 2026).

The unsubscribe signal

A small steady rate is healthy and normal.

A spike after one campaign tells you exactly what not to repeat.

Treat it as feedback rather than as loss.

🕰️ How Long Before It Works

Expectations here are usually wrong in both directions.

阶段	What to expect
First 3 months	List building, little revenue
Months 3 to 6	First consistent responses
Months 6 to 12	Patterns become readable
Year two onward	The compounding starts

The averages you read describe programmes in the fourth row.

Comparing a three-month-old list against them guarantees disappointment.

Why year two is different

By then the list is larger and the audience knows you.

You also know what they respond to, which took a year to learn.

Most of the reported return comes from that accumulated knowledge.

The mistake that resets the clock

Stopping for six months undoes much of the progress.

People forget who you are and complaint rates rise on your return.

Consistency matters more than volume or polish.

✉️ What to Actually Send

The commonest reason small businesses stop is running out of ideas.

类型	Effort	Usually works?
Answering a common question	低的	是的
Recent work or a finished job	低的	是的
A genuine offer with a deadline	低的	是的
Seasonal or timely advice	缓和	是的
Company news about you	低的	很少
Long formal newsletters	高的	Often not
Reused social media posts	Very low	不

The three easiest types are also the three that work.

Company news is the exception people default to first.

The question bank approach

Keep a note of every question a customer asks you.

Each one is a message, and you never run out.

You already know the answer, so writing takes minutes (Christopholivier Consulting, 2026).

Length is not the point

Three useful paragraphs beat a designed newsletter nobody finishes.

Short messages are also faster to write, so you keep sending them.

Sustainability matters more than any single message.

Write to one person

Imagine a single customer and write to them directly.

Messages addressed to an audience read as broadcasts and get ignored.

The word you want is “you”, not “our valued customers”.

🔁 Automated Sequences Worth Setting Up

Three automations do most of the work in small programmes.

Sequence	Trigger	价值
Welcome message	Someone subscribes	Highest engagement you will get
Abandoned basket	Checkout not completed	Strong, if you sell online
Post-purchase follow-up	After delivery	Reviews and repeat orders

The welcome message is opened more than anything else you send.

Most businesses waste it on an empty confirmation.

What the welcome should do

Tell them what you will send and how often.

Give them one genuinely useful thing immediately.

Set the expectation you intend to keep, then keep it (Mailneo, 2026).

Why these three are enough

Elaborate automation suits large catalogues and large teams.

A small business gets most of the benefit from these three.

Add more only when these are working reliably.

🔀 Where Email Fits Beside Everything Else

Email works best as the destination, not as the whole strategy.

Channel	Job it does	You control it?
搜索	Brings strangers who are looking	Partly
社交媒体	Builds familiarity over time	不
付费广告	Buys attention immediately	While you pay
您的网站	Converts interest into contact	是的
电子邮件	Keeps the relationship alive	是的

Only the last two rows are genuinely yours.

The other channels rent you attention on their own terms.

The sensible sequence

Use the rented channels to bring people to something you own.

Convert that visit into an email address wherever you reasonably can.

Then the relationship continues without paying for reach again.

Why this matters more each year

Reach on rented platforms has become steadily more expensive.

Organic visibility fell as those platforms monetised attention.

An owned list is insulated from that trend entirely.

The one-line summary

Rent attention, then convert it into something you own.

That principle survives every change in platform economics.

🚫 这些数据无法告诉你什么

Publishers sell email software. Litmus, Omnisend and Klaviyo all do.

Figures are self-reported. Companies estimate their own returns.

Averages hide huge variation. Your sector and list maturity dominate.

Attribution is not causation. Credited revenue overstates created revenue.

New lists behave nothing like mature ones. The averages assume established programmes.

The variation point deserves weight

These averages combine established retailers with new newsletters.

A business with 40 subscribers is not in the same distribution.

An average across a wildly varied population describes nobody.

To their credit

The sources themselves suggest treating these as order-of-magnitude confirmation.

That caveat is usually stripped out by the time you read the statistic.

The publishers are more careful than their popularisers (Litmus, 2026).

🏁 简短版

The $42 per $1 claim is a 2019 UK figure measured in pounds.

Quoting it as current US dollars compounds two separate errors.

The current figure of $36 is better but still self-reported and attributed.

Attributed revenue counts sales that would have happened anyway.

Email still deserves your effort, for a reason no multiple captures.

You own the list, and no algorithm sits between you and your readers.

Measure your own return, include your hours, and hold back 10% to see the truth. 📧

🚀 Browse Marketing Trials by Category →

❓ 常见问题

Is the $42 per $1 email ROI figure accurate?

It is a 2019 United Kingdom figure denominated in pounds. Quoting it as current dollars is doubly wrong.

What is the better figure?

Around $36 per $1, which is more recent. It remains self-reported and based on attribution.

What is wrong with attribution?

It credits email with sales that would have happened anyway. Credited revenue overstates created revenue.

So is email marketing worth doing?

Yes, but for a different reason. You own the list, and no algorithm can reduce your reach.

How do I measure my own return?

Include your hours, not just the subscription. Time usually costs several times the software.

How do I find the real effect?

Hold back 10% of your list from a campaign and compare purchase rates. That difference is incremental.

Why do platforms publish these numbers?

They sell email software. That does not make the figures false, but it explains the emphasis.

Do the averages apply to a small list?

Not really. They combine established retailers with new senders, and a 40-person list is not in that distribution.

Is email better than social media?

For reliability, yes. Social reach depends on an algorithm you do not control and cannot appeal.

Does list size change the economics?

Yes, favourably. Platforms charge in bands, so a growing list often costs nothing extra to reach.

What should I do first?

Start collecting addresses, with permission, before you worry about which platform to use.

📚 参考资料

Litmus. (2026). Email marketing guide on email ROI2026年8月8日检索自 https://www.litmus.com/resources/email-marketing-roi

Mailneo. (2026). Email marketing statistics 2026: 50+ key benchmarks2026年8月8日检索自 https://www.mailneo.co/blog/email-marketing-statistics-2026

Tabular. (2026). 578 email marketing stats you can use in 20262026年8月8日检索自 https://tabular.email/blog/email-marketing-stats

Angarum Media. (2026). The 2026 email marketing benchmark report2026年8月8日检索自 https://angarummedia.com/research/2026-email-marketing-benchmark-report/

Eggknite. (2026). Email and deliverability 2026 report2026年8月8日检索自 https://www.eggknite.com/reports/email-deliverability-2026

Christopholivier Consulting. (2026). Email marketing benchmarks 2026: verified data2026年8月8日检索自 https://christopholivierconsulting.com/email-marketing-benchmarks/

本站相关阅读

我们的 small business website research covers the same causation problem in a different field. See also our page speed analysis on separating correlation from controlled tests, and the 免费试用会员目录 for testing email platforms free.

关于本次分析

This article traces a widely repeated statistic to its origin and identifies two compounding errors in how it is quoted: currency and date. The more current figure is presented with its methodological limitation, which is that attributed revenue is not incremental revenue. All the sources sell email software or services, and that is stated. The case made for email marketing deliberately does not rest on any return multiple. Checked on August 8, 2026.