OwnlyticsCold email tracking from send to sale, with UTM tags
To track cold email to a sale, put utm_source, utm_medium and utm_campaign on every link, with one campaign name per angle you are testing. Then compare each angle on visitors, signups and revenue per visitor, not on opens or replies alone.
Fig.Cold email sequences compared by revenue, not opens. Sample data, not a real site.
It is a small habit that turns "that email felt good" into a comparison you can act on.
Why are opens and replies not enough?
Open rates are unreliable because many mail clients preload images, and they say nothing about whether the person acted. Replies are valuable, but most buyers never reply. They click, read, and either leave or buy. The click is the first step you can attribute, and the sale is the last.
What naming scheme works for cold email?
| Field | Convention | Example |
|---|---|---|
| utm_source | where the list or sender comes from | outbound, coldlist-a |
| utm_medium | email | email |
| utm_campaign | month, segment, angle | 2026-11-agencies-pricing-angle |
| utm_content (optional) | the variant | subject-a, cta-demo |
Rules that save pain later:
- One campaign name per angle. If two emails make different promises, they get different campaigns. Otherwise you cannot say which promise worked.
- Lowercase, hyphens, no spaces.
Agenciesandagenciescan show up as two rows. - Include the segment. An angle that works for one audience can fail for another.
- Keep a one-line log of what each campaign name means and when it was sent.
A full link: https://yoursite.com/?utm_source=outbound&utm_medium=email&utm_campaign=2026-11-agencies-pricing-angle. Our UTM parameters guide covers the basics in more depth, and sources and attribution explains that UTM tags override the referrer, which is what you want here since email clients rarely send one.
How do you check that a sale will carry the tag?
Tags on the landing URL are only useful if the sale can be tied back to that visit. Click your own tagged link, go through a test purchase, and look for the campaign on the sale. Revenue attribution setup explains how sales are linked, and the Stripe Checkout UTM guide covers keeping tags through to the payment on a custom checkout.
One caveat: people often forward your email, or click days later on a different device. A purchase on another device may not be linked to the original click, and may show up as an unattributed or estimated sale. Do not expect perfect coverage.
Example (made-up numbers for illustration)
You send 300 emails to each of three angles, all to a similar kind of buyer.
| Campaign | Emails | Visitors | Signups | Sales | Revenue | Revenue per visitor |
|---|---|---|---|---|---|---|
| 2026-11-agencies-pricing-angle | 300 | 48 | 9 | 3 | $147 | $3.06 |
| 2026-11-agencies-time-saved-angle | 300 | 71 | 6 | 0 | $0 | $0.00 |
| 2026-11-agencies-case-style-angle | 300 | 30 | 8 | 1 | $49 | $1.63 |
The time-saved angle got the most clicks and no sales. The pricing angle got fewer clicks and the most revenue per visitor. If you had judged on clicks, you would have scaled the wrong email.
The decisions this triggers:
- Scale the pricing angle to a larger list, keeping the same segment.
- Rewrite or retire the time-saved angle. It attracts curiosity, not intent.
- Keep testing the case-style angle with a larger sample before judging. Thirty visitors is thin.
- Read “Setup guide”x.comNov 1 · 3 min on page
- Came back from a newsletternewsletterNov 6 · read pricing
- Opened checkoutNov 12 · 14:02
- Paid $49Dodo PaymentsCredited to the newsletter visit (last non-direct)
Fig.One buyer from a cold email. Sample data, not a real site.
What should you look at, and when?
- Within 48 hours: visitors per campaign. This is how many people clicked. If it is near zero across angles, the problem is deliverability, subject or list quality, not the offer.
- Within a week: signups and checkout starts per campaign. Goals let you set these up with a data attribute on a button.
- Within two to four weeks: sales and revenue per campaign. Cold email buyers can take a while, so see how long buyers take to purchase before declaring a loser.
Also look at the path. The Path to purchase view shows which pages buyers saw before paying. If every buyer from a campaign visited the pricing page, point that angle's link straight at it.
What decision should each number trigger?
- Clicks low: fix subject lines, sender name, list quality, or timing before touching the landing page.
- Clicks good, signups low: the email promised something the landing page does not show. Match the page to the email.
- Signups good, sales low: check pricing, trial, or onboarding, not the email.
- Sales good, volume small: scale the angle, carefully, and watch for the same revenue per visitor on the larger send.
When this does not apply
If your list is under a few dozen contacts, you cannot compare angles meaningfully, so use the tags for record keeping and rely on replies and conversations instead. Also, unsolicited email is regulated in many places. Rules vary by country and by whether you are writing to businesses or individuals. This is general information, not legal advice, so check with your own counsel or regulator before you send.
At a glance
| Field | Value |
|---|---|
| utm_source | list or sender name |
| utm_medium | email |
| utm_campaign | month, segment, angle |
| utm_content | optional variant |
| Early read | visitors per campaign, 48 hours |
| Real read | revenue per visitor, 2 to 4 weeks |
| Test first | click your own link through a test purchase |
To see campaign-level revenue on sample data, the live demo shows how it looks.
FAQ
Do UTM tags hurt cold email deliverability?
We do not claim a tag changes deliverability. Very long or heavily decorated links can look odd, so keep the tags short and the number of links low.
Should I use a link shortener?
You can, but it adds a moving part and one more thing to break. A plain link with short tags is easier to debug.
What if the buyer forwards the email?
The click still carries the tag, so a purchase on the same device and browser can be linked. A buyer who switches device may not be linked to the click at all, so treat results as a floor.
How many emails per angle do I need?
Enough to get dozens of clicks per angle. With fewer, treat differences as hints. Revenue per visitor needs more volume than click rate does.