Email Analytics Tools: Beyond Your ESP’s Dashboard
What your provider will not tell you, and the three questions that need data from somewhere else.
On this page
The short version
- Three questions your sending platform cannot answer, and all three need data from another system.
- The join is the whole problem. Every analytics arrangement depends on matching a subscriber to a person in another system, usually on email address.
- Before buying anything, check whether your existing analytics tool already answers the question with a properly tagged link.
A sending platform reports on sends: delivered, opened, clicked, unsubscribed. Those are facts about the email and they stop at the moment the reader leaves it.
Three questions live beyond that boundary, and each requires joining email data to something else. That joining is the actual work, and no tool removes it.
The three questions
| Question | Needs | Usually answered by |
|---|---|---|
| What did this send earn? | Revenue joined to the click | Web analytics, or the commerce system |
| What is a subscriber worth over time? | Cohorts tracked across months | A warehouse, or a spreadsheet |
| Which acquisition route produces buyers? | Lead source joined to purchases | Both of the above, plus source at capture |
Check your analytics tool first
Most of the first question is answerable with consistent link tagging and the analytics tool you already have. Every link carrying a medium of email, a source naming the programme and a campaign that sorts chronologically gets you a channel report with revenue attached.
That covers a great deal, and it fails in the two places described in every conversion tracking guide: redirects that strip query strings, and journeys that cross a device boundary. Neither is fixed by buying an analytics product.
So the sequence is: fix the tagging, confirm the parameters survive to the destination, and only then consider whether the remaining gap justifies a tool.
Cohort value needs no tool at all, at first
The second question — what a subscriber is worth over time — is answerable in a spreadsheet for most programmes. Take everyone who joined in one month, track what that group spent over the following twelve, and divide.
Repeat it by acquisition source and you have answered the third question too, which is the one that decides where to spend on growth.
It is tedious and it is the highest-value analysis in the whole area. A tool automates it; it does not make it more true, and doing it manually once tells you whether automating it is worth anything.
What a dedicated tool actually buys
Aggregation and continuity. It maintains the joins, keeps the history, and produces the cohort view without anyone rebuilding a spreadsheet each quarter.
That is worth real money for a programme where several people need the answers regularly. It is worth very little for one where the answers are consulted twice a year, because the manual version costs a day twice a year.
The honest test: how often has someone asked one of the three questions in the last six months, and how long did answering take? If the product of those two numbers is smaller than the annual licence, the spreadsheet is winning.
Analytics check
- Every email link is tagged with a consistent scheme
- Parameters survive every redirect to the destination
- The subscriber-to-customer join is normalised on both sides
- The match rate is measured, not assumed
- One cohort analysis has been done by hand
- The frequency the questions are actually asked is known
- Attribution model and window are stated on every revenue figure
Frequently asked questions
Why do my platform's numbers disagree with my analytics tool?
They measure different things — the platform counts clicks it recorded, the analytics tool counts sessions it attributed. Both are internally consistent. Pick one as the number of record and use the other for diagnosis.
Do I need a data warehouse?
Not until the joins are being rebuilt manually often enough to be annoying. Before that, exports into a spreadsheet do the same work at a fraction of the setup cost.
How do I attribute revenue for a long sales cycle?
Session-based attribution will not do it. The email address has to appear in whatever records the sale, so the join happens on the person rather than the session.