An acquisition campaign has a natural ending. The creative work is approved, the offer runs and the results arrive in a form that can be compared with the previous campaign. A reader's relationship with a publication does not follow that timetable. The important decision may come weeks later, when the introduction ends, when the reader has had time to use the product or when an ordinary bank statement makes the continuing payment visible. I think publishers learn too little when they stop examining the relationship at the point where the campaign team has finished its work.
This does not make acquisition unimportant. A publication needs ways for people to discover what it offers, and a lower-cost introduction can be a reasonable invitation. The problem begins when the introductory response is asked to stand for a durable business outcome. A person can accept an inexpensive trial without having formed the habit or the understanding needed to continue. That distinction is uncomfortable when starts are the clearest number in the performance meeting. It is also essential if the company wants to understand the product it is actually selling.
What did the reader agree to?
I begin with the offer itself. Can a reader understand what is included, what will change after the introduction and how the continuing relationship works? This is a product question as well as a communication question. If the newsroom publishes valuable material but the subscription is difficult to explain, the difficulty may lie in the way access, newsletters, audio and local editions have been assembled. Better campaign language cannot fully resolve a confusing product. The executive team needs to be willing to examine the underlying arrangement.
At Tessara, we tested a clearer twelve-week introduction across four titles. The change reduced the discount and made the continuing price more direct. Our acquisition colleagues anticipated the immediate consequence correctly: fewer people started. The observed decline was 18%. If that number had been the entire review, the experiment would have looked like an obvious failure. It was therefore important to agree in advance which later observations would inform the decision, rather than selecting a favourable measure after the result became known.
Paid retention at week twenty-four rose from 41% to 58% in the cohorts we examined. Net subscription receipts per starter over that period increased from A$47 to A$61. I do not treat those numbers as a complete profit statement. There were fewer starters, campaign costs still existed and shared editorial expenditure had not disappeared. They told us something narrower and useful: the people entering through the revised offer showed a different continuing pattern in that observed period. That was enough to justify further work, not enough to declare that we had solved subscription economics.
Follow the cohort beyond the campaign
A cohort is useful because it keeps the original group in view. If a publisher mixes new promotional accounts with long-standing subscribers, a change in the mix can look like a change in loyalty. If it reports only the percentage retained, the reader cannot see whether a much smaller starting group has produced fewer continuing accounts. I want the underlying counts and the period beside the rate. I also want a clear definition of paid status, because a registered account or an unexpired trial can otherwise slip into a more flattering description.
The commercial discussion should include people who left. Cancellation comments are not a representative survey simply because they are available, but they can reveal a mismatch worth examining. Some of our readers wanted more material specific to their immediate area. Others used only one part of the offer and did not consider the bundle worthwhile. A clearer price did not answer those concerns. It made them easier to distinguish from dissatisfaction about the introductory arrangement, which gave publication teams a more useful question to work on.
There is a risk that the organisation turns retention into another isolated score. A team can improve a short-term measure in ways that do not strengthen the publication, just as it can improve starts without building a lasting relationship. I therefore ask what changed in the reader's experience, which periods have been observed and what work remains unmeasured. The aim is not to surround every result with so many caveats that no decision can be made. It is to keep the decision proportional to what the evidence actually establishes.
Commercial colleagues sometimes argue that a more demanding review slows experimentation. I understand that concern. A publication cannot wait a full year before making every change. We can distinguish an early signal from a settled conclusion, however, and decide what level of commitment each supports. A limited extension may be reasonable while a cohort matures. A group-wide claim about durable value requires more. The reporting language should make that distinction visible so that temporary confidence does not become an irreversible assumption in next year's budget.
Leave room for journalism that is not a conversion device
Subscription analysis becomes harmful if every editorial decision is judged as though it were a direct-response advertisement. Some reporting matters because it maintains a reliable public record, follows a subject over time or explains something readers did not know they needed to understand. It may support the publication's value without producing an identifiable conversion on the day it appears. I want editors to use audience evidence intelligently, but I do not want the commercial dashboard to become a substitute editor.
This creates a real management tension. The enterprise must fund its journalism, and editorial purpose cannot excuse an indefinitely unclear account of cost. Equally, attaching an immediate revenue requirement to each story can narrow the product until it no longer deserves a lasting relationship. The CEO's work is to help the organisation examine both obligations together. That includes discussing the overall offer, the capacity to produce it and the funding choices needed to maintain work whose value is not captured neatly by a campaign attribution model.
For me, the reader who stays is not simply a better number in a sales report. Continued payment is one form of evidence that the publication remains useful enough to choose again. We should learn from it without confusing an account with a person, a period of retention with permanent loyalty or a subscription receipt with profit. The most useful commercial conversation begins after the introductory celebration, when the organisation asks what the reader found worth returning for and whether it can keep providing that value on honest terms.