Why hiring research does not explain firing
The framework page uses jobs to sharpen acquisition: an invoicing tool rewrites its homepage around the job of chasing late payers and wins more trials. That research looks at the moment of hiring. Retention needs the opposite moment, when a customer decides the product is no longer the best way to make progress. The interview subjects are different, because the useful people have already gone. The questions are different, because the story runs from settled use to doubt to switching. And the output is different: not a message, but a list of jobs the product is failing, each with someone who can fix it.
Satisfaction measures do not reach this. A customer can rate a product well and still leave because their situation changed and another option fits the new situation better. The voice of the customer guide covers how relationship and transactional surveys work; they show how people feel about the product, while switch interviews show what they did instead and why.
Decisions that change for retention
| Framework step | When jobs serve acquisition | When jobs serve retention | Decision it forces |
|---|---|---|---|
| Who is interviewed | Recent buyers about the decision to buy | Recent leavers about the decision to stop | How to recruit leavers quickly and honestly |
| What is mapped | The job and its functional, emotional and social layers | The forces: push, pull, habit and anxiety around leaving | Which force is strongest and which can be changed |
| Job inventory | Main and related jobs to message | Jobs never done versus jobs stopped being done | Whether a loss is a fit problem or a decline |
| Output | Positioning and content brief | A lever per lost job, owned by product, service or messaging | Who owns each lever and how it is measured |
| Scope | Which jobs to claim | Which jobs to stop serving | Which customers to let go |
Interview recent leavers about the switch rather than surveying satisfaction
The switch interview walks backwards from the moment of cancelling or lapsing to the first sign that something was wrong: what happened, what they tried, what they turned to, who else was involved. It avoids asking what they disliked, because that question produces opinions rather than events. The customer interviews guide covers the discussion guide and recruitment; for leavers, add a clear promise that nobody will try to win them back during the call, and keep consent and data handling to the standards set out by research bodies such as the one listed under sources.
The trade-off is depth against volume. Interviews reach a small number of people deeply, while a cancellation survey reaches many shallowly. Choose interviews to find the jobs, and use the survey later to size them.
Map the forces: push of the current product, pull of the alternative, habit and anxiety
Every switch is a balance of four forces. The push is what made the current product feel inadequate. The pull is what made the alternative attractive. Habit is the comfort of the existing routine, and anxiety is the worry about switching. For retention, habit and anxiety are allies: they explain why customers stayed through early doubts, and strengthening them is a legitimate lever. Map each leaver's story onto the four forces, then look for the force that recurs most often.
Choose to act on push first when it comes from the product failing a job; act on habit when the product works but customers never formed a routine around it.
Separate jobs the product never did from jobs it stopped doing
Some leavers hired the product for a job it was never designed to do; they were mis-sold or misread the offer. Others hired it for a job it did well until something changed: a feature removed, a price raised, a schedule moved, or the customer's own situation shifted. The first group is a positioning and onboarding problem, often solved by setting honest expectations. The second is a decline problem, usually solved in the product or service. Tag every interview with which group it belongs to before drawing conclusions, because mixing them leads to fixing the wrong thing.
Turn each lost job into a retention lever owned by product, service or messaging
A lost job becomes a lever only when someone owns it. Product owns jobs the product stopped doing. Service owns jobs lost to poor handling, where service recovery and complaint handling can win back a customer who would otherwise have gone. Messaging owns jobs the customer did not know the product could do: a welcome series that shows the job in the first days, or lifecycle email and SMS flows that prompt the habit before it lapses. The customer retention strategy guide sets out how these levers fit into a retention plan. Choose the owner by who can change the cause, not who noticed the symptom.
Decide which jobs the business will stop serving and let those customers go
Some lost jobs are not worth winning back. Serving them would pull the product away from the job its best customers hire it for, or cost more than those customers are worth. The trade-off is short-term numbers against focus: letting a group go makes the churn figure worse for a while and the product better for everyone else. Decide explicitly, write it into positioning and onboarding so new customers with that job are not recruited, and track the group separately so its departure is not mistaken for a new problem. The STP framework helps decide which segments the product exists to serve.
Decision points in a worked retention study
Lingomere finds the job its language school stopped doing
Lingomere runs online language courses with live conversation classes and self-study lessons, sold as a monthly subscription. Monthly lapse rose from 6% to 9% over two quarters. A cancellation form said "no time" for 54% of leavers. The team ran 18 switch interviews with learners who had lapsed in the previous six weeks.
| Lapsed learner group | Job they hired Lingomere for | Push | Pull | Habit and anxiety | Job status | Lever and owner |
|---|---|---|---|---|---|---|
| Evening learners, 9 of 18 | Practise speaking after work to keep progressing | Live classes moved to daytime slots after a timetable change | Free conversation-exchange apps in the evening | Habit broke when the slot went; anxiety about speaking with strangers in apps | Stopped doing | Product restores evening classes |
| Exam takers, 5 of 18 | Pass a certified language exam | No exam preparation track | Specialist exam tutors | Little habit formed | Never did | Positioning makes clear Lingomere is not exam prep |
| Travel learners, 4 of 18 | Get by on a trip in a few weeks | Course pace too slow for the date | Phrasebook apps | Planned to return after the trip | Never did, but close | Messaging offers a short travel module and a pause option |
The "no time" answers were mostly evening learners whose slot had moved. Lingomere restored two evening class times, sent lapsed evening learners a message naming the change, and reactivated 210 of 1,400 within a month. It chose not to build exam preparation, and updated its onboarding to say so. Lapse returned to 7% within a quarter, with the exam-taker group tracked separately.
The example shows how the stated reason hid the job. "No time" was true, but the cause was the product removing the time slot that fitted the learners' lives. Separating stopped jobs from never-done jobs kept the team from building exam preparation for a group it was never designed to serve.
Measuring retention work with jobs: KPIs and cadence
Measure retention by job group rather than in aggregate: lapse and reactivation rates for each group, the share of new customers who arrive with a job the product does well, and lifetime value, LTVTotal value a customer brings over time., by job. Each lever gets its own KPIA metric tied directly to an objective.: usage of the restored feature, contacts resolved, or completion of the habit-forming step in onboarding. Read lever measures monthly, lapse by job group each quarter, and rerun a small round of switch interviews every two quarters to see whether new jobs are being lost. Where retention reports are organised as a FunnelModel of the customer journey stages., add the job as a dimension so that a fall in one group is not averaged away by another. The customer journey mapping framework helps place each lost job at the moment it was lost.
Where jobs research misleads on retention
Leavers tell a coherent story after the event, and coherence can be invented. A small set of interviews can over-represent the articulate and the angry. Some leaving has nothing to do with jobs: a payment fails, a customer moves, a budget is cut. And a job the product stopped doing may have stopped because the customer's situation changed, not the product, in which case no lever will bring them back.
Common mistakes
- Trusting the cancellation form. Its reasons are the last step, not the cause.
- Interviewing stayers about leavers. Only people who left can describe the switch.
- Ignoring habit and anxiety. They are the forces that kept customers through early doubts.
- Treating every lost job as a product gap. Some customers hired the product for a job it never claimed.
- Levers with no owner. A finding nobody owns stays in the research report.
- Trying to keep everyone. Choosing which jobs not to serve is a retention decision too.
