How to get hired as a lifecycle marketing manager in 2026-27

Four jobs share this title, and the documents are not interchangeable

"Lifecycle marketing manager" describes a scope, not a job. Four distinct jobs ship under it, plus the UK and European habit of calling the same work CRM marketing, plus a handful of companies using retention marketing, engagement marketing or customer marketing for it. The stack, the metrics, the cadence of the work and the stories that land in the interview are different in each. Decide which one the posting means before you rewrite anything, because a DTC retention resume sent to an enterprise CRM team reads as junior, and the reverse reads as slow and process-bound.

The tell is in the nouns. Flows, segments, SMS, AOV and Shopify means ecommerce. Canvas, push, in-app, trial, D30 and store subscriptions means consumer app or subscription. Activation, PQL, nurture, expansion, Salesforce and net revenue retention means B2B SaaS. Contact policy, decisioning, next best action, consent, journey, and a named enterprise platform means a large regulated CRM team. If the posting names a platform, that single word tells you more about the job than the paragraph of responsibilities above it.

There is a fifth use of the title, and it is the one to catch before you accept. Some postings use "lifecycle marketing manager" for an email production role: build, QA, schedule, maintain the calendar, support the growth team. The signals are a responsibilities list with no metric in it, no mention of experimentation, a reporting line into brand or creative rather than growth or product, and no stated access to data. That job exists and can be a fine step, but it is an email marketing manager with a better title, it pays like one, and taking it while expecting ownership is how people spend two years not building the resume they need. Ask in the screen: which metric is this role accountable for, who else is accountable for it, and will I have query access to the data behind it?

• DTC and ecommerce retention. Stack: Klaviyo, Attentive or Postscript for SMS, Shopify or another commerce platform as the event source, sometimes a loyalty or reviews tool. Metrics: repeat purchase rate, second-order rate and time to second order, purchase frequency, AOV, revenue share from owned channels, discount margin given away. The work is flows versus campaigns, a promotional calendar negotiated with merchandising, and a constant argument about discounting. The hiring manager is often a Director of Retention, a head of ecommerce, or the founder.
• Consumer subscription and app lifecycle. Stack: Braze, Iterable, MoEngage, CleverTap, OneSignal or Airship across push, in-app, email and sometimes SMS; RevenueCat, Stripe Billing, Recurly or Chargebee on the billing side; Amplitude or Mixpanel for behaviour. Metrics: trial-to-paid and free-to-paid conversion, D1 or D7 or D30 retention, resurrection rate, voluntary versus involuntary churn, ARPU. Store subscription mechanics (grace periods, billing retries, the fact that Apple and Google own parts of the cancel flow) are specific knowledge employers screen for and candidates rarely have.
• B2B SaaS and product-led lifecycle. Stack: Customer.io, HubSpot, Marketo Engage, Braze or Iterable, with Salesforce as the system of record, Pendo or Appcues or Intercom for in-product messaging, and a warehouse behind it. Metrics: activation rate against a defined action, time to value, product-qualified lead volume and quality, trial and freemium conversion, net revenue retention, expansion and seat growth. Half the job is negotiation with product and sales about who owns which touch.
• Enterprise and regulated CRM. Banks, insurers, telcos, airlines, grocers, utilities. Stack: Salesforce Marketing Cloud, Adobe Journey Optimizer or Campaign, Braze, Insider, Bloomreach or Emarsys, with a real CDP and a real data team. Metrics: contact policy and frequency cap compliance, next-best-action acceptance, cross-sell and product holding, churn and save rate on retention offers, consent and preference coverage. Legal and compliance review is a scheduled dependency, release cycles are slower, and a candidate who cannot talk about contact policy and consent reads as someone who has only worked in a tool with no guardrails.

How lifecycle marketing hiring actually works in 2026-27

Nothing licenses this role. There is no board exam, no registration, no mandatory certification and no required degree, which means the screen is entirely evidence-based and the pile is enormous. Recruiters are reading for four coarse signals in the first fifteen seconds: which platform you have operated, how big a base, B2C or B2B, and whether there is a measured retention number anywhere on the page. Everything else you have written is read after those four, if at all.

The loop is normally five stages, and only two of them are conversations. A recruiter screen checks those four signals plus compensation and location. The hiring manager conversation (a Director of Lifecycle, CRM, Retention or Growth; at a startup the VP Marketing, head of growth or founder; in enterprise a CRM lead inside a larger marketing organisation) is where you walk the lifecycle you own stage by stage and get interrupted. Then a case exercise. Then a cross-functional panel, which almost always includes a product manager and a data analyst, often a customer support or CX lead, sometimes a creative or brand partner, and sometimes a lifecycle peer who will ask the most technical questions in the loop. Then occasionally a final with a VP or CMO that is about scope and judgment rather than craft.

The case exercise is the stage that decides it, and it comes in three recognisable shapes. First, the diagnosis: here are our numbers (a funnel, a cohort table, a churn breakdown), tell us where you would intervene and why. Second, the design: design the first thirty days for our product, or design the winback programme. Third, the critique, which is the most revealing and increasingly the most common: here are the three onboarding emails we actually send, tear them apart. Some companies hand you a sandbox in their real Klaviyo or Braze account for an hour.

What separates a strong case from a weak one is the order of operations, not the ideas. Strong candidates diagnose before they design: they state what the numbers do and do not tell them, name the two or three data points they would ask for first, and say which hypothesis each would kill. Then they prioritise by expected value against effort and say out loud what they are choosing not to do. Then they design the measurement before the creative: the control group, the metric, the window, what would count as a failure. Then, last and briefly, the messages. Weak candidates arrive with a beautiful journey diagram, fourteen emails, and no way to tell whether any of it worked.

Two things get a case rejected even when the thinking is good. Designing a journey the company's data cannot support, because you never asked what events exist, is the first: a journey keyed on an event nobody fires is a plan that cannot ship. The second is proposing a discount as the first lever in a business whose margin you have not asked about. Both are failures of asking rather than failures of marketing, and both are avoided by spending the first five minutes on questions.

On timing: two to five weeks end to end is typical. Small DTC brands and agencies can run the whole thing in a week and will sometimes hire off a portfolio conversation alone. Enterprise CRM teams take six weeks or more, add a panel presentation, and may be gated on a headcount approval that has nothing to do with you. Contract and freelance work is a genuine parallel market in this discipline, more than in most of marketing, because platform migrations create well-defined projects: an ESP or CDP migration, a Braze or Marketo implementation, a deliverability recovery, a lifecycle audit. Specialist retention consultancies and agency retention practices hire largely on having done one of those.

One thing to establish before you accept, because it decides whether the job is doable. Lifecycle is frequently a team of one with agency support, and the gap between owning a metric and being able to move it is engineering time and data access. Ask who fires the events, how long a request to the data team takes, whether you can query the warehouse yourself, who owns the billing emails, and what happened the last time lifecycle asked product for a change. A role accountable for activation with no route into the product is a role that cannot succeed, and the answers to those questions tell you that in ten minutes.

One thing about the market rather than a claim about it. Retention budget is easier to defend than acquisition budget because its return is measurable inside a quarter, which makes this a good title to hold and a crowded one to apply for: a large share of the applications for any lifecycle opening come from email marketing managers hoping the title transfers. It does not transfer by itself.

• Recruiter screen, 20 to 30 minutes. Screens for: platform names that match the posting, base or list size, B2C versus B2B, team size if the role manages, salary expectation, location and work pattern. Name the platform in your first sentence. Give the base size as a number.
• Hiring manager, 45 to 60 minutes. Screens for: whether you own a metric or a channel. The opening question is nearly always a version of "walk me through the lifecycle you own". Answer stage by stage, with the metric at each stage, the current value, and what you changed. Expect to be interrupted on any number you say.
• Case exercise, take-home or live, two to six hours of real work. Screens for: diagnosis before design, prioritisation, and measurement design. Ask clarifying questions in writing before you start; the questions you ask are part of what is being assessed.
• Cross-functional panel. The product manager is checking whether you will treat product problems as email problems. The data analyst is checking whether your numbers survive contact with a definition. The support or CX lead is checking whether you have ever read a complaint. Prepare a different answer for each.
• Data check, where it exists. Usually reading a cohort retention table out loud and saying what it means, or writing a query for 30-day retention or repeat purchase rate. Sometimes a spreadsheet exercise instead. Rarely algorithmic, never a whiteboard puzzle.
• Optional final, 30 minutes with a VP or CMO. About scope, ownership and how you handle being told to email the whole base. Not about craft.
• What to ask them, in every loop: which metric is this role accountable for and who else is; what share of churn is involuntary; is there a holdout today; who fires the events and how long a data request takes; what the last product change lifecycle got shipped was; whether there is an agency and what it owns; and if there is a bonus tied to retention, exactly how it is calculated.

The retention metrics you will be asked to define, and the ones that get you marked down

The most predictive thing about a lifecycle interview is whether you can define your own metrics precisely, unprompted, including the denominator and the window. Interviewers push on definitions because they are a cheap proxy for rigour: someone who says "churn was 5%" and cannot immediately say whether that is logo or revenue, monthly or annual, gross or net, voluntary or total, is someone whose results cannot be trusted either. This is the easiest part of the interview to prepare and the most commonly failed. Every number used as an example below is a shape to show the form of a good answer, not a benchmark to quote.

Start with the churn family, because that is where the confusion is worst. Logo churn counts customers lost; revenue churn counts money lost; they diverge sharply when your losses skew small or large. Gross revenue retention measures what you kept from an existing cohort and can never exceed 100%. Net revenue retention adds expansion, upsell and price increases and routinely exceeds 100% in healthy B2B, which is exactly why quoting NRR without GRR hides a leaky bucket behind expansion revenue. Voluntary churn is a customer deciding to leave. Involuntary churn is a payment failing: an expired card, a declined transaction, a hard decline on a retry. In many subscription businesses involuntary churn is a large enough share of total cancellations to be the cheapest thing on the roadmap to fix, which is why "how much of your churn is involuntary" is one of the best diagnostic questions in the discipline and a very good thing to ask the interviewer.

Then the retention curve. A cohort retention table shows, for each acquisition cohort, the share still active or still paying at each subsequent period. Two things matter when you read one: the shape of the early drop and whether the curve flattens. A curve that flattens means you have a retained core and the problem is the early period. A curve that keeps declining with no asymptote means you do not have product-market fit with that cohort and no email sequence will fix it. Being able to say which of those two situations a table shows, and what each implies for where you spend, is the most useful five seconds of analysis in the job. For apps the same thing is expressed as D1, D7 and D30 retention; for subscriptions, as M1 through M12.

Then the value metrics, where the house rule is to state your assumptions before anyone asks. Customer lifetime value is not a fact, it is a model, and the number changes entirely depending on the horizon, whether you used gross revenue or contribution margin, whether you discounted future cash flows, and whether you included refunds and returns. Say the assumptions out loud: "24-month CLV on contribution margin after returns and shipping, not discounted". Note that a platform's predicted CLV has its own fixed horizon baked in, so it is not interchangeable with yours. The same discipline applies to CAC payback period, which is the metric a CFO actually cares about and which lifecycle work directly shortens. A candidate who presents an undiscounted lifetime-revenue-times-three figure as CLV has told the room they have never had this conversation with finance.

Then the stage metrics specific to your variant. Activation is the one people get wrong most often, because activation is only meaningful when it is defined as a specific action that correlates with retention, measured within a specific window: three uploaded documents in seven days, two shifts logged in fourteen days, a first transfer in 48 hours. "Signed up and logged in once" is not an activation definition, it is a login. Interviewers ask how you chose the definition, and the right answer involves looking at what retained users did early and what churned users did not, not picking the event that was easiest to instrument.

Finally the channel metrics, including the two that died. Open rate has been unreliable for consumer email since Apple's Mail Privacy Protection began prefetching images in 2021, which inflates opens for Apple Mail users and decouples the number from human attention; image proxying and caching at other providers blunted the signal further. Click-to-open rate inherits the problem in its denominator. A candidate who in 2026 reports open rate as a success metric, or who triggers an automation off "opened but did not click", has dated themselves. You can still use opens as a weak aggregate engagement signal, directionally, for something coarse like a sunsetting policy, and you should say so with that qualification attached. What you report instead is further down the funnel: click rate on a clean denominator, conversion rate, revenue or conversions per recipient, and the incremental version of each.

The habit that holds all of this together: every number you say in an interview arrives with four things attached. The baseline it moved from, the denominator, the time window, and how it was measured. "Repeat purchase rate went from 21% to 26% over two quarters, measured as the share of first-time buyers in a monthly cohort who placed a second order within 90 days, against a 5% holdout" is a lifecycle marketer talking. "Increased repeat purchases by 24%" invites exactly the question you do not want.

• Gross revenue retention (GRR): revenue retained from an existing cohort over a period, excluding expansion. Caps at 100%. The honest measure of leakage.
• Net revenue retention (NRR): GRR plus expansion, upsell and price. Can exceed 100%. Always report it next to GRR, because NRR alone can hide heavy churn behind a few growing accounts.
• Logo versus revenue churn: customers lost versus money lost. Name which one you mean, every time. They diverge whenever customer value is unevenly distributed.
• Voluntary versus involuntary churn: a decision versus a failed payment. Involuntary churn is addressed with retry schedules, pre-dunning expiry notices, dunning sequences, a card account updater and a self-service payment-update path, and it is often unowned.
• Dunning recovery rate: the share of failed payments eventually recovered. Report it with the window and the retry schedule, because the schedule is the intervention.
• Cohort retention curve: retention by acquisition cohort over time. Read the early drop and whether it flattens. Flattening means fix onboarding; no flattening means fix the product or the acquisition source.
• Activation rate: the share of new users completing a defined action correlated with retention, inside a stated window. State the action and the window, and say how you chose them.
• Time to value: elapsed time from signup to that action. Shortening it is usually a bigger lever than adding messages.
• Trial-to-paid and free-to-paid conversion: with the trial length, whether a card was required up front, and whether you count the first successful charge or the first non-refunded one.
• Repeat purchase rate and second-order rate: the share of buyers who buy again, and specifically the share of first-time buyers who place a second order, within a stated window. Second-order rate is the sharpest early signal in DTC.
• Customer lifetime value: a model, not a fact. State the horizon, whether it is margin or revenue, whether returns and refunds are netted out, and whether it is discounted.
• CAC payback period: months to recover acquisition cost from contribution margin. The metric that justifies a lifecycle team's budget to finance.
• Incremental revenue per recipient: revenue per message recipient measured against a withheld control. The only revenue number that survives senior questioning.
• Sunset or engagement policy: the rule by which you stop messaging someone who has not engaged. Having one is a positive signal; not having one, on a large consumer base, is a deliverability problem waiting to happen.

Measurement: holdouts, incrementality, and the claim that ends interviews

There is one sentence that reliably ends a senior lifecycle interview badly: "our winback flow drove 1.2 million in revenue". It fails because the interviewer knows where that number came from. Platform attribution, whether it is Klaviyo's attributed conversion value, a Braze conversion window, Iterable's attribution settings or a last-click email credit in GA4, counts a purchase if the person received or clicked a message inside a lookback window before buying. It does not ask whether they would have bought anyway. For a winback or a cart abandonment flow, the people in the audience are self-selected toward buying: that is why they are in the audience. A large fraction of that attributed revenue is revenue you would have had with the messages switched off.

This is not a theoretical quibble, it is the core professional competence of the role in 2026, and it is where the hiring bar has moved. When every candidate can build a flow and every platform can generate the copy, what distinguishes a lifecycle marketing manager is being able to say truthfully how much money the programme makes. Employers have been burned by the shape of this: a retention team reporting a large share of total revenue as email-attributed, in a business whose total revenue did not move, is an expensive and familiar experience, and it is why holdouts are now asked about by name.

A holdout is a randomly selected group withheld from treatment, measured on the same outcome over the same period. Two forms matter. A global or programme holdout withholds a small share of the base, commonly in the low single digits up to around 10% depending on base size, from all lifecycle messaging for a meaningful stretch (a quarter is a reasonable unit) and tells you what the entire programme is worth. A campaign or journey holdout withholds a share from one specific treatment and tells you what that treatment is worth. You want both, and they answer different questions: the global one justifies the team, the campaign one directs the work.

Holdouts cost revenue in the short run, and defending one to a VP is part of the job, so have the argument ready. The framing that works: the holdout is not forgone revenue, it is the price of knowing which of our thirty journeys to keep. It is usually cheaper than one quarter of running a journey that does nothing, and it is the only way to find the journeys that are actively negative, which do exist (over-messaged winbacks that drive unsubscribes, discount flows that cannibalise full-price orders, push notifications that drive uninstalls). Say that, and say what size holdout you ran and why that size.

Know the invalid designs, because being able to name them is a strong signal and being caught using one is fatal. Pre-post comparison with no control confounds everything seasonal, promotional and macro. Comparing people who opened to people who did not, or clickers to non-clickers, is selection on the outcome and will show a gigantic fake lift every single time. Comparing a treated segment to an untreated segment that was not randomly assigned measures the segment, not the treatment. Removing people from the analysis after they were assigned breaks the randomisation, so analyse on intention to treat, including the people who never opened. And stopping a test the moment it turns positive inflates false positives; if you need to look early, use a method designed for it rather than peeking at a fixed-horizon test.

Be honest about statistical power, because that is where most real lifecycle testing quietly fails. With a list of 40,000 and a 2% conversion rate, you cannot detect a 5% relative lift; the test will come back inconclusive and someone will read the direction of the noise as a result. Calculate the minimum detectable effect before running, and if the answer is that you cannot measure what you are about to do, either change the design (test a bigger swing, measure a higher-frequency outcome, run longer, pool similar tests) or run it as a judgment call and label it as one. "We shipped it without being able to measure it, and here is why that was the right call" is a respectable answer. A confident percentage from an underpowered test is not.

If you have never had a holdout, say that plainly and then say what you would do. "We measured on platform attribution, I did not trust it, and here is the global holdout I would set up in the first month: 5% of the base withheld from all lifecycle sends, measured on revenue per user and churn over a quarter, with these known limitations." That answer beats pretending, because the interviewer's next question was always going to be about the control group.

• Global programme holdout: a small random share of the base withheld from all lifecycle messaging for a quarter or more, measured on revenue per user, retention and churn. Answers "what is this team worth".
• Campaign or journey holdout: a random share withheld from one treatment. Answers "should we keep this journey". Run it on every journey you inherit before you optimise any of them.
• Analyse on intention to treat: include everyone assigned to treatment, whether or not they opened, clicked or were even deliverable. Dropping non-openers is the single most common way a real lift becomes a fake one.
• Pre-register the metric, the window and the decision rule before you launch. Writing down what would make you kill the idea is the cheapest credibility you will ever buy.
• Compute the minimum detectable effect first. If your base and base rate cannot detect the lift you are hoping for, say so and change the design rather than running an unmeasurable test.
• Never compare openers to non-openers, clickers to non-clickers, or an engaged segment to a disengaged one, and be ready to explain why to a stakeholder who wants you to.
• Watch the negative outcomes alongside the positive one: unsubscribes, spam complaints, push opt-outs, uninstalls, discount margin given away, and cannibalisation of full-price orders. A journey can raise attributed revenue and lower total contribution.
• Keep incremental and attributed numbers in separate, labelled columns in your own reporting, so that nobody downstream confuses them and so that you never have to.

The stack: what to learn, in what order, and what moving audiences into the warehouse changed

Platform experience is a hard screen in this role, which is annoying but easy to work with because the concepts transfer almost completely and the vocabulary does not. A segment is a segment whether Klaviyo, Braze or Marketo is drawing it; what differs is whether it is called a segment, an audience or a smart list, whether journeys are canvases, flows or programs, and what the templating language is. Recruiters screen on the noun. So if you have run Iterable and the posting says Braze, write a line that maps your experience explicitly: the concept, then their word for it. Do not claim daily use of a platform you have only watched a demo of, because panel interviewers ask platform-specific operational questions (how do you handle a user who qualifies for two journeys at once, what happens to someone mid-canvas when you edit it) that cannot be bluffed.

Learn it in this order. First, one execution platform to real depth, chosen to match the variant you are targeting. Second, SQL, to the level of writing a cohort retention query and a repeat-purchase query against a real schema, with joins, window functions and date arithmetic. Third, the event and identity model: what events your product fires, what properties they carry, how an anonymous visitor is stitched to a known user, what happens when the same person has two records. Fourth, deliverability and consent mechanics. Fifth, experiment design. Templating (Liquid or Handlebars), responsive HTML email and basic rendering QA sit alongside the first item and are assumed rather than impressive.

SQL is now the dividing line, and it is worth being blunt about why. The job used to be to request a list from analytics, build the campaign, and report what the platform said. The teams hiring now expect the lifecycle owner to pull their own cohort, check their own result, and notice when a number looks wrong, because the alternative is a week of queue time per question and a programme that is never measured. You do not need to be an analytics engineer. You need to be able to write a query with a join and a date window without help, read someone else's query without being frightened of it, and know what a cohort table is doing. That is a few weeks of deliberate practice, and it changes which jobs you can apply for.

The structural change in this discipline over the last few years is that audience building moved out of the email tool and into the data warehouse. The old pattern was a customer data platform (Segment, mParticle, Tealium, RudderStack) collecting events and piping audiences into the ESP, with the ESP's own segment builder as the place where logic lived. The pattern now common, often called composable or warehouse-native, is that Snowflake, BigQuery or Databricks holds the customer model, dbt defines the metrics and the audience logic as version-controlled SQL, and a reverse ETL tool (Hightouch, Census) syncs those audiences and traits into Klaviyo, Braze, Iterable, Salesforce Marketing Cloud or the ad platforms. Both patterns are live in the market and plenty of companies run a hybrid.

Why this matters to you specifically: it changes what a senior lifecycle marketer is expected to be able to do. In a warehouse-native team, "I can build any audience I need inside Klaviyo" is a limitation rather than a skill, because the audience that matters depends on a margin calculation, a subscription state and a support-ticket history that never reach the ESP. The candidate who can say "the segment is defined in dbt, I wrote the model with the analytics engineer, it syncs through Hightouch, and I can tell you exactly which fields the ESP sees" is operating a level up, and that is the level the better-paid version of this job sits at. If your experience is entirely inside one ESP's own builder, learning this pattern, even on a free-tier warehouse against a sample dataset, is the highest-leverage week you can spend.

Know the ecosystem by category rather than memorising logos, and know which categories a given employer owns. Execution and orchestration (Klaviyo, Braze, Iterable, Customer.io, Salesforce Marketing Cloud, Adobe Journey Optimizer, Marketo Engage, HubSpot, Insider, Bloomreach, Emarsys, MoEngage, CleverTap, Airship). SMS (Attentive, Postscript, Twilio). The data layer (Segment, mParticle, RudderStack, Snowflake, BigQuery, dbt, Hightouch, Census). Product analytics (Amplitude, Mixpanel, GA4, plus Looker or Mode for reporting). In-product messaging (Pendo, Appcues, Intercom). Subscription and billing (Stripe Billing, Recurly, Chargebee, RevenueCat, and the payment-recovery specialists). Deliverability and QA (Google Postmaster Tools, Microsoft SNDS, Litmus or Email on Acid). Experimentation (Statsig, Eppo, GrowthBook, plus the platforms' own test tooling).

A migration is the single most valuable concrete project on a lifecycle resume, and worth volunteering for if one is going. Moving an ESP or standing up a CDP forces you to document every journey, audit the data model, rebuild logic you inherited without understanding, negotiate scope with engineering, warm new sending infrastructure and hold deliverability steady through the switch. It is a contained, datable, unarguable piece of work with a before and after, it is the standard brief in the freelance market, and it is the kind of thing a hiring manager will spend ten minutes asking about because it reveals how you work.

• Pick one execution platform matched to your target variant and learn it to operational depth: Klaviyo for DTC, Braze or Iterable for consumer subscription and apps, Customer.io or HubSpot or Marketo Engage for B2B, Salesforce Marketing Cloud or Adobe Journey Optimizer for enterprise.
• Get SQL to the point where you can write a 30-day cohort retention query and a second-order-rate query unaided, against a schema you have not seen before. This is the most common hard gate in the loop after platform names.
• Learn your product's event and identity model: which events fire, what properties they carry, how anonymous and known identities are stitched, and what a duplicate profile does to a journey. Journey bugs are nearly always identity or event bugs.
• Learn the warehouse pattern even if your current employer does not use it: a warehouse, dbt for the audience logic, reverse ETL into the ESP. Free tiers and sample datasets are enough to be able to speak to it honestly.
• Be able to name what you personally built: number of journeys live, the trigger for each, the suppression and exit rules, frequency caps, and who owns the template library. "Managed Braze" means nothing; "owned 34 live canvases including the dunning sequence, with a global frequency cap of three per week and a 90-day sunset" means something.
• Treat a platform migration as a career artefact. If one is happening near you, get onto it; if you are freelancing, it is the brief that gets commissioned most.

Deliverability, consent and cancel flows: the part of this job that can actually be checked

Lifecycle marketing is one of the few marketing disciplines with genuine external constraints, enforced by parties who do not care about your campaign calendar: mailbox providers, mobile carriers, app stores and regulators. Candidates who know this surface stand out immediately, because most do not, and the cost of not knowing it is the kind of failure that makes a company distrust the whole channel.

The mailbox-provider rules tightened materially and recently, and they are now settled conditions rather than advice. Google and Yahoo's requirements for bulk senders took effect in February 2024: senders above roughly 5,000 messages a day to their consumer domains need SPF and DKIM authentication plus a DMARC record on the sending domain, one-click unsubscribe implemented in the headers to the RFC 8058 standard and honoured within a couple of days, and a reported spam rate kept below 0.3% as measured in Google Postmaster Tools, with Google advising you stay well under it. Microsoft followed with equivalent authentication requirements for high-volume senders to its consumer domains (outlook.com, hotmail.com, live.com, msn.com), enforced from May 2025, where failing mail is junked and can be rejected outright at the SMTP level. Verify the current specifics in each provider's postmaster documentation rather than quoting them from memory; treat the direction as settled.

The practical consequences for how you work are concrete. List growth that adds unengaged addresses is now actively harmful rather than neutral, because complaint rate is a ratio and the denominator does not protect you. A sunset policy (stop mailing people who have not engaged in N days, with N chosen from your own purchase or usage cycle rather than copied from a blog) is a deliverability control, not a vanity exercise, and having one is a positive interview signal. Sending infrastructure needs warming when it changes, which is why migrations are risky. And someone needs to be watching Google Postmaster Tools and Microsoft SNDS weekly; if you have been that person, say so, and say what you did the time the complaint rate spiked.

Apple's Mail Privacy Protection, introduced in 2021, prefetches images for Apple Mail users, which is why open rates are inflated and decoupled from attention. Apple went further with Apple Intelligence, which summarises emails in the inbox list and in notifications for users who have it enabled, and which has been standard on supported devices since iOS 18. For those users a machine-written summary may be what they see instead of the preview text you wrote. The practical response is unglamorous: write subject lines and opening content that survive being summarised, do not hide critical information in a preheader, and check how your sends render in that context rather than assuming.

Consent rules differ by channel and jurisdiction, and this is a place to know the shape precisely and the detail cautiously. US email runs on an opt-out regime under CAN-SPAM: honour unsubscribes promptly (the statute allows up to ten business days), include a valid physical postal address, and do not use deceptive headers or subject lines. Canada's CASL is far stricter, requiring express or carefully documented implied consent with records to prove it, and it attaches liability to officers. In the EU and UK, email marketing generally requires consent under GDPR and the ePrivacy rules, with a limited soft opt-in for existing customers that varies by member state. Preference centres, consent capture provenance and suppression hygiene are operational work that falls to this role, and in enterprise CRM teams they are a large part of it.

SMS in the US is the highest-risk channel you will touch, and the asymmetry is worth internalising: the TCPA carries statutory damages per message and attracts class actions, so an SMS consent mistake is a legal event rather than a marketing one. Marketing texts require prior express written consent, A2P traffic must be registered for 10DLC with the carriers via The Campaign Registry or it gets filtered, and carrier content filtering will block some categories regardless of consent. The rules here keep moving: the FCC's one-to-one consent rule, due to start in January 2025, was vacated by a federal appeals court that month; most of the FCC's consent-revocation rules took effect in April 2025, while the piece requiring a revocation sent through one channel to stop all messages has been repeatedly waived and is not currently in force. The honest position in an interview, and in practice, is that you know consent and revocation are tightly regulated and still in flux, you know which reviews and which counsel you would check with, and you do not improvise. Saying that is a better signal than reciting a rule that may have changed.

Cancel flows and save offers have their own regulatory surface, and it is the one retention marketers most often miss. The FTC's negative-option rule, the "click to cancel" rule, was vacated by a federal appeals court in July 2025 on procedural grounds, and the Commission has since moved to restart that rulemaking. That is not a licence to add friction: the FTC can still act on deceptive or unfair subscription practices under existing authority, state automatic-renewal laws (California's among the strictest) govern renewal notices and how easy cancelling has to be, and on mobile the store controls the cancel path anyway. If you are asked to design a save flow, the right instinct is to ask which jurisdictions the base sits in and who signs off, not to copy a dark pattern you saw somewhere.

Push and in-app have their own version of the same discipline. Push permission is a scarce, one-shot resource on iOS and a revocable one everywhere, and the cost of over-messaging is an opt-out or an uninstall, both harder to reverse than an email opt-out. A candidate who talks about push volume without talking about opt-out rate and uninstalls has not run push at scale. A single cross-channel contact policy, frequency caps and a view of total messages per user per week are the controls, and they are standard questions in enterprise CRM loops.

None of this requires you to be a lawyer or a mail administrator. It requires you to know which constraints exist, which ones you own, which ones you escalate, and what evidence has to exist before a send goes out. It is the easiest place to be visibly more serious than the other candidates.

• Authentication: SPF, DKIM and DMARC on the sending domain, with DMARC moving from monitoring to enforcement over time, plus BIMI if the brand wants the logo. Know whether you own this or your platform and IT do.
• One-click unsubscribe in the headers (RFC 8058) for bulk consumer sending, honoured quickly, in addition to the visible link.
• Complaint rate: watch it in Google Postmaster Tools and Microsoft SNDS, keep it far below the 0.3% threshold, and know what you would cut first when it rises.
• A written sunset policy with a threshold derived from your own purchase or usage cycle, plus a re-permission path for the people it suppresses.
• Suppression and hygiene: hard bounces removed immediately, role addresses and known spam traps excluded, and a rule against importing a list you did not collect.
• Consent provenance: for every subscriber, where and when they consented and to what. Without it you cannot defend a CASL or GDPR question, and you cannot safely migrate platforms.
• SMS in the US: prior express written consent, 10DLC registration, carrier filtering, documented revocation handling, and legal review before any change to the consent flow.
• Cancel and renewal mechanics: who approves the save flow, what renewal notices the relevant state laws require, and the fact that on iOS and Android the store owns the cancel path.
• Push and in-app: treat permission as scarce, track opt-out and uninstall alongside conversion, and enforce a cross-channel frequency cap so email, push and SMS do not each independently decide a user can take three messages.

From email or CRM execution to lifecycle ownership, and the resume that shows it

If you are an email marketing manager, a CRM executive or a marketing automation specialist trying to make this step, the gap is specific and it is not seniority. It is five things: you own a channel rather than a metric, you report attributed rather than incremental results, you cannot pull your own data, you have no evidence of influencing anything outside your own channel, and you have never documented an end-to-end lifecycle. Each of those is closable inside your current job, mostly without permission, and closing two of them is usually enough to change which interviews you get.

The highest-yield move, in any business with subscriptions or repeat purchase, is to take ownership of involuntary churn. Failed payments, retry timing, pre-expiry card notices, the dunning sequence, the self-service payment-update path and the card account updater sit in the gap between billing engineering, finance and marketing, which in practice means nobody owns them end to end. The work is unglamorous and the result is cash: recovered payments, measurable against a baseline, with a control group that is easy to justify because nobody can claim the messages are brand-building. It is also a story that travels, because every subscription business has the same problem and most handle it badly. Go and find out what share of your cancellations are payment failures. If nobody knows, you have just found your project.

The second move is a lifecycle audit of a product you do not work for, written up and published. Pick a company you would like to work for, or one in the same category. Sign up with a fresh address and a real intent, then go quiet, then abandon a cart or a trial, then churn. Capture every message with timestamps and channels, map it as the journey it is, and write the diagnosis: which stage has no messaging at all, where the sequence contradicts itself, where the timing is wrong for the purchase cycle, what they are clearly measuring, what they appear not to measure, and the three changes you would make first with how you would measure each. It takes a weekend, it demonstrates every skill the case exercise tests, and it beats a folder of email designs as an answer to "do you have a portfolio".

The third move is to run one experiment properly, end to end, in your current job: a written hypothesis, a pre-registered metric and window, a randomly assigned control, a power check, and a result you accept even if it is negative. One such experiment, described precisely, is worth more in an interview than three years of shipping campaigns, because it is the thing the role is actually being hired for. If you can only get one, make it a journey holdout on something your company believes in, because a credible "we found this flow was worth less than we thought" is a memorable interview story.

Then the applying itself, which has a shape worth following rather than spraying. Read thirty real postings in your target market and variant before you write anything, and tally which platforms, metrics and seniority words recur; that tally is both your keyword list and your reading of the market, and it takes an hour. Source openings from the careers pages of the ten companies whose lifecycle programme you can already describe, the DTC and retention agency boards, the job boards and partner directories attached to the platforms, the retention communities and newsletters where hiring managers post first, and recruiters who specialise in CRM and growth. Treat the contract market as a parallel route rather than a fallback: migrations, deliverability recoveries and audit briefs are scoped, paid work, and one finished migration is a better interview asset than six months of applying.

Then apply like someone who has already done the work. Send the audit of their product with the application, as a short note naming the three things you would do first and how you would measure each, rather than a cover letter restating the resume, and use a referral wherever one exists, because the pile is deep. A workable four weeks: the thirty-posting tally and the resume rewrite below; SQL practice to cohort-query standard; the audit of one target company; then ten applications with the audit attached, and one new audit a week for as long as the search runs.

Then rewrite the resume around metrics rather than channels, which is mostly an act of deletion. A lifecycle resume has one job: to make a stranger believe you have owned a measurable part of a customer base at a scale and in a context close to theirs. Verb lists fail at this, and most email and CRM resumes are verb lists: managed the email calendar, crafted compelling copy, collaborated with stakeholders, grew the list, executed campaigns. Those sentences are true of everyone with the title, so they carry no information and the screen skips them.

Replace them with a fixed fact block per role, in the same order every time so a reader can compare: what the business was and who the customer is, the base size, the platforms, the channels you owned, the stages you owned, and then the metrics you moved with baselines and measurement method. Every number in the examples below is a shape, not a benchmark. Use your own, and expect to be asked how each was calculated, including any you borrow from an article like this one.

Two notes on honesty, because this role gets caught more than most. First, if you cannot disclose a base size or revenue figure, give the shape and say why: "mid-seven-figure monthly revenue, DTC apparel, client under NDA", "a base in the low millions of monthly actives". Bands are fine; vagueness that hides whether the base was five thousand or five million is not, because scale is the first thing screened and a missing number is read as a small one. Second, do not claim a result you measured on platform attribution as an incremental result. Label it: "platform-attributed" or "measured against a 5% holdout". The candidate who labels their numbers is trusted on all of them.

• Lead each role with the base and the ownership, in units a stranger can calibrate: "DTC skincare, 1.4M email subscribers and 310K SMS, Klaviyo and Attentive, owned welcome through winback plus the promotional calendar" or "B2B workflow SaaS, 68K free accounts, Customer.io and HubSpot with Salesforce as source of truth, owned trial activation and expansion nurture".
• One metric per line with a baseline, a window and a measurement method. "Second-order rate 19% to 24% over two quarters, measured as first-time buyers placing a second order within 90 days, against a 5% holdout" beats "improved retention by 26%" in every room that matters.
• Put the involuntary churn work in writing if you have it: "rebuilt the dunning sequence and retry schedule with billing; share of failed payments recovered up from one figure to another, worth a stated amount per month, measured against the prior quarter's recovery curve". Hiring managers read this line twice.
• One decision per line, with the option you rejected named. "Cut the welcome series from seven emails to four and moved the discount from email two to email five, because the early discount was cannibalising full-price first orders; full-price order share up, revenue per new subscriber flat." This sentence type is rare and disproportionately persuasive.
• Name the experiment, not just the win. "Ran a programme-level holdout at 5% for one quarter; found two of eleven journeys were not measurably incremental and retired them, reclaiming send volume and cutting complaint rate." Retiring a journey is a senior act.
• Show cross-functional reach, with the artefact. The product change you got shipped because your data showed activation was gated on a step nobody could complete. The support themes you turned into a message. The pricing or billing change you argued for. Lifecycle candidates who have only ever changed emails cap out.
• A technical line that is specific: SQL (and against what: Snowflake, BigQuery, Redshift), dbt if you have used it, reverse ETL if you have, Liquid or Handlebars templating, the CDP, the analytics tool, and what you personally built in each.
• Deliverability and consent, briefly and concretely: DMARC enforcement you took part in, a sunset policy you wrote and what it did to complaint rate, a migration you warmed, a preference centre you specified, a consent model you worked within.
• Platform migrations with dates and scope: "migrated 1.1M subscribers and 34 journeys from one ESP to another in nine weeks, zero sending gap, deliverability held at or above the prior quarter's inbox placement".
• What to cut: open rates, "crafted compelling copy", design-tool lists, "grew the list to 500K" with no engagement figure, campaign counts with no outcome, and any adjective about yourself.

Pay, levelling, and the interview: the questions and what gets people rejected

On pay, resist the aggregator averages, because this title spans an email production role at a 40-person DTC brand and the owner of a bank's entire customer communications programme, and an average across those tells you nothing. Build the number from three sources. The BLS OES codes named above, read at state and metro level rather than nationally, because the metro spread matters more than the median. Posted ranges under pay-transparency law, filtered to your metro and your variant, read as thirty real postings rather than as somebody's summary of them. And for technology companies, self-reported compensation sites, treated as directional and biased upward, and never quoted back in a negotiation as if authoritative.

Then level yourself honestly, because the variance within the title is driven by scope rather than years. The factors that actually move the band: base size and whether it is consumer-scale; whether the programme's revenue is material to the company; how many channels you own (email only is the bottom of the range, email plus SMS plus push plus in-app is the middle, add paid audiences or a decisioning engine for the top); whether you own the platform contract and budget; whether you manage people or agencies; and whether you are accountable for a company-level metric or for shipping a calendar. In B2B, complexity and sales alignment substitute for raw scale. Enterprise CRM in regulated industries pays for the compliance and stakeholder load, not the craft.

Two structural notes. Agency and consultancy retention work usually pays less in base than the in-house equivalent at the same level, and buys you breadth and migration experience faster, which is a reasonable trade early and a poor one late. And if a lifecycle role offers a variable component tied to churn or retained revenue, ask exactly how it is calculated before you accept: tied to attributed revenue it is a bonus you cannot control and could inflate, which is bad in both directions; tied to a holdout-measured or company-level metric it is defensible.

Now the interview. It is less structured than a software loop and more structured than most marketing loops, and it tests four things in roughly this order: whether you own a metric or a channel, whether your definitions hold, whether your results survive a measurement question, and whether you can work with product and data without turning every problem into an email. Prepare three stories at different scales (one programme you built or rebuilt, one experiment with a clean result, one failure you caught and corrected) and make sure each has numbers, a baseline and a measurement method attached.

The questions below are the ones that come up repeatedly. Rehearse the definitional ones out loud, because the failure mode is not ignorance, it is being imprecise under mild time pressure and watching the interviewer write something down.

Finally, the rejections. The most common are not knowledge gaps. They are: presenting attributed revenue as incremental and then defending it when pushed; being unable to state your own baselines, which reads as never having owned the number; proposing more volume as the answer to a retention problem; treating an activation problem that is clearly in the product as something a three-email sequence will solve; showing no awareness that deliverability and consent exist; and talking about customers entirely as segments, with no evidence of ever having read a support ticket, a cancellation reason or a reply. Interviewers in this discipline want someone who is both quantitative and actually curious about the person on the other end, and candidates tend to arrive with one of those and not the other.

• "Walk me through the lifecycle you own, stage by stage." Answer with the stage, the metric, the current value and the last thing you changed. This is the question the whole loop hangs on.
• "How is your activation defined, and how did you choose that definition?" The right answer involves looking at what retained users did early, not at which event was easiest to instrument.
• "Our trial-to-paid is 18%. Where do you look first?" Diagnose before designing: cohort the trials by source and by whether they hit the activation action, and find out whether the problem is the trial or the traffic.
• "How do you know your winback worked?" The holdout question. If your answer does not contain a control group, you have answered no.
• "What is the difference between gross and net revenue retention, and which would you show the board?" Both, next to each other, and say why.
• "How much of your churn is involuntary?" If you do not know for your current job, say so and that it is the first thing you would measure, then describe the dunning work.
• "How do you decide send frequency, and what is your sunset policy?" Tie both to complaint rate and to your own purchase or usage cycle, not to an industry benchmark.
• "Deliverability drops next Monday. What do you do in the first hour?" Check authentication and the postmaster tools, identify what changed in the last send (volume, list source, content, infrastructure), pause the riskiest sends, segment back to the engaged core.
• "How would you measure whether SMS is worth adding on top of email?" An incrementality test on top of the existing email programme, not a comparison of SMS revenue to email revenue.
• "A VP wants to email the whole base about a launch. What do you say?" The answer that gets offers: yes to the launch, no to the whole base, here is the engaged segment and the frequency cap, here is the complaint-rate risk, and here is the holdout that will tell us what it was worth.
• "Which metric would you stop reporting?" Open rate, with the Apple Mail Privacy Protection reasoning, and what you would report instead.
• "Tell me about a result you later found was wrong." Have one. The candidates who do not are either inexperienced or not paying attention.
• A cohort table read aloud, or a SQL query for 30-day retention or second-order rate. Practise against a real schema until you can do it while talking.
• The case: a journey design, a funnel diagnosis, or a critique of their real onboarding emails. Diagnose, prioritise, design the measurement, then write the messages, in that order, and say what you chose not to do.

What a screen is looking for

These are the terms that a resume screen, human or automated, is matching against for this role. Use the ones that are true of you, in the words the posting uses.

• Lifecycle marketing
• Lifecycle marketing manager
• CRM marketing
• Retention marketing
• Email marketing
• Marketing automation
• Customer lifecycle management
• Journey orchestration
• Customer journey mapping
• Onboarding journey
• Welcome series
• Activation
• Time to value
• Trial-to-paid conversion
• Free-to-paid conversion
• Churn reduction
• Voluntary churn
• Involuntary churn
• Dunning
• Payment recovery
• Card account updater
• Winback campaign
• Reactivation
• Cross-sell
• Upsell
• Expansion revenue
• Customer lifetime value (CLV)
• LTV:CAC
• CAC payback period
• Net revenue retention (NRR)
• Gross revenue retention (GRR)
• Logo churn
• Cohort retention analysis
• Retention curve
• D1 D7 D30 retention
• Repeat purchase rate
• Second-order rate
• Purchase frequency
• Average order value (AOV)
• ARPU
• MRR
• ARR
• RFM segmentation
• Predictive CLV
• Churn propensity model
• Uplift modeling
• Incrementality testing
• Holdout group
• A/B testing
• Experiment design
• Minimum detectable effect
• Segmentation
• Personalization
• Dynamic content
• Triggered campaigns
• Behavioral triggers
• Event-based messaging
• Push notifications
• In-app messaging
• SMS marketing
• 10DLC registration
• TCPA compliance
• CAN-SPAM
• CASL
• GDPR
• Consent management
• Preference center
• Sunset policy
• Suppression list
• List hygiene
• Deliverability
• Sender reputation
• SPF
• DKIM
• DMARC
• One-click unsubscribe (RFC 8058)
• Google Postmaster Tools
• Microsoft SNDS
• Inbox placement
• Spam complaint rate
• Email rendering QA
• Frequency capping
• Contact policy
• Next best action
• Klaviyo
• Braze
• Iterable
• Customer.io
• Attentive
• Salesforce Marketing Cloud
• Adobe Journey Optimizer
• Marketo Engage
• HubSpot
• MoEngage
• CleverTap
• Segment
• mParticle
• RudderStack
• Hightouch
• Census
• Snowflake
• BigQuery
• dbt
• SQL
• Amplitude
• Mixpanel
• Looker
• GA4
• Liquid templating
• HTML email
• Stripe Billing
• Recurly
• Chargebee
• RevenueCat
• Pendo
• Appcues
• Intercom
• ESP migration
• CDP implementation
• Revenue per recipient
• Cross-functional collaboration with product and data

Mistakes that cost people this job

Questions people ask