The First 90 Days in Customer Success
The 30-60-90 plan that wins a final round, and the different, harder one you actually run after joining. I have watched hundreds of my own hires succeed or stall in their first quarter, and the difference is almost never talent. It is what they chose to earn in each of the three phases.
In This Guide
Every 30-60-90 template says the same three things: learn in the first month, contribute in the second, lead in the third. It is the plan equivalent of wishing everyone well. It cannot fail, it cannot be checked, and it tells the hiring manager nothing about how you will actually behave when you are handed a book of accounts with three renewals already on fire.
A plan worth writing names what you will have earned at each gate: context by day 30, a visible win by day 60, a running system by day 90. Earned is the operative word, because each phase produces something another person can inspect. This is the standard I hold my own hires to, and the one I hand them on day one.
Why Most 30-60-90 Plans Say Nothing
Learn, contribute, lead is a horoscope, not a plan.
The generic plan fails for one structural reason: none of its phases has an exit condition. "Learn the product" is never done, so nobody can say you did not do it, so it commits you to nothing. The fix is to replace activities with artifacts. Not "learn the customers" but "produce a one-page read on my book: every account, its renewal date, its last health movement, and the one thing I do not yet understand about it." The first is a mood. The second can be put in front of your manager on day 30, and either it exists or it does not.
A phase without an exit condition is not a phase. It is a month with a label on it.
What Each Phase Is Actually For
Context, then a win, then a system. In that order, without skipping.
Understand before you touch anything
The goal is to know your book and the machine around it better than anyone expects a new person to. Read every account's history before you meet it. Sit in on support queues and listen to recorded calls, because the tickets tell you what customers actually struggle with, which rarely matches what the onboarding deck says. Map the lifecycle as it really runs, not as the playbook describes it, and write down every gap between the two without fixing any of them yet.
The exit artifact: a one-page book read (every account, renewal date, risk level, open question) and a private list of the five things that seem broken. Private matters. Month-one diagnoses are half wrong, and the ones that survive to day 45 are the ones worth saying out loud.
One visible fix, finished, with a before and after
Not five initiatives. One. Picked with the criteria in chapter 05: small blast radius, short feedback loop, visible to someone senior. A stalled onboarding unstuck, an at-risk account stabilized with a documented save play, a renewal rescued by starting it ninety days early instead of thirty. Finish it completely and write down what the before state was, because in sixty days nobody will remember and the delta is your credibility.
The exit artifact: the win, done, plus the measured delta, communicated in one short written update. Not a deck. A paragraph with numbers in it.
Leave something running that does not need you
This is the fifth beat from the hiring playbook, applied to your own onboarding: take the win from phase two and turn its mechanism into something that runs without your attention. The save play becomes a documented trigger and checklist the whole team uses. The early renewal start becomes a calendar rule, not a personal habit. The point is that by day 90 your value is no longer only your effort, and everyone senior can see it.
The exit artifact: one process, documented and adopted by at least one person who is not you.
The Interview Version
The same plan, presented with one crucial disclaimer.
When a final round asks for your 90-day plan, they are not grading the plan. They are checking whether you think in outcomes and whether you have the humility to know the plan is provisional. So present the three-gate structure above, made specific to their business from your research, and end with the sentence that separates serious candidates from template users: "this is built from the outside, and I would revise it in week two once I have seen real data."
That sentence matters because the biggest risk a hiring manager prices into a senior hire is the person who arrives with answers before questions. The caveat shows them you will not be that person, while the specificity shows you did the work. Both signals, one page. If you also sent a proof-of-work artifact with your resume, the 90-day plan is its natural sequel, and the two together carry the final round.
Days 1 to 30, Week by Week
The context month, scheduled, because unscheduled learning becomes meetings.
| Week | What to do, and what it produces |
|---|---|
| Week 1 | Access, tooling, and the paper trail: read your accounts' contracts, past QBRs, escalation history and renewal notes before meeting anyone. Meet your manager and agree what day-30, day-60 and day-90 success looks like, in writing, because the plan you were hired against and the one your manager carries are not always the same document. |
| Week 2 | The machine: sit with support for half a day, listen to five recorded customer calls, shadow one onboarding and one renewal conversation. Map the lifecycle as actually run. Produces the first draft of your gap list. |
| Week 3 | The customers: introduction calls with every account above your risk line, listening ratio four to one. Ask each what almost made them leave, and what they would fix first. Produces the real health picture, which will disagree with the health score somewhere. |
| Week 4 | The synthesis: the one-page book read, the five-item private gap list, and the shortlist of three candidate wins for phase two. Review all three with your manager; let them veto, because they know where the political tripwires are buried. |
Picking the First Visible Win
Three filters, applied in order. Most candidate wins fail the first one.
| Filter | What it rules out |
|---|---|
| Small blast radius | Anything that requires another team to change first. Your first win must be deliverable with the authority you already have, because you have no credit to spend yet on cross-functional persuasion. |
| Short feedback loop | Anything whose result shows up next quarter. NRR is the wrong first target precisely because it moves slowly; a stalled onboarding or an escalation backlog shows a delta in weeks. |
| Visible one level up | Anything only you will notice. The win should be legible to your manager's manager without explanation, which usually means it touches an account or a number they already watch. |
A test that rarely fails
Ask your manager in week one: "what is the thing everyone knows is broken but nobody has had time to fix?" The answer is usually your phase-two win: pre-validated as real, pre-approved as wanted, and pre-loaded with an audience that will notice when it is fixed.
The Day-90 Scorecard
Grade yourself before anyone else does, with questions that have yes-or-no answers.
By day 90, is each of these true?
- I can describe every account in my book in two sentences each: where they are, what they are at risk of, what happens next.
- One thing is measurably better than it was at day 30, and I have the before number written down.
- One process I built is being run by someone who is not me.
- My manager has received a short written update from me every two weeks without asking for it.
- I know which of my month-one diagnoses turned out wrong, and I have said so out loud to the person I said it to.
- The three customers most likely to churn this year know my name and have my direct line.
Six yeses is an exceptional first quarter. Four is solid. Fewer than three means the next month is about focus, not effort, because the usual cause is having started eight things rather than finishing two.
How First Quarters Go Wrong
From watching hundreds of my own hires: four patterns account for nearly all of it.
Rebuilding on arrival. The new hire who announces in week two that the health score is wrong and the playbooks need rewriting is usually right, and it does not matter. Diagnosis without context reads as arrogance, and the organization quietly stops helping. The same observations, delivered at day 45 with evidence, land as insight.
Hiding in the learning. The opposite failure: still "ramping" at day 75, no win attempted, perfectly pleasant, invisible. Learning without an artifact is indistinguishable from not learning, and quiet first quarters are how strong hires end up on performance plans that surprise everyone.
Promising the metric you cannot move. Committing to lift NRR by quarter one is committing to weather. Renewal cycles are long and your influence lands late; promise the leading indicators you actually touch, time to value, escalation resolution, early-warning coverage, and let the lagging number follow.
Treating the book as a queue. Responding well to whatever arrives feels productive and cedes the quarter to chance. The accounts that decide your year are usually quiet, and the ones making noise are often already decided. Phase one exists so you work from the map, not the inbox.
Variants: IC, First CS Hire, Leader
Same three gates, different artifacts.
| Situation | How the gates change |
|---|---|
| IC CSM joining an existing team | The plan above as written. Your main added risk is inheriting a book mid-crisis; if so, the phase-two win chooses itself, and phase one compresses to two weeks for everything except the burning account. |
| First CS hire at a startup | There is no machine to learn, so phase one maps the customer base and the founder's expectations instead, which are usually three jobs wearing one title. Your day-60 win is making one lifecycle stage repeatable, almost always onboarding. Your day-90 system is the first playbook, however thin, because you are not building your workload, you are building the function. The playbook generator exists for exactly this seat. |
| Leader taking over a team | Phase one adds the team itself: one-on-ones with every member, and the question "what would you fix first" asked of each. Your win is removing one thing the whole team hates, fast, because it buys the trust the bigger changes will spend. Your system is the operating cadence: the forecast, the escalation path, the weekly rhythm. Resist reorganizing anything for the full 90 days. |
The Bottom Line
Context, then a win, then a system, each with an artifact someone else can inspect. That is the whole plan, in the interview and in the job. The candidates who present it win final rounds because it is checkable, and the hires who run it succeed because by day 90 their value is no longer only their effort. Learn, contribute, lead promises a mood. This promises evidence.
Questions This Article Answers
What should a 30-60-90 day plan for Customer Success contain? +
Three gates, each with an artifact someone else can inspect: context by day 30, meaning a one-page read of every account with renewal dates, risk and open questions; one visible win by day 60, finished, with the before and after measured; and a running system by day 90, one process documented and adopted by at least one person who is not you.
What is wrong with the learn, contribute, lead template? +
None of its phases has an exit condition. Learn the product is never done, so nobody can say you did not do it, so it commits you to nothing. Replacing activities with artifacts fixes this: not learn the customers, but produce a one-page book read by day 30. Either the artifact exists or it does not.
How do I pick my first win in a new Customer Success job? +
Three filters in order: small blast radius, so it is deliverable with authority you already have; short feedback loop, so the delta shows in weeks rather than quarters, which rules out NRR as a first target; and visible one level up, so your manager's manager notices without explanation. The shortcut is asking your manager in week one what everyone knows is broken but nobody has had time to fix.
How should I present a 90-day plan in a final interview? +
Present the three gates made specific to their business from your research, and close with the sentence that separates serious candidates from template users: this is built from the outside, and I would revise it in week two once I have seen real data. The specificity shows you did the work; the caveat shows you will not arrive with answers before questions.
What are the most common first-quarter failures in Customer Success? +
Four patterns: rebuilding on arrival, announcing in week two that everything is wrong, which reads as arrogance even when correct; hiding in the learning, still ramping at day 75 with no win attempted; promising a metric you cannot move yet, usually NRR, whose renewal cycles land your influence too late; and treating the book as a queue, responding to noise while the quiet accounts that decide your year go unexamined.
How does the plan change for a first CS hire or a new leader? +
A first CS hire has no machine to learn, so month one maps the customer base and the founder's expectations instead, the day-60 win is making one lifecycle stage repeatable, usually onboarding, and the day-90 system is the first playbook, because you are building the function rather than the workload. A leader adds the team to phase one, wins trust by removing one thing the whole team hates, builds the operating cadence as the system, and reorganizes nothing for the full 90 days.