Arca
Explore the Finance interviews hub and the finance careers guide.
A 30-60-90 day plan is your answer to the question that shows up near the end of every final round: what you'd actually do once you're in the seat.
For a finance seat, the honest answer has a fixed order, analyst or controller. Learn how the numbers are built and who uses them. Get one number to tie. Then improve one thing and show the before and after.
It's one of our finance interview guides, with a template you can copy.
What is a 30-60-90 day plan?
A 30-60-90 day plan is a one-page plan for your first 3 months in a role, split into three phases with a different goal for each.
Most templates call the phases learn, contribute and lead, which fits any job anywhere. A finance plan should read like someone who has lived through a month-end close wrote it.
The 90-day window isn't arbitrary. Michael Watkins wrote a whole book on it, The First 90 Days.
I never had a 90-day plan going into anything. I developed one as I grew up in my career, and it's the same in any finance seat. Learn everything. Make sure it's all stabilized and you can do every part of it. Then start to improve. A 30-60-90 plan is just those three moves with dates on them.
What should a 30-60-90 day plan include?
A 30-60-90 day plan should include six things: a goal for each phase, who to meet and why, what you'll learn first, one early win, a measure for each phase, and the questions you still need the hiring manager to answer.
Most plans skip the last one.
Write down the 2 or 3 things you can't know from the outside, like who signs off on the forecast, or whether the close lands on the day it's supposed to. Then ask them in the interview. It gives the hiring manager something to answer instead of something to grade.
Why do interviewers ask for a 30-60-90 day plan?
They ask because it's the cheapest way to watch you do the job before they hire you.
It shows three things.
- Whether you understand what the seat is for. A controller plan protects the close and an FP&A plan protects the forecast.
- Whether you'd rush. Someone who promises a new forecast model in week 2 hasn't seen the data yet.
- Whether you can prioritize. Ten actions in each phase tells them you can't choose, and choosing is most of the job.
Gallup found only 12% of U.S. employees say their company does a good job of onboarding, and HBR adds that logistics isn't fitting in. A candidate who arrives with a plan isn't counting on the company to have one.
When I'm the one across the table, I'm listening for whether the candidate plans to look before they touch anything.
What should you do in the first 30 days?
Spend the first 30 days learning how the numbers are built, who uses them, and what they do with them.
Listening doesn't look impressive on a slide, and it's where the expensive mistakes get made.
I learned that on a five-year planning process that took inputs from around 60 people across finance. I redesigned the intake to make it easier, and I'd gone overboard and simplified it a little bit too much. The cuts the VP needed downstream, like compensation, IT and depreciation, weren't there anymore.
The fix was a meeting I should have had first. "Let's have a meeting just so I can understand how you use it, the lens you look at it from." Then I simplified it again without breaking the output.
Listening, then building. Now I run a small loop. Map who uses the output, get a mini buy-in, put in the change, see the effect, and follow up with the same people again.
The lay of the land means four things.
- Who reads each report, and what decision they make with it.
- Where the data comes from, and where it gets typed in by hand. The hand-typed cells are where the errors live.
- The close and forecast calendar, and which dates actually move. Every calendar has a day 3 that's really a day 5.
- Which number people argue about in meetings.
The first one matters most, because the ask you get is rarely the whole question. When a new boss asks me for something, my first question back is the background, the nuance and the timeline they need it in. I was once asked for sales against target by rep, maybe ten minutes of work. It was actually feeding a headcount decision, and the way the ask was framed pointed at the wrong person. I showed both views, and the decision changed.
All of that assumes there's someone to ask.
When I started a senior manager seat, my plan for the first 30 days was to learn the ropes. But the person before me had left, and my boss was new too. So I learned from old files, old models and questions, which put me inside somebody else's system. It was foreign and inefficient. I was stumbling through the role, and it made me uncomfortable.
The goal got smaller. Recreate what they did in their old system, then improve it.
If that's your seat, rebuild one month the last person's way before you change a cell. That's how you find out which of their odd steps had a reason.
Either way, sit through one close and change nothing.
A good day-30 goal: "I know who uses what, I've seen one close end to end, and I've fixed one small thing that was costing someone time."
What should you do in days 31 to 60?
Days 31 to 60 are for making the baseline dependable, which means reconciling the number people argue about and owning one process.
If the numbers don't tie, a faster report gets you the wrong answer sooner. That's modelling 101. Every number ties back to the source. It's our own version of the scientific method: if somebody rebuilt your model with the same assumptions, they should get to the same number, and you can only do that with a data source that's fixed, as opposed to manual typing here and there.
Sometimes the baseline is more basic than a number.
The most extreme version I've lived through was bringing a small acquired business into a larger finance team. It came with paper cheques, an assistant doing the books in QuickBooks, and one person close to retirement who held most of the knowledge. The first thing we had to do before we could even get up and running was figure out the banking. Nothing on the improvement list mattered until cash came in the same way every month. About 80 percent of customers moved to automatic payments within 6 months, and the time to get paid dropped from 40 to 50 days to under 30.
If you spent month one in somebody else's system, month two is when the improving starts. Begin with the number people argue about, and trace it back to where it's first recorded. Somewhere along that path two teams are calculating it differently. Get them in a room, agree one definition, write it down, and name an owner.
Standardize before you automate. When 7 teams were each building the same scorecard their own way, fixing the inputs first is what took errors from about 10 percent to under 1, and the full story is in FP&A interview questions.
I believed the opposite once, about my own AI tools. I thought a rough process would sort itself out once AI was on it, and what came back was work I had to keep checking. Now my line is that "AI isn't going to fix a broken workflow." That story is in the controller guide's automation answer.
So month two writes the process down, and month three has something worth automating.
A good day-60 goal: "The number ties, everyone uses the same definition, and I run one process without anyone checking behind me."
What should you do in days 61 to 90?
Days 61 to 90 are for one measured improvement you can show, and a ranked list of what comes next.
By now you know which report nobody reads and which process eats a week of someone's month, so fix one, with a before and after.
One I did was a monthly pack nobody read, moved into one live view, and the FP&A interview questions guide tells it in full.
Make sure it's a fix and not a patch. When something goes wrong, I ask three questions. What went wrong? What can we fix? Are we going to run into this again? The last one decides it. A patch you redo next month is just next month's manual step.
End with a ranked list of the next 3 improvements and a rough effort for each. That turns a 90-day plan into a reason to promote you later.
A 30-60-90 day plan template for finance roles
The plan fits on one page, so copy the structure and swap in the posting's specifics.
A finance 30-60-90 day plan
- Days 1 to 30Learn. Meet the people who use the numbers, map where the data comes from, sit through one close, and deliver one quick win.
- Days 31 to 60Make the baseline dependable. One reconciled source of truth, agreed metric definitions and owners, and one recurring process you now run.
- Days 61 to 90Improve and show it. One measured improvement with a before and after, plus a ranked list of what's next.
For each phase, write three lines:
What each phase needs
| Line | What to write |
|---|---|
| Goal | One sentence. What's true at the end of this phase that isn't true now? |
| Actions | 3 or 4, each tied to something in the posting. |
| Measure | One thing you could count or show, like a cycle time, a pack date or a reconciled number. |
The last line is your finish line. Leadership has one trusted number, and the pack ships faster than it did in month one.
How does the plan change by level?
The phases stay the same, but the scope grows with the seat.
Same three phases, different scope
| Seat | What the plan should emphasize |
|---|---|
| Financial analyst | Learning the data and the recurring reports, owning one report end to end, catching errors before they go upstairs. |
| Senior analyst or FP&A | Owning a forecast cycle, partnering with one business leader, improving one model. |
| Finance manager | The close and forecast calendar, a team's workload, one reporting cycle rebuilt, and who on the team you'd develop. |
| Controller | Close accuracy and timing, controls, the audit calendar, and agreed hand-offs with FP&A. |
For a controller seat, the close is the plan, and the questions before you get there are in controller interview questions.
The analyst seat's questions are in financial analyst interview questions, and the FP&A seat's in FP&A interview questions and what they're really checking.
How do you present a 30-60-90 day plan in an interview?
Present it as a hypothesis, and say what you'd do, why in that order, and what would make you change it.
A 60-second version
I'd spend the first month learning who uses which numbers and for what, I'd sit through one close, and I'd look for one quick win.
Then I'd make the baseline dependable. One reconciled number, clear definitions, and one process I own.
By month 3 I'd bring one measured improvement and a ranked list of what's next, based on what I learned, not what I guessed today.
Then stop and ask a question back. In my last final round, mine was close to this. "What would you want the new hire to own right away in the first 30, 60, 90 days? What would you expect to be taken off your plate in the first three months?" That turns a recital into a conversation.
They'll push, so have the pushback answered before you walk in.
When they push on the order
Don't defend the plan. Show that it bends.
If it's a take-home, keep it to 3 slides, one per phase, plus a title slide with the finish line.
I've built one of these for real: a 90-day plan pitched to a CFO, as a deck plus a recorded walkthrough. They hadn't asked for one. The role was exactly what I wanted to do, building out a finance function, and it was such a strong fit that I had to get my pitch in. I didn't hear back. That pitch is where this framework came from. I work from a script when I record ("I work better that way"), and the brief for this one was four beats: "I understand your problem." Then: "This is how I would approach it over the next 90 days, and this is why you should hire me, and this is my ask." There's "something that they'll actually read," and there's "what I'm going to show on screen." The on-screen one gets fewer words. You point at it and talk.
What are the most common 30-60-90 plan mistakes?
The most common mistake is promising results before you've seen the data.
Same plan, two versions
Rebuild the forecast model and cut the close by 3 days in the first month.
Learn how the forecast is built in month one, own one cycle in month two, and bring a shorter version with a before and after in month three.
Four others show up often:
- Too long. A 5-page plan tells them you wrote it before you understood the seat.
- No measures. "Improve reporting" can't be checked. "The pack ships 2 days earlier" can.
- Skipping the close. If the plan never mentions month-end, the interviewer assumes you've never sat through one.
- Criticizing what's there. There's usually a reason it looks the way it does, and you find it by running it their way once.
One more shows up once you're in the seat. Don't schedule your asks on close days. When a team pushed back on me during that acquisition, the first thing I checked was whether it was a genuine capacity constraint or they didn't see their own dependency. There were month-end and quarter-end deliverables in the way, so I timed mine around those.
How to build yours in one evening
You can write a solid plan in about an hour, posting open beside it.
Your 30-60-90 in an hour
- Highlight the 3 priorities in the posting, in its own words.
- Write one goal sentence for each phase.
- Add 3 or 4 actions per phase, each tied to a highlighted priority.
- Add one measure per phase you could count or show.
- Write the finish line for day 90 in one sentence.
- Say the 60-second version out loud twice.
If the new seat owns a planning cycle or a team, show that on your resume too, and our finance resume guides cover how.
The first time you say it out loud will be rough. I'm weak at the start, to the point where I don't even want to begin. Once I do a few reps and get a little success, I'm all in. That's why it's twice. Don't time the first pass. It's for understanding. Time the second one.
Write the finish line first.
The rest gets easier.
