Yesterday morning was PowerDay, 9:00am-12:30pm Eastern Time. I'm in Vancouver, so I dragged myself out of bed at 5:20am for the interview. Luckily it went well — less than 50 minutes after the interview ended I got an email from the recruiter with an offer writeup, and now the next step is team matching. I've been sending out résumés for a year and a half and honestly it made me want to die, constantly falling apart, occasionally feeling stable. I finally don't have to send out résumés and rejection letters anymore. Along the way I got a lot of help and support from people who'd gone through it before me, so I'm writing this experience report hoping it helps more people, and wishing everyone an offer soon.
Timeline:
- 11/4 referral submitted
- 11/6 received OA invite, required to submit within 14 days
- 11/13 OA passed
- 11/15 recruiter phone call
- 11/17 received the take-home data challenge (DC)
- 11/25 submitted the DC
- 12/7 notified the DC passed, scheduled the earliest PowerDay slot; also had a call on 1/8 where the recruiter introduced and helped prep for PowerDay and let me ask interview-related questions
- 1/17 PowerDay
OA: Same as other write-ups I'd read — camera on, I think it was 15 questions, one SQL and the rest multiple choice, analyzing datasets, all doable in Excel. Time was very tight and I didn't finish at all. I only fully thought through and finished the first 7 or 8 questions; the last 7 or so needed four datasets, if I remember right, and I didn't even get to look at the last few — pure guessing :) But the bar probably isn't that high, getting more than half right should be enough.
RC phone call: Simple questions about experience and skills, asked how proficient I was with Tableau, told me about the possible office location and salary expectations, then introduced the DC and asked when I wanted to receive it. No technical questions.
DC: Airline case. The prompt said the average time was 8-10 hours, but I spent way more than that. I spent the whole Thanksgiving break hunkered down at home scratching my head over it — honestly it was pretty complex, and at first I had no idea where to start. When I submitted, besides the Jupyter notebook I also turned in an Excel metadata file for the new fields I'd created, plus a PPT. I ended up using that same PPT for the PowerDay presentation too, with only small edits from the submitted version. My suggestion is to treat it like a real business project — do the data cleaning and processing carefully (I don't know if it's because I was thorough that my explanations were also thorough, but during the interview they only challenged one small point, and I recovered quickly). Make sure your assumptions are solid, and every step should make business sense. I also built a rough storyline that tied the different questions together. Read the prompt carefully, think through the logic and method for each part clearly — it doesn't need to be complicated, but it needs to be well-reasoned, and you should know why you chose each method so you can explain it if challenged during the presentation. When I presented, I went through the PPT first, and partway through he asked me to go deeper on a metric I'd defined myself for judging route profitability. After the PPT, the interviewer said he wanted to look at the code again, so we went through it, and he'd interrupt with questions wherever something came up.
PowerDay
Order: DC - Case - EQ - Case
Overall experience: very good. Every interviewer except the last case interviewer was smiling, warm, and gentle. The last one wasn't hostile or difficult either, just a bit cold and expressionless, like a robot — not much small talk, straight to business, no filler. What's interesting is that in write-ups I'd read, the BQ section and the DC both seemed to end early a lot, but for me all four rounds ran almost the full 45 minutes, and a couple even ran a minute or two over.
Case advice: The case question bank I'd studied from was more than enough to prep from — focus especially on the high-frequency SDA case questions. Personally I feel the SDA case pool is basically fixed, and it's a subset of the SBA case pool — SBA interviews can pull from the SDA pool, but SDA doesn't necessarily pull from the SBA pool. Whether or not you've already prepped a given case, treat it as new and re-derive the numbers, in case the interviewer uses a different context or different data.
Case 1: A common case, the "coder" case. Maybe because this case is short, this interviewer spent a good chunk of time chit-chatting with me first — about the weather, snow, his two daughters, how he'd just moved to Texas, and so on — I was a little confused, wondering if I was actually here to do a case. The question was exactly the same, and the numbers seemed to be the same too. My interviewer actually simplified it further — we never even got to the "hire a contractor" option; he only had me calculate the scenario where existing staff work overtime, and asked for my recommendation. If you've prepared and perform normally you shouldn't run into trouble. There was still time left at the end, he complimented my work, and we chatted with some reverse questions.
Case 2: Amusement Park. Also a case that shows up a lot in write-ups. This question seems to have a few variants, and the version I got happened to be one I hadn't prepared for — but if you get a case you haven't prepped, don't panic. Calm down and think it through, treat it like an elementary or middle school word problem, say out loud every method or possibility you can think of, and keep actively talking with the interviewer. Usually they'll give you hints. My interviewer didn't give many hints, but he did answer whenever I asked something — anyway, the more you talk, the better; talking is basically winning.
This question had a lot of numbers. Partway through he shared a spreadsheet with an income-statement-like table, and I needed to calculate revenue, cost, profit, and so on.
Roughly, the questions were:
Q1: What's the amusement park's business model? Basically, what are the revenue and cost sources?
Q2: For the first question he gave roughly this data: an amusement park, 2,000 acres, 1 million annual visits/entries, 3 ticket types — single-day $80, 5-day $300, annual pass $1,000. Asked for revenue. The interviewer didn't tell me how many of each ticket type were sold, so I had to ask, and he told me: single-day 250k, 5-day 100k, annual pass 10k.
Q3: Asked, for people who bought the annual pass, what's the average number of annual visits. I got completely stuck on this one. I tried not to panic and spun through options in my head, and after going back and forth with the interviewer, he told me the current information was enough to estimate it. I eventually came up with a method — made a small mistake along the way too — and finally got 25 days/year. He asked whether that number made sense, and I said it was reasonable (honestly not sure if it was right, I just forced out a few justifications).
Q4: Then he shared a screen with a spreadsheet showing an income-statement-like table, with various revenue and cost sources, and gave me time to copy it all down, and I needed to calculate revenue, cost, profit, and so on. Worked out a profit of 15 million.
Q5: New condition — the neighboring landowner offers an olive branch: there's a 1,000-acre plot available, on condition that the land owner fee rate goes up from 5% to 10%. Some revenue and cost variables scale proportionally with land size, while others stay fixed. Asked whether to take the offer and recalculate profit. Found the new profit was higher than before — my answer was 25 million — and my recommendation was to sign.
Q6: New option: you can enter a bid. If the bid succeeds (80% probability), the land owner fee rate stays at the original 5%; if it fails, you lose the land entirely. Recalculated the profit if the bid succeeds, then calculated the expected profit of taking the bid — success profit × 80% plus loss-case profit × 20% — which came out higher than both previous options, so my recommendation was to take the bid.
Q7: Do you see any gaps or underlying risk in your calculations and analysis?
BQ/EQ: The usual three — failure/mistake, helping others, accomplishment. The interviewer explicitly said they were looking for situation, action, and result. I'd prepared my stories in detail using the STAR method, but they still dug deeper — for example, when I talked about handling an urgent last-minute report request from a client, the interviewer asked roughly how long it took, what role my manager played, whether I got support, and what kind of support. Luckily all my stories were based on real experiences, so they held up well under scrutiny and I could answer follow-ups fairly easily. My suggestion to everyone is also to write your own stories based on real experience, and practice with several people — see if they can follow what you're saying, and whether you can quickly respond and expand on details.
That's about it!
Discussion
Loading comments…