Citadel (Surveyor) Case Study/Modeling Test

I’ve done a few of these with different teams there and they were all quite different. You get a big binder with Ks, Qs, transcripts decks, and maybe peers. I ran out of time on every one, but passed about half and got an offer for a couple. Not having the PDFs to copy the historical is a really a time killer. I did much better, always passed, when I did NOT finish the 3 statement model but rather had a full income statement with thoughtful revenue and builds and cost builds, P/E multiple. You don’t have time to build a beautiful 3 statement model and be able to discuss the model. They want to know if you can model technically and if you can understand a business and quickly identify key issues, risks, upside/downside drivers. My thinking was I could only prove one or the other given the time limit - if you build a good earnings model they’ll be probably willing to assume you could build a 3 statement with more time. If you can build robust model but are at a total loss when they ask you to discuss - I’m sure anyone who gets there can do that and you’ll spend the whole discussion saying “I didn’t read that part”. That was my thinking but depending on what the team is looking for it could certainly be dead wrong.

Other thoughts. Don’t format. Read the transcript questions. Model each segment. Try to have a unit/price build and an incremental edit margin at least or better, cost lines tied to units/employees/store count/etc

Do not build a PPE schedule etc. balance sheet working capital items - just forecast as same % of sales as last year. Add some seasonality and comps fluctuations in your quarterlies. Just use adjusted numbers, don’t waste time typing in the whole reconciliation. If FX matters, include a dummy line where you would put BBG FX if you had it. I’ve had an offshore drilling company (modeled monthly activity), retail store (beginning/opened/closed/ending store same store sales/opening and closing costs/per store cost items) and , media conglomerate (dunno, depends), diversified industrial company (backlog build, incremental margins, FX, SotP), pharma (mkt size, penetration, probabilities, patent expirys)

There will probably be some big “thing” they want you to pickup on (big risk or big opp) - I don’t think they usually pick a name they’re lukewarm/neutral on.

Fastest 4 hours of your life, stick to a schedule. 30 mins read 1qs/investor day deck, identify segments, drivers, issues. 1hr input historicals, 1 hr full income statement and multiple build and rev drivers with dummy assumptions, 15 minutes change drivers to guidance and make sure it’s not fucked up. 30 minutes tweaking drivers and making a bullet list of why you picked assumptions and other things you noticed and addl research or data you would do. If you feel comfortable after that build a quick balance sheet (cash,ar,inv... then type in total current assets and back into the the plug for everything else. Repeat for lt assets with PPE and other; repeat for cl, lt liabilities. Shareholder equity as one line. Cf includes builds for capex, buybacks, total debt pay down/raise, abc, dividend. Run everything off the is/bs and enter total cfo/cfinv/cfin for the hostricals, plug for “other”. Obviously the particular company has important issues that require more detail. Don’t build a DCF. Go in knowing a reasonable multiple range for each subsector, hopefully you’ll get a comp sheet.

For you discussion copy all tabs and so you can tweak assumptions real time as you go through it with them but you’ll still have your original assumptions so it’s not like you’re covering up errors or anything.

This was my approach, I don’t doubt people will disagree with all or parts, but hopefully it’s helpful. Good luck. It is not fun.