I am looking to hire a Financial Modeler for modeling projects. Other than the usual
What interview questions to ask when hiring a Financial Modeler?
Answers
The best way to be sure you are getting what you need is to request a modeling sample. Base your discussion on what you see. There is a danger in asking targeted questions as you may exclude someone based on a function they have used seldomly. I have modeled for about 15 years and have used V-look only a few times. There is never only way to get an answer. Asking for a sample allows you to understand the thought process.
1
Agree with Regis.
And, answering a technical question correctly doesn't mean they have the chops to actually perform the technical task.
I made the mistake of accepting answers provided by candidates as to their technical skills at face value during one of my last hires and paid the price.
The person is actually incompetent and not motivated to do any more than they absolutely have to.
Now, the CEO won't let me get rid of said person. And now I'm stuck with a body, filling a slot, sucking up a manager's salary and contributing very little to my department.
I either have to look the other way and do the work myself - which I have been doing for the last year - or expend extraordinary amounts of my own time and energy
Frankly, she's gotten away with murder and my hands are tied.
If I had used some specific, targeted testing for technical skills as I have at other employers, I would have "smoked" her out right away.
I agree with Regis, asking technical questions is hit or miss. Better to ask about problems the interviewees have solved with data modeling. Ask about the source of data, the volume of source data, what needed to be done to the source data to make it usable, and how was the source data or model results summarized. In this way, you'll get a feel for how technical the interviewee is as well as his/her ability to communicate results.
I'm known way too many people who over-model; spreadsheets that you need to F9 to recalculate because the complexity parks the machine, even for very simple problems.
I tend to shy away from the technical aspects (per Bob / Regis). Although I have always hated mgt consulting interviews, their bkms I do appreciate in this area. Ask the "how many dumptruck loads would it take to move Mt. Fuji" question. Get into how they think, and very critically, how they identify uncertainty. IMHO, in modelling the greatest danger is either thinking you are certain about a variable or missing a variable entirely.
Note (pet peeve); we all recall I'm sure the discussion of "significant digits" in science class in 8th grade. Ask a question that brings significant digits into play, or gets them to say "10.378 apples", and see if they say "about 10" or communicate the concept of std deviation, etc.
A few are here: note some seem totally irrelevant, but if they don't offer their thinking process, that *is* a red flag (part of giving an answer is communicating the constraints and inputs), so the penguin question is totally fair game.
http://www.geekwire.com/2013/cows-canada-24-wacky-interview-questions-amazon-google/
2
Keith - I could not answer most of these questions with a straight face.
My answer to the penguin question: He says, "Hello, my name is Juan, I am replacing the annoying Chihuahua for the Taco Bell commercials."
Thanks Regis for the suggestion on asking for a sample. It is indeed a very direct way to see how a candidate conceptualize a model.
Keith, I believe the F9 was meant to recalculate Data Tables which are notorious for grinding Excel to a halt. Despite this disadvantage, it is still a common way and perhaps best way(?) to perform scenario analysis.
Thank you for the list of market-sizing type of questions. Keith, you made a good point. I agree they are good for testing the thought process of a candidate. I will squeeze in one such question.
I tend to think of these market-sizing and behavioural questions to be similar to asking a pianist his thoughts on a piano piece. All good if he can discuss in a clear and clever manner about the piece. But in the end I am really interested if he can play and compose.
I do need a modeler who can play and compose and thus my interest on the technical questions to ask.
Further sharing of possible technical questions to ask a financial modeler is much appreciated.
Thank you.
I do modeling about 50% of my time. Ask the consultant for a sample or two of his work that he has sanitized for name and industry references. I suggest asking for a challenging and a simple solution. Ask him what his charge was from the client and where he ran into pitfalls and most importantly, how he solved them. I often add a new function based upon a specific client need.
I would not focus much on the excel functions. A question or two on pivot tables and the statistics and random number functions.
Modeling and forecasting is about uncertainty and delta's. I would ask about methodologies for example "What are the differences between Monte-Carlo, Markov Chain and Dynamical System modeling?" or "What is more important in forecasting, precision or accuracy?" and he importance of the intercept and the slope of a linear regression. . I would also be sure to ask them about their biggest frustration with Excel.
1
[on the lighter side]
My biggest frustration with Excel (more specifically with the Intel Chip) is the rounding error that is out to the 17th place (if memory serves me well) right of the decimal point.
Can't tell you how many times spreadsheets don't foot because of it!
As mentioned by several above, it's better to look for proven APTITUDE in modeling than quizzing on specific techniques. Look also for an innate CURIOSITY for finding better techniques. The combination of aptitude and curiosity ensures they can quickly fill any current technical gaps as well as adapt to future needs.
If you really want to swing for the fence, look also for:
- Business acumen. Can they understand how your business works and what is needed to be successful? If they can, they will deliver not just information, but also INSIGHT.
- Relationship skills. Can they build genuine trust and RAPPORT across all departments? This is critical to get the information they need as well as to deliver messages folks don't want to hear.
- Communication skills. Can they SIMPLIFY complex ideas for non-financial folks to quickly absorb? Even the best Excel guru is impotent if they can't convey their work's implications.
Having spent many years doing such work, both internally and on a project basis, I wholeheartedly agree with David W's comments above. Knowing excel formulas and tricks is of course nice, but this should really be a given if they have any volume of experience in this area and are putting themselves out there for such work...
I think more important though is getting a sense of how a person interacts with others and that they can quickly develop a rapport with you and your colleagues. There needs to be a "fit" to foster a good working relationship, to ease sharing of information and developing the key drivers. They should know what questions to ask and know about your particular business/industry, which leads to a stronger modeling engagement.
You don't want strictly an excel robot, but instead a trusted business partner. If it were me looking to hire a consultant for such work, I would weight higher someone that has both public accounting experience (auditor type) and a number of years doing controller /
I suggest you also ask about the candidate's general understanding of and approach to managing spreadsheet
As JP Morgan recently learned with the "London Whale" fiasco, failing to manage spreadsheet risk can be very costly.
http://en.wikipedia.org/wiki/2012_JPMorgan_Chase_trading_loss
Be sure that your questions are consistent with what is needed from the position.
* Some financial modeling projects involve client-facing work, such as interviewing the client for data gathering as well as facilitating pitches, and as such necessitate a level of panache and interpersonal skills.
* Some require prowess with communicating complex data in an at-a-glance manner and therefore require visual design skills and the judgment to select the most powerful information to summarize and present.
* Some emphasize the results and only need the answers to the questions for which the model is needed.
Varying degrees of flexibility should be taken into account as well. Are the models one-time projects or can assumptions change in the middle of an investor meeting?
In a vacuum, we can't advise you to ask the VLOOKUP/INDEX-MATCH question or the Monte Carlo question. I invite you to share more about how the models will be used and then I will offer more specific suggestions.
In short, you'll need to be crystal clear about what you need from the modeler in order to know what questions to ask.
And don't overlook a demo, no matter what.
While agree that asking technical questions can be hit or miss, especially if you are not in a position to evaluate the answers very well, here are a few questions I would ask if I were looking for a someone to do financial modeling.
How about:
1. What are the design principles of a good financial model?
2. What is the difference between NPV and XNPV? when would you use either? What are the limitations of both? When would you run a NPV without either?
3. What is an array function and why would you use it?
4. Please walk me through a model you have built.
5. How would you summarize a data set of 50,000 rows? 5 million rows?
6. Which is more useful LOOKUP or VLOOKUP? When should each be used? And what are the pitfalls of each?
7. How would you use a macro? How many have you written? What do they do? How would you characterize the different kinds of things you can do with a macro? When would you not use them?
8. How would you dedupe a range of values?
9. Financial modeling is not my forte. What mistakes am I likely making? (In your case, if you're using Index/Match, there's probably a simpler way to do things).
10. Walk me through your process for running sensitivities?
11. How do you make sure your numbers are right? What kinds of checks do you build in to the model?
1
I like number 11. Can't tell you the number of spreadsheets I've seen (or made) that didn't foot or proved to be the right answer to the correct question...
To answer your Excel question, first let me start with that vlookup is a bit simpler and quicker to put together and has its time and place.
Sometimes though, we need to lookup a value that might be to the left of the search column (vlookup only searches from left to right).
Rumour has it that Index Match also works faster in big spreadsheets but I don’t notice much of a difference.
Here’s the structure of the Index Match formula:
=index(WhichColumnToReturnValueFrom , (match(SearchForValue, SearchColumn, 0)
=INDEX ( Column I want a return value from , ( MATCH ( My Lookup Value , Column I want to Lookup against , ZeroForExactMatch)
Filed Under:
Human Capital