Hedge Fund Modeling Detail?
I'm curious how HFs (talking fundamental L/S HFs) do models and projections. Coming from banking, unless we have the management model, we're almost always just taking consensus and extrapolating. It's never that useful other than to put numbers around "value creation" or expected valuation in a deal.
Curious how you traders actually go about modeling. How do you efficiently construct projections for a public company when you are completely outside-in? I tried once with a CDN company (e.g. Fastly, Limelight) and got completely lost in the sauce with the revenue projections.
Do you do quarterly models? How do you project revenues? Do you do full 3-statements or is that useless?
I get that a lot of it is dependent on your investment thesis, so I guess as a corollary, what piques your interest about a specific company that leads you to the actual process of modeling the company?
Thanks for any insights!!
Bump
Great question. It entirely depends on the strategy, what company you're valuing and the current industry, but in general terms for our L/S HF our models follow this format:
- Quarterly 3 Statement Model;
- DCF
- Comparable Universe (For multiples)
- TAM
- Unit Economics (Most important tab for us)
- Consensus (sanity test)
- Stress Test (I have a Big4 M&A background where ALL assumptions were stress tested -> I guess it became a habit).
The contents within these tabs vary for each company, therefore it doesn't make much sense going into it in-depth. However, typically I don't touch companies if I can't 'draw' the business model on a whiteboard or a piece of paper. Rule of thumb -> if you're reading about the company for 10 mins and it still interests you...then read for 30 mins. If the company is still interesting to you after 30 mins, read for 2 hrs. If the company has your interest after 2 hrs -> build the model and initiate your due diligence.
This is great advice ^ if not of interest. And if cannot model / take a view on fundamentals gtfo
Can you give an example of what type of analysis you would include in your unit economics sheet? Is this actually being used as a driver for the model or is it a side calc used to calibrate your modeling assumptions? Are you looking at both current unit economics as well as trying to forecast “steady-state” margins?
Good question. It varies for each company but for this example let's assume we're looking at a SaaS or E-commerce co.:
Here's what it looks like at a current state:
- ARPU
- GPPU
- Churn
- Fixed / Variable Costs Per User
- Contribution Margin Per User
- CAC
- CapEx Per User
- LTV Per User
- GMV per user
- etc.
- etc.
These are a few of the metrics we use. However, if the company is young (majority of the time), then we rely on Unit Economics as the driver for the valuation. If it is an older company that has ~8 years of operating financials (I.e Shopify or CRM), then we simply use it to calibrate our modeling assumptions.
For a younger company, we also forecast these economics. We ask questions like "Can this company benefit from economics of scale? If so, what would that do to margins at scale? Did their older / larger competitors benefit from this? If they did, what do their unit economics look like? Why / How will our target company benefit from this?"
We go a lot deeper into the financials than most funds, a lot of funds just extrapolate consensus and rely on quarterly earnings as their catalysts. We very rarely do that, unless it's an extremely mature company (i.e KO) where the future is far easier to predict.
If you look historically, consensus got quarterly earnings wrong nearly 80% of the time (to the upside) for younger tech companies. Take a look at SHOPs early public earnings, consensus almost always got quarterly's wrong for the first 3 years. This is commonly the case.
I hope this helps.
how much are you guys down this yr?
hey quick question. what does TAM stand for?
Total addressable market
True An@l Master
Thanks for the detail! A few follow ups if you don't mind:
1) I'm surprised you do a DCF. I've always felt that this was a pretty useless method for trading in general. How do you utilize this analysis exactly? Is it just another sanity check on your price expectations?
2) How do you build TAM forecasts? Does some of this data come from proprietary sources? Or are you building largely from public info?
3) Similarly for unit economics, is it mostly proprietary, or is it built from limited public info with a lot of assumptions?
Good questions. Sorry if my response is a bit lengthy, but you bring up some valid points.
1. People often take the output of DCFs too literal. They're simply a tool that should point you in the right direction. No DCF is 100% accurate, regardless of how much time you spent creating the model. When we use a DCF model, we often cross-check it with our multiples model to see if they're similar. If our multiples pricing model and DCF both indicate that the the stock has 100% upside over the next 12-24 mos, then this is pretty confirming evidence. It's even better if the unit economics produce a similar outcome. If you have 3 different models pointing to the same outcome...now comes the fun part....We try to determine how we could be possibly wrong. It's a very arduous process, but when you're managing other people's money, you want to be sure that you're making the right investment.
2. There are a variety of ways you can arrive at TAM. We have relationships with various marketing companies that will give us the TAM research that they've conducted for a company in the same sector, or in most cases, the target company in question has it published in their 10-k or S1 filing. It's not necessarily proprietary, but sometimes it can be hard to find. When marketing companies don't have the research, and the target doesn't have the research, we simply take the NAICS code of the business and compute it ourselves. When it comes to forecasting, it's very straightforward. If the industry maintained a CAGR of 15% over the last 20yrs, we simply forecast it at a CAGR of 15% for whatever our discount period is.
3. When it comes to unit economics: It's not proprietary. 99% of the time we can compute it using the financials that we find in the financial statements. It usually built with pure public info. There were a few times that we had to essentially 'pry' information out of management to compute unit economics.
hope this helps.
Cap structure + summary financials can get you 90% of the way there, why over complicate this
Try the WSP Premium package. It is not exactly like the real thing, but it helps you learn and understand the fundamentals. Then it becomes all about inputs, and that is something that is not entirely objective.
L/S SM analyst here. Almost never do forecasts … not worth the effort. Predicting the future is hard
What strategies do you use?
Id dolore ea qui. Atque nisi non iure in cum qui aut. Sunt ullam in quod vel eius. Occaecati culpa est maxime atque.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...