How Technical is MF PE?
Like the title says - wondering how technical / complex the work is for MF PE associates in particular.
For context, currently a first year analyst in M&A at an EB and signed an MF PE offer on-cycle. Currently not getting a ton of technical work and honestly hours have not been bad (50-70 most weeks). Don't think it is a reflection of my performance, my reviews have all been very positive, just a slow time for my group. That said, I'm worried that I'm not getting the reps / experience necessary to be prepared for or succeed in my next role.
If anyone could speak to how technical their experience in MF PE was (are you frequently building models from scratch or running complex analysis without much guidance or support?) as well as how they prepared, that would be much appreciated.
Comments ( 52 )
Following
Almost all models are made from scratch and it gets fairly detailed. You're expected to know how to model with limited guidance, but I don't think it's anything earth shattering. It's difficult work, but as long as you know how to build models, you can figure it out. It takes longer for some vs. others, but that's how you realize who is more cut out for this.
What would you say are the types of models you're building from scratch most frequently? Operating models? LBOs? Something else?
How technical was your group in banking? Anything you did in particular to prepare?
From what I've heard, most large institutional MF type places have templates and precedents - are models built from scratch to meet these, or completely from scratch in the format and style the associate prefers?
This aligns with what I've heard too, especially for LBO models , which are usually from templates, but I guess operating models might have to be more bespoke (understanding they likely also leverage precedent models for similar companies)
-
-
Learn more
-
|
Suggested Resource
Learn More300+ video lessons across 6 modeling courses taught by elite practitioners at the top investment banks and private equity funds -- Excel Modeling -- Financial Statement Modeling -- M&A Modeling -- LBO Modeling -- DCF and Valuation Modeling -- ALL INCLUDED + 2 Huge Bonuses.
Learn moreI mean every place has precedents, but you almost always build from scratch and just reference other models. LBOs are easy to build, but it's the operating model that's nuanced for every deal.
I'm an AS2 at a MF and I don't understand why you wouldn't just make it from scratch every time. The "templates" no matter how well they are built will have issues or add unnecessary bloat that slows down the file. The more toggles and switches you have built in the more likely this is to be true. I usually copy the general IC templates (which are ~95% the same every time), the "LBO output" part of the model (which you just link to the operating model, and then piece-meal reference old models / deals for specific types of businesses .
Some of the comments here are insufferable about using a template for everything. Of the 2 platform deals I've been lucky to work on since I started, there's a ~0% chance the templates referenced in this thread have the functionality to build in the structures we put in place for those two situations. If a firm is relying on a template to do 90% of the model then that's telling me they only do the same type of deal over and over again and are not very creative/scrappy. I'm also curious which firms these could be as I have friends at most of the MFs and I've never heard of a model being as simple as "dropping in a few cash flow items from the QoE".
Getting to OP's question, the type of modeling you use in PE is not like in banking. Unless you are ripping MM sell-sides, most of the time in banking you focus on below the line numbers (EBITDA / Net Income , CF items etc.) and in my experience in PE most of the model work is refining everything above that. This is really not learned in banking except in rare instances. Best way to learn it is to learn is to take a random company and try to build an operating model for it. You won't have the data but that's not important for getting quick in excel / clean formatting. Just try to build the infrastructure and inputs in a clean way. The first two times I did this were in college and I'm glad I had to do it back then as I look back on how much better my modeling is now. Every time you make on focus on finding ways to improve things, better formulas, better format, better operating logic, etc.
Some businesses that could make sense to start with: TravelPort pre-LBO 10-K had very clean financials, pick a software business of your choosing, TransDigm / Mr Car Wash / POOL (services roll-ups) - for these try to create an M&A bolt-on template, within our firm we sometimes call them tuck-in profiles or new business line profiles if it's a business around new service launches.
Just a quick question - for your operating model are you factoring in any sort of transaction considerations? I'm assuming not, right? This is just preliminary to get a sense of what drives the business/ ebitda /margins, etc
I think it depends on how you define a template.
I make music on the side. I start from a "blank" workspace every time. However, I have templates for what I am trying to for tracks. For example, I may start from a blank template but once I determine I want to use a "lofi piano" in the song, I pull up a piano and the "lofi" template I already developed. I think there is utility in having templates for commonly used things. If you commonly use a "lofi piano" then you should probably save yourself time and have a template for that. However, that doesn't mean you should template everything. If everything you use is a "lofi piano" your music is going to get very boring very fast. If all you are doing is using a template for a deal then you are probably missing something.
To bring this back to a more concrete finance example, take the... "summary level" financial statements . The financial statements that would be presented "as the financial statements" that don't have the detailed schedules and the like that actually drive the model. You could probably have a template for this "summary level" financials. You could also probably have templates for specific schedules. For example, in the aerospace industry most suppliers work based off of the end plane. For example, Boeing may build 52 737s a month and the supply chain would be based off this 52 shipsets a month number. You could build a template for this in revenue. You can do an "ASP x QTY" schedule based around plane models if you are doing aerospace deals where you can pull up this specific schedule as needed. For example, if the supplier only works on Boeing commercial jets then you'd pull this template for the 737/787/777/etc and just stop there. Or you may work on all Boeing and Airbus commercial jets you just add this... sub-template to account for the additional Airbus content. What I am basically trying to say is that you can have both high level templates (ie - just pull up this entire model and just change the numbers) and lower level templates (ie - having a template for a specific schedule but not the entire model).
But yeah. Broadly speaking, using one template for something like 90% of the model is generally indicative of either the firm doing virtually identical deals or they are missing something important.
Facts what I have first hand experience from 2 MFs and 3 UMMs (reviewed models, co invested, worked at, roommate worked at) and I've never heard of this. We have best practices templates but we substantially change them.
Anyway the technical nature of the job isn't from building some random model is from deeply understand how cash would flow in a particular scenario to the various investors, what that scenario means from a business and exit standpoint and how it would interact with your security and every implication thereof (incentive alignment, downside protection, upside capture etc) and then be able to negotiate all that on the fly.
not sure if that makes sense to folks but that's where technical expertise comes into play - visualizing the distribution of outcomes and structuring an approach that captures the best part of those outcomes while still incenting management and making the seller happy. Of course if you're just doing vanilla buyout maybe all less relevant
Based on the past two years when we run hard at a deal it probably goes like:
65% Operating Model (from scratch)
20% LBO math
15% IC material outputs
Do you feel the operating model actually gives accurate forecasts?
Accurate in what sense?
In that it accurately predicts business performance during the hold period? Over the last 3 years? Hell no, not accurate.
In that it helps you think clearly about drivers of returns and what will make / break a successful deal? Yes, largely accurate.
This breakdown is really helpful, thanks
Currently in IB but I did an internship at an UMM PE firm before. We had an LBO template model that we always used and had an operational model that fed into the LBO model . How deep we went depended entirely on the situation. If we had just received an IM that was semi-interesting we might just build a very simple operational model that is based on headline growth, margins and % of sales for key items. Such a model might end up being 100 lines and even as an intern I could build those in 20-30 mins.
When we got more interested in a potential target we would build more detailed operating models where we modelled by country, division etc. Such a model might end up being 2-3,000 rows and take a few hours to complete. However, most often the models were not very complex, there were just a lot of rows.
When we pursued a target or modelled for portfolio companies, things could get really complex. For example, we had a software company where the operational model ended up being 20-30,000 rows. That model was also quite complex as we did a monthly bottom-up model with ARR builds, measured sales per FTE in sales with scale-ups etc. Two VPs took over a month to build that model.
It is also worth mentioning that the modelling is extremely dependent upon the type of information you have. When you are at the early stages of looking at a target, you will simply not have enough information to build a very detailed model. As such, you will most often encounter very detailed and complex models when working with portfolio companies or have access to a data room as that is when you can get enough information to model at a very detailed level.
Ridiculous when you could get 95% of the precision with an EV / EBITDA or EV /EBITDA-Maintenance Capex on one page
I agree. The operational models were typically extremely subjective at the first stage. Most often you would find some extreme numbers in the IM (let's say 25% growth p.a. for a company that has been pretty much flat historically, with margins going from 10% to 15% in the forecast period). Then my VP would ask me to start with some numbers pulled out of the air like 5% growth and 12% margin to see where we land.
It would either be:
- "The assumptions give us a 3.2x base case"
- "That's too high, this should be more of a 2.5x case. Let's keep profitability flat and see what happens"
or
- "The assumptions give us a 2.2x base case"
- "That's not gonna fly with the partner who is heading the deal. He really likes this case. We need to get at least 3.0x for a deal like this if it's gonna fly with IC down the road. Let's see what we can adjust"
Can anyone give an example of these detailed operating models and what goes into them that makes them so big/complex? Coming from a pretty chill product group to MF PE and really only see the higher-level 3-statement type of stuff so really curious to understand how an operating model differs from a 3-statement LBO .
Basically, you do very detailed builds for each line item. Let's take revenue for a project-based business as an example. The company is operative in 15 countries and has 6 divisions. The average revenue per project is very different between the 6 divisions as the complexity is different. You might also have different prices in different countries. You could model this as avg. revenue per project * number of projects for each division in each country. Your avg. revenue per project might be driven by inflation in each country and price increases that differ for each division and country.
You will then model costs. Obviously, your number of employees will increase with the number of projects. However, the cost for an additional FTE will differ between countries and divisions if the complexity of the work varies. As such, you will also need to model direct costs by division and country. These things aren't necessarily very complex to model but the number of line items will result in a lot of moving parts.
Then there are also things that are more complex. For example, we had a model on a SaaS company where we had a scale-up for new sales employees by month. We would model new sales by the number of FTEs in sales. However, we assumed that a new FTE wouldn't be very efficient for the first few months. So the number of sales per FTE increased with the tenure of the FTE. As the sales personnel were on commission, our costs were also tied to the number of sales made by each FTE. We then had to calculate the effect of new sales on ARR. We also had to model upsell and churn to figure out ARR.
This is just what I came across during my time as an intern. I'm sure there are better examples of models that are far more complex than the ones I came across.
Very helpful -- thank you. So yes it sounds like they aren't that "hard" but more like detailed intuitive builds that take a lot of time/moving parts. Interesting.
Just to chime in, in MM London and I built one which looked at essentially the value of healthcare cases, the length they'd be open, depending on the complexity of case etc.and then any inflation / market growth.
Then had the other divisions / segments which were a bit simpler.
Didn't have different countries though.
Do most MFs use FactSet or CapIQ ? Both?
Unlikely to be both, probably one of the two
This has been discussed many times before, but from my perspective, the models end up being essentially useless the moment the deal closes. If every deal went according to the model, every deal would pretty much be a 3.0x return. The models are incredibly detailed and well-built, but GIGO applies here - Garbage In, Garbage Out. No one is able to accurately predict how many customers will be won/less, the expenses incurred on projects … synergy numbers are completely estimated and are usually way off … it is all just one big game to give you false confidence in your investment. Afterall, people just make the model say exactly what they want it to say - if the Partner likes the deal, expect to continue to adjust the assumptions until the returns meet your hurdle. Almost never in my career have I heard someone say: "Wow, this seems like a really great company and I really want to invest, but the numbers just don't work … darn … let's move on to the next one."
CompBanker’s Career Guidance Services: https://www.rossettiadvisors.com/
so wtf is the point of using models then
They serve all sorts of purposes:
1 - If you want a bank to give you debt to finance the acquisition, you need to demonstrate to them that the company can meet its obligations.
2 - If something goes wrong, no one can accuse you of 'winging it.' Models and comprehensive due diligence is a great way to "cover your a.." Limited Partners are expecting you to do the modeling work.
3 - The process of modeling is actually pretty educational in terms of learning about the business. It drills into your head which divisions / products / etc. are the most profitable, where management thinks the growth is coming from, etc.
… tons of other similar reasons.
CompBanker’s Career Guidance Services: https://www.rossettiadvisors.com/
The technical element is really more organizational than financial or even Excel-based. We're not out here re-configuring CAPM or using VBA .
But the ask when it comes to modeling is:
A. When you get asked to replace a blue cell assumption with a gigantic data dump in diligence, can you replace the assumption with something dynamic that's 1. accurate, 2. readable / reviewable, and 3. easily editable in case underlying mini-assumptions change? And can you cut that data quickly?
B. Can you make the transition from the "plugging in management / bear / bull cases to get the first-bid return outputs out" stage to the "if Brazilian productivity is growing in 2024 at a slightly different level because we changed our assumptions for one account" while making sure that every related output both links correctly and is updated by the time it hits the new PPT deck?
Those aren't unrelated, and may sound simple, but again people just want to know you can handle mechanics correctly even with organizational complexity so they (and you!) can move to understanding drivers / risks correctly at a high level.
I'm moving from a less technical coverage group (at a bank with a separate M&A team) to a large PE fund next year and want to do some prep on my own time. What prep would you recommend I do before I join so I can get up to speed technically most efficiently?
You passed the model test? You're good. Seriously there's no need to put yourself through boot camp if you know how LBOs work
Adipisci ut est numquam qui doloribus. Repudiandae voluptatem impedit dolorem qui in esse qui aut. Vel dolorum et voluptatem qui veniam qui.
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...
Quisquam a perspiciatis natus eligendi. Voluptatibus porro esse assumenda tempore nulla perferendis illum. Architecto dolorem voluptatem qui sunt. At eveniet et blanditiis at voluptatum.
Consequatur iure consequatur soluta. Nemo cupiditate cumque inventore minus.
Reiciendis officiis doloribus suscipit et ut. Et id ex et eos sint. Minima ad ipsa enim non iusto. Repellendus harum fugiat voluptate sequi ut qui illo doloremque. Et expedita aut blanditiis.