delete delete delete delete

it's the illusion of work so that they can charge what they charge. they're rather pointless otherwise

They are a complete joke and I agree with the sentiments of everyone else, but I want to provide a bit of a counterpoint. As a buyer, I obviously don't trust or care about the assumptions/outcome, but it helps me understand the business. I can look at the model to get a sense of how the finances work: what drives revenue, what are the big cost items, are they variable or fixed, what are cash flow needs, etc. I know they're put together by analysts that don't know that much, but it's still helpful to level set how I should think about the business.

Also, I’m going to have to make one, and it’s much easier for me if I can rip a lot of it from the bank model.

Also find it funny how nobody ever talks about this

This is a pretty common criticism from juniors in finance actually...

Is it that the concept of modelling in general is useless, or is it just sell-side models that are useless? People seem to suggest that it's important to get modelling experience for the buyside, so does that mean the models are more actionable/insightful in PE?

It's also p interesting to see how investors like Buffett hardly use modelling in the traditional sense (DCFs, Comps, LBOs), and they rely more heavily on qualitative aspects of a company's business model

In my views, Models are especially helpful when looking at companies with contract backlog or to understand the areas of growth of a business (e.g. opening of a warehouse in X or Y country)

If you're using a model purely to generate a final number of how much a company should be valued, then you're using it wrong. The value of the model isn't in predicting the future - everyone knows the assumptions are basically worthless and nobody, least of all a fresh university graduate has any idea how a company might grow with any amount of accuracy.

What it's actually useful for is being able to adjust certain drivers and inputs and seeing how it might affect the valuation of the company. This way, you can easily calculate what happens to the business if revenues do increase by 10% next year, or if certain material costs end up decreasing by 5% next year, rather than just leaving it to estimation and having no idea how a certain scenario might actually affect the company. This is why all the best practices always point to ensuring the operating model is linked properly and you don't leave hardcodes everywhere - it's supposed to be more of a scenario analyzer rather than a crystal ball.

If you're sell-side, maybe yes. But on the buy-side things are different. For instance, just to give you an example, I'll take Howard Marks's thesis on risk (see below) which is one that I like. The graph, in a nutshell: The higher the potential return on an investment, the higher its risk deviation.

When you forecast 3 potential scenarios: Best, mediocre, and bad, you should take into account that all of those are possibilities and because two of them didn't end up happening it doesn't mean that they couldn't have happened (Fooled by Randomness is a great book about this topic). Ideally, a good financial model should allow you to put a number (%) on the possibility of different scenarios and calculate how much money you would be willing to invest to bear X% of risk for a Y% return. 

Alternatively, if you forecast, let's say, 5 different scenarios (good, half-good, medium, half-bad, bad), now you're closer to predicting where your investment may go, and if you apply a second level of thinking, you can prepare preventively on how to react when the numbers take an undesired path. The more different scenarios you forecast, the greater the chance of nailing the scenario that ends up happening. Also, ideally, when you forecast the undesired scenarios, you should consider preventively what options may be available in that circumstances (e.g. if you forecast a "bad" scenario model with certain numbers, you could already know who would be interested to buy the asset at that price range). Some may say that all of those steps are too much, but I don't like uncertainty so always having a number prepared upfront and an exit strategy for A, B, C ... Z scenario allows me to sleep better.

Also, in private equity and real estate, forecasting may be even more relevant (compared to public markets) because certain aspects of your investments fall within your control (improving operations, cutting costs, acquisitions, etc.) or understanding (business cycles) so you can forecast some numbers based on the pre-planned decision on the business's financials or how you expect the economy to go.

Of course, I ain't saying that modeling is the ultimate decision-maker in investments, but it's just an instrument to support your investment decision. Not necessary, but could be helpful.

The point is not the model itself but the sensitivities you run, that's the useful part imo

Sellside models are often inaccurate - just reverse engineering assumptions and inputs to goal seek to a predetermined output

a-non-hardo

Is the only reason we model to try and persuade people with non-finance backgrounds the company will do well?99% of models never come to fruition. We just do them to show what the business COULD produce? Find it funny how many different models people in M&A will put together when they legit never come to be true.Do we do it to justify our high salaries and large bonuses?Just think it's kind of funny is all. When someone will spend days on end modeling, then the MD will say it should have a higher valuation, so then the associate / analyst will juice up their projections even more, then in the end, none of it actually occurs post-close.Is this the reason we have jobs? Cause we always juice up businesses and say they're going to increase revenues by 10% the first year? Then 12% the 2nd? All backed up by some horse shit assumptions?

Also find it funny how nobody ever talks about this

Disagree here with you on this. I'd say I've seen my models come to fruition over the following 1-2 years from model creation with about 60-70% accuracy. I was in banking  for quite some time as well. 

Bankers are glorified salespeople. All of the work done during the deal process (making of the CIM, management presentations, valuation models) is to help justify the highest valuation for the company. Whether or not this actually comes true is an entirely different matter - both banks and the client want to sell the company at the highest price possible. So to answer "why do we even model?" - it's because modeling is a way of justifying the highest price so that the firm and its client can get paid the most.

Compare it to real estate - investment bankers are like investment sales teams working under brokers. Have I ever read a broker's OM forecasting 5% rent growth for the next 5 years on some dogshit "value-add" property and interpreted it to be an accurate prediction of an asset's future performance? Fuck no. Anyone who does deserves to lose money. But the underwriting serves as a sanity check and a justification of valuation - a starting point for the buyer and seller to find middle ground on price.

It is an insane amount of false precision. I truly believe that looking at historical income statements, hearing Management's growth projections and taking a non-analysis driven guess of what next year's Net Income will be, is more accurate than a 20 tab model the majority of the time.

The other thing that drives me crazy about models is sometimes the intangibles really are more important than historical trends. If your latest product/software is game-changing in terms of product value-prop and/or if you add one or two massive key accounts to your client roster, hockey-stick level growth really is possible. And it isn't going to follow some mathematical formula  like "Okay for every additional $1M of marketing spend we're going to increase our page views by A%, increase our page-view-to-customer-purchase conversion rate by B% and increase our average order size per customer by C%. It's just not how things shake out in real life. And from my personal experience, the bottom-line estimates that management gives you are often times far closer to correct than are the ultra-formulaic growth projections, with dozens of ultra-specific assumptions embedded, that bankers put together. Just my $0.02.  

CYA

IMO it is useful for scenario analysis / understanding a range of outcomes. What do you need to hit an IRR if X? What level of debt should we be able to service for this add-on? However really detailed models are stupid and false precision (both sell and buy-side). 

There's more to investment banking than sellside M&A. 

Models matter - a lot - in buyside advisory and lev fin. And even in sellside M&A they matter (again, a lot) in carveouts/divestitures, and LBOs. 

Ok, fine. Let's accept your premise that they're bullshit. Then what do you replace it with? Give us a better mousetrap. Until you come up with a better way to put a value on something, keep modeling.

Come model an E&P company, pretty grounded model comparably to DCFs / etc. with a myriad of far fetched assumptions

Honestly, you work in a casino and they keep changing the rules so wealthy people never have to pay up on a bad bet.

Analyst at a MMPE shop and sort of agree - I do think it can speak to the reputation/quality of a bank/bankers if, for example, they show something bad in their model. For example, saw a CIM the other day with a projected decline in revenue. Of course, EBITDA was trending up, but pretty ballsy (or maybe the MD was just stupid idk) to put declining revenue in the CIM. If we were to pursue the deal with this bank, and the MD didn't present as totally dumb on the phone, then this level of honesty in their model speaks to their character, which could be helpful with process dynamics. If they told us they had another buyer at $50m and we needed to get to $52m to sign it up, I would be more convinced that the banker was telling the truth and not just lying to squeeze extra $ out of us. 

But also agree with OP - have seen plenty of CIMs with historical revenue/EBITDA CAGR of 10-20%, projecting 60% for the next 5 years (often with not much backing up these claims). Screenshots of these instances flow through our teams messages with plenty of eyeroll and laughing emojis.
 

On a separate note, one of our VPs was telling us over lunch that some IB MD he knew went and sold some company to some PE group. And then the PE group hired the MD as CEO because he knew the ins and outs of the business better than anybody else. Could you imagine building that model, and then having to go execute on it? 

buy side credit here for different view - simple models that focus on key drivers are helpful. complex models with a million tabs are not. and we are focused almost entirely on downside scenarios, unless we have mezzanine and/or equity coinvests. 
 

when you are lending, you basically want to know “what needs to happen for this company to break, and when do I need to be at the table to make sure I don’t lose money here” - so it is super, super helpful to know that you are setting your covenants correctly and protecting your loan for your investors. the deeper in the capital stack you go, the more important it is. 

For the laughs. As a lender and previous buy-side analyst, it cracked me up seeing the sellside broker or investor CIMs with growth rates double above any historical references. Not that that isn't doable, but that without any context, it wasn't all that useful. The value comes from understanding what drives the models, as others have commented here. Is it a gross margin increase, is it wages, are there other levers that can be pulled that help bottom line cash flows, etc.... The models facilitate the understanding of these movements.