My explanation of our valuation expansion

So, even though development costs aren’t considered in calculating Gross Margin, they need to be considered.

From my perspective as an accountant one BIG FLAW in GAAP is expensing R&D for software development instead of capitalizing it. Here is why:

If you build a steel mill for a $10 million and depreciate it over ten years, you add $1 million to Cost of Good Sold (each year) which reduced Gross Profit by $1 million.

If you develop Windows for $10 million it gets expensed as R&D, you add ZERO to Cost of Good Sold which overstates Gross Profit (each year) by $1 million.

In other words, according to GAAP, Windows has no value! Does that make sense?

Denny Schlesinger

1 Like

Again, I disagree. Since R&D costs for developing software aren’t included in COGS

It has been years since I took financial accounting, but is not the cost of producing the software part of the cost of goods sold. From what I read, if the programmers efforts and hardware is associated with one product/project, the cost is then part of COGS.

From a site discussing software development:
"What is software cost made up? Software development cost is made up of time and effort estimation that are spent on the project. There are several components that make up a total cost. First of them are direct and indirect costs.

Direct costs comprise the income of every team player, payments for the hardware that is used, spendings on other things required for work. We can also state some sums that are spent on specific risks related to the project. Indirect costs comprise a control of the quality, audits, security issues. In simple words, they are not addressed for a specific project or its part.

Here are a few more things that make up a software cost:

Efforts that are spent by people on the needs of the project.
Tangible resources required for completing tasks.
Purchase and maintenance of equipment.
Software purchases and maintenance.
Renting an office, paying the bills."

Direct costs are computed in the COGS.

It would seem at first blush that the COGS would contain most of all costs for SaaS. Admin, executive pay, some hardware, depreciation, etc, would go toward Net profit.

What am I misunderstanding?


What am I misunderstanding?

In GAAP, R&D is expensed so it does not get into COGS. In Proper Accounting R&D to develop software should be capitalized and the amortization of the capital investment gets into COGS.

Funny thing, if a third party develops the software and expenses it according to GAAP, the cost of the software is not included in book value. Buy the company and you’ll pay the “real” price which is above book value. The difference between book value and the price you pay (some or most from the R&D expensing) goes into Goodwill and is then expensed, effectively making it’s way into COGS – sort of.

For investing purposes, FORGET ABOUT GAAP, it’s CRAP.

Denny Schlesinger


From my perspective as an accountant one BIG FLAW in GAAP is expensing R&D for software development instead of capitalizing it.

Actually, I have seen a few companies now starting to capitalize SW development and amortize it over time. If you look at their balance sheets and cash flow you will notice it.

1 Like

In other words, according to GAAP, Windows has no value! Does that make sense?

In my Ikea example, the design of the table is also expensed. The first reason it isn’t COGS is that the cost of design doesn’t vary with the number of tables being sold. It’s a one and done. So, while clearly that tangible thing - the table’s design - has value since Ikea will be selling tables with that design for years, the cost to create it is not a COGS. And for software there’s the additional issue I brought up that while a table design may be a one and done, software design almost never is. The company has to keep developers on the project to improve it and continue to make it relevant or sales will drop. That also makes it hard to consider as a COGS - harder than the Ikea example.

Even if you were to attempt to include design or R&D, the company may not be able to accurately predict when and if the benefits of developing will occur. The first year Ikea sells that table they don’t know if it’s going to be a hit or a dud. Do they include 10% in year 1 since it’ll be a product sold for 10 years, or 75% of the design cost since they’ll discontinue it next year?

Now, while what does/doesn’t go into COGS is fun for accountants and maybe produces inflated Gross Margins, it misses most my original point, which is that high Gross Margin rates are not a guarantee of company profitability. Factor One does not stand on its own, at least for me.


VEEV may be an interesting case study here.
“Only” 25% or so in rev growth, yet a high P/S of about 26 or so.
Their net income and operating income grew at over 50%.

So they appear as an example of a SaaS company showing earnings power as it ages. VEEV is about $1b runrate now.

ServiceNow (NOW) is another example.…
Sub revenue “only” 36% y/y, but FCF was up 40% y/y. They are forecasting $3.2b full year rev.

Whether AYX, ZS, ESTC, SMAR and others can get to $1b-2b/yr runrate is a separate question, but I think the market is modeling a future state that looks something like what they see from CRM, NOW, VEEV and other more mature companies.

If you are still growing 25-35% y/y, why not continue to grow your base at the expense of profits, so that you can eventually have larger profit dollars off a larger rev base down the road. FB, GOOGL, and AMZN all have shown how dramatically their profitability can grow as they scale, whether you look at it as profit dollars or profit %.

My quibble is just in how far-future-baked some of these valuations are, like with ZM. If you go into a stock hoping to get an eventual double or triple over X amount of years, and the starting point is $20b mkt cap at 60 P/S, that just seems a lot less likely to happen vs a $5-10b mkt cap at 16-30 P/S.



I have wanted to opine on this thread, but I’ve been very slow in collecting my thoughts. Unfortunately it’s veered into a GAAP / COGS discussion, but maybe we can get it back.

You see, it’s not about Gross Margin. That’s one component of the value of SaaS companies: they are software companies. But it’s not the only component, or even the most important. So let’s not dwell on it. What Saul has laid out is a cumulative case. It explains why these companies are valuable. I think we all believe that is true.

One way to look at this, as some have pointed out before, is the LTV to CAC ratio. With all SaaS companies, the LTV (life time value) of a customer is (or at least should be) several times that of the CAC (Cost of Acquiring a Customer). So we need to consider the LTV of a customer. Will the customer spend more in the future? SaaS customers usually do. Will they remain a customer for the foreseeable future? SaaS customers often do.

Also, there are other levers SaaS companies can pull, the most obvious being reducing OpEx and thereby increasing profitability. Just a few minutes ago Dreamer pointed to VEEV and the profits they are generating. Most SaaS companies focus on maximizing revenue growth, but apparently revenue growth in the 20’s and rapidly growing profits works too. The key is how long the revenue will be able to grow into the future, and how much leverage the company still has to drive more and more profit to the bottom line. This method has worked for PAYC too.

I guess I just take what I consider to be a common sense approach. Yes, our SaaS companies are amazing. Yes, they will probably continue to be. But yes, they are also about twice as expensive on average than they were two years ago. Maybe they are more fairly valued now than they were then. But they are certainly less of a bargain. So the issue is how much future growth you have to pay for now. In short, I very much agree with what Dreamer just said:

Whether AYX, ZS, ESTC, SMAR and others can get to $1b-2b/yr runrate is a separate question, but I think the market is modeling a future state that looks something like what they see from CRM, NOW, VEEV and other more mature companies. [Larger companies with a SaaS model have shown] how dramatically their profitability can grow as they scale, whether you look at it as profit dollars or profit %.

My quibble is just in how far-future-baked some of these valuations are, like with ZM. If you go into a stock hoping to get an eventual double or triple over X amount of years, and the starting point is $20b mkt cap at 60 P/S, that just seems a lot less likely to happen vs a $5-10b mkt cap at 16-30 P/S.

My common sense approach is to own some that are growing like weeds, but with PS ratios in the 20s, and some growing not as fast, but with PS ratios in the low to mid teens. (If I could find companies growing like weeds with PS ratios in the low to mid teens, obviously I’d prefer that…but those don’t seem to exist right now.) What I don’t seem to be able to convince myself to do is to own companies with a PS over 30 or 40 or 50 or 60. There are just too many other good companies on offer where you aren’t paying for so much future growth, even though I’m still expecting the growth (MDB, AYX, TWLO, ESTC, etc).

Just another Fool’s perspective. Good discussion, Saul.


PS As JAF pointed out, Fish’s reminder is crucial: If you invest like [Saul] you need to be on top of things. Having a concentrated portfolio does not allow you to sit back and buy and forget. Please remember this. You can’t just buy companies we discuss and hold them. You can’t even just buy what Saul is in. You have to evaluate your holdings daily.

I consider my method very similar to Saul’s, except that I have taken to holding some cash as valuations of all our companies have risen. There’s no shame in trimming a position as it grows over 10% or 15% and then considering buying some back if it drops. If it just keeps going up, you’ll still have enough. Trust me – that’s what’s been going on since late 2016.


No worries Saul!

Thanks for pushing back. I always appreciate that!

Right, my point wasn’t so much in getting the numbers spot-on with ZScaler as it was just thinking through what is priced in.

Your numbers reveal that it, in fact, may be pretty likely that ZScaler could surpass 35% growth for the next five years.

Assuming 45% growth instead, and keeping everything else the same, we get 16% CAGR. Much better, but still a tall task for any company. Especially in cyber-security where the larger a company is, the more of a target it becomes for hackers.

ZScaler will face more competition in the future so that’s why I chose 35% versus a higher number. But the key is, as I said before, you’ll be able to see that competition coming and get out before most people can if growth slows too much. That’s why I think staying on top of everything is so important.

Thanks Saul, no need to apologize. You’re just making me better.

All the best


While I tend to agree with Denny about the idea of capitalizing software, it may be that it makes less difference than it would seem. If Windows was something that one developed once and then sold in that form forever, then one would clearly want to capitalize it. But, for many types of software it is often the case that the investment in the software is actually pretty continuous. I.e., in year 1, one invests N in the software and perhaps releases and early version. But, then in year 2, one invests N again and releases a more full featured version. This tends to continue indefinitely, with N actually going up year after year as the company gains more revenue. The one place it conspicuously does not work is when a company is acquired and the acquiring company makes a minimal investment in the software for the purpose of harvesting the maintenance income.


Hi Saul,

First of all, let me tell you that I love your post, its content, and clarity of writing. I realize that this conversation has gone pretty far already, so I hope my post still makes sense in the context.
Here are my thoughts reading your post:

I always had to shrug a bit when you declared in the past that you don’t look at valuation metrics at all in your investments. But it was also very clear to me that you had a framework in your mind comparing the companies you invest in with “traditional” businesses that made you believe that these new companies have so much better fundamentals that make them many times more valuable. I would assume that the gap used to be so big in the past (when EV/S ratios were well below 20x) that you didn’t even bother trying to put a number on how much more worth these businesses should be. Now, that valuation multiples have expanded substantially again in the past months, it seems like this valuation gap is getting smaller and smaller, which makes it necessary to look a bit deeper into valuation considerations. I think you did a great job doing that in your post.

While I agree with everything you said in your post, I also feel we should not always focus on the positive traits of our companies (that justify a higher valuation multiple) but also talk about some risks that they entail (which could suggest a discount to the valuation multiple). Here are some risks that I sometimes wonder about:

  1. Profitability and investor sentiment changes – While I agree that our companies could probably turn a profit anytime they want (most of them are already cash flow positive for a longer time), it is still a fact that our companies, for the most part, are not profitable yet. That is certainly a risk to consider, especially in a more bearish environment. If everyone panics, they will more likely sell “speculative” positions which could start a negative swirl. I still don’t know what I would do in a recessionary environment, although if fundamentals stay intact I would probably do nothing and just hold/add. Investor sentiment is something we cannot really control or predict but it will probably impact our stocks more than other more “stable” companies. Also, I think our companies are not just trading on fundamentals, they are also held widely among momentum traders who are riding the wave of high returns. If they jump ship that could again negatively impact the stock prices.

  2. Share-based compensation and frugality – I’m a bit ambiguous on this part. On the one hand, I think it is a smart way to pay employees through SBC and it helps these growth companies achieve their goals quicker. On the other hand, I think they sometimes overdo it; i.e. they are much too generous with SBC almost to the extent of being wasteful – which makes me wonder how prudent they are in their general business affairs. Someone posted a video about some Okta-conference recently, where Jeff Lawson (Twilio CEO) was also on the stage. I don’t quite remember the context anymore, but I think they talked about some software perk some kid developed for free for Okta and everyone thought it was great. To which Lawson was saying something like “Who is that guy? You should write him a $50,000 check immediately! If you won’t do it, I will!” I have to say that I did cringe a bit when I heard that. Maybe I’m reading too much into that single event (they were on stage and all was fun and games, so he was probably joking), but I guess a frugal guy like Warren Buffett would never even think of a joke like that – it would never cross his mind.

  3. Disruption is happening faster and faster these days. The times where SAP and Oracle locked-in their customers around the globe for decades are over in my opinion. Sometimes I wonder how strong the lock-in-effect of our companies really is and what their other durable competitive advantages are. It’s easy for me to see these traits in companies like Netflix, Amazon, Google, Mastercard, and even Apple – which is why I can (and do) hold these positions with high confidence for the long-term. I love my AYX, ZS, MDB, TWLO, etc. but I also think these are much more fickle. We saw that with the TWLO/Uber-incident, the MDB/Amazon-worries. While I’m a strong believer in the current superiority of these SaaS/SW companies, I also think that the future for these companies is not as predictable as their current market caps might suggest.

So, yes, today a company might be worth 30 times sales, probably much more if the future plays out as planned. But tomorrow might be a completely different story. What I try to remind myself is that these companies’ outlook can change on a dime. +20-30 times sales is always a best-case valuation, but best-case is not reality; it doesn’t always play out like that. In that regard, I also try to remind myself that (growth) stock valuations are nothing absolute – they are inherently relative and bound to swing strongly over time.

Best wishes,


First of all, let me tell you that I love your post, its content, and clarity of writing. I realize that this conversation has gone pretty far already, so I hope my post still makes sense in the context.
Here are my thoughts reading your post…

Thanks Niki, excellent analysis and addition to my thoughts.


1 Like

As usual, I’m a few days behind the curve here. While there are several lengthy and interesting posts in this thread, I wanted to respond the Niki’s post in particular.

IMO, your first point is simply observing that these stocks tend to be more volatile than older/larger companies. You more or less acknowledged that in the end, assuming the fundamental story had not changed, you’d probably take little action as your portfolio declined in value. Possibly, you would add to certain positions. Volatility increases with the rate of growth. I don’t have that stats to prove that, but it seems pretty clear as a heuristic. As the vagaries of market sentiment drive wild swings in stock prices causes ulcers for you, probably the best thing to do is to avoid these investments altogether. It comes with the territory. When the next recessionary market hits (could be sooner than we think), it’s going to take some stamina to sit by for the duration (average is 18 months) and do little.

Your second point does not apply to all our positions equally. Each company handles this differently. As such, I’m not sure what to make of it. I’ve not tried to calculate SBC/employee for each of my positions to see how it stacks up against an industry average. Never thought of it in those terms. Even if I did, I’m not confident I would know how much weight to give it when it comes to buy/sell decisions. As Duma pointed out in a different thread, there’s at least some evidence that much of the information taken into consideration is actually of little value. A lot of folks discounted his observations as statistically invalid due to sample size, companies selected, etc. All valid criticisms, but the basic point he tried to illustrate I think valid. Used to be, getting relevant and timely information for the purpose of making investment decisions was difficult. The reverse is true today. We’re inundated with information. Separating the important from the trivial is no mean task. Is SBC even worth considering? I’m not sure . . .

Finally, your third point, I think is born from a lack of knowledge about what makes an IT product sticky. It goes to the heart of the land and expand strategy practiced by most of not all the SaaS companies discussed here. I was a little slow to get into AYX. Frankly, I didn’t see the moat. From a product perspective, I still think their moat is not all that deep or wide. A competitor could offer a similar product. But, it wouldn’t be easy. There’s some serious R&D behind the Alteryx offerings. But, the thing that really cements their market leadership is the network effect. Their products are complex and sophisticated, they come with a serious learning curve. The more customers they gather, the more they enhance the overall available skill set. And the more they penetrate their market, the more they enhance the user blogs, non-company training, tips and tricks, etc. If you’re in the market for data analysis tools (a fast growing market at that) an Alteryx competitor has to have great offerings at lower prices to compensate for the dearth of experience and available skills. This is a very big deal.

Our two cyber-security companies, ZS and OKTA have a different kind of stickiness (although they also benefit from the network effect). These companies have products that become embedded in operations of their clients. As disruptors, it is not too hard for them to dislodge those companies that are currently in the field as they have offerings that are more effective with less burden on end users as well as the IT organizations that install and maintain the products. MDB has already won the no-SQL war, of course they have competitors, and for some use cases Cassandra or some other no-SQL DBMS will be a better fit, but Mongo is pretty much in the same position Oracle was in from the 1980s up until recently. Few products get more embedded in business operations to the extent that a DBMS does.

What I’m getting at is, s/w stickiness is based on a lot more than price. And once certain s/w products are brought in house their stickiness goes way beyond functionality. As with SBC there is nothing universal or inherent about this. It needs to be evaluated product by product, company by company. I have much less confidence in SMAR or ZM than I do in ZS, OKTA or MDB. I try and keep my position percentages in line with my confidence levels.


Implicit in the background of this as well as many other threads on this board; the relevance of a fairly simple but extremely powerful research concept known as convergent validity. I see it all the time here. So, we have rational/logic based considerations such as product “stickiness of product“ or “disruptor” status or “moat”. And we have quantitative considerations such as yoy gross margins growth and PEG. And we have, of course, the on-going testing of hypotheses in the form of results of Saul’s (and others’)investment theses (aka portfolio results). There are many other examples. The point is that convergent validity always compels high(er)confidence in predictions. And here it is coupled with the general willingness of board members to share and embrace scrutiny and divergence of opinion. It is hard to imagine a more robust(in a statistical frame) approach than Saul’s community and system.



I had to google “convergent validity” in order to understand your post. I agree with you. I’ve learned a great deal since I started to follow Saul’s board which was not long after its inception. This board with it’s mixture of numerical analysis along with more squishy analysis of business and product attributes (and sometimes even evaluations of specific executives) is my primary investment research resource. My portfolio results attest to the fact that it is a great resource.

In the end, I still make the decisions to buy/sell/allocate. I just make much more informed investment decisions than I ever did before I became a member of this cyber-community.


I try and keep my position percentages in line with my confidence levels. My problem has been that confidence levels do not necessarily correlate with future performance of the stock. google definition of “confidence” .

the feeling or belief that one can rely on someone or something; firm trust.
“we had every confidence in the staff”
synonyms: trust, belief, faith, credence, conviction; More
the state of feeling certain about the truth of something.
“it is not possible to say with confidence how much of the increase in sea levels is due to melting glaciers”
note the word "feeling"in there. Confidence is thus mostly emotional in basis

1 Like

Aren’t you all glad right now that you didn’t get panicked into selling out on Monday when the world was coming to an end?


It’s crazy!
An average investor like myself a year ago would have lost a return opp of one year returns in the last three days…if one year and not five or nine which is the case for IPSA Chile Index.

Thanks Saul, MF and everybody here, u all have change my life.

Best from Chile,



Yes and no. I’ll grant that there is no objective measure for “confidence.” You can’t buy a confidence meter on Amazon. So, yes, it’s subjective, even “emotional” is you like to use a word that makes it sound like an arbitrary judgment.

I would suggest that their attributes that contribute to one’s “feeling” of confidence. Some of those attributes are objectively measurable, other are not, but have shown over time to have a pretty good correlation to a company’s performance. Things like executive bios. Things like the tone and temperament of conference calls. Things like product positioning and the competitive landscape. Things like regulatory oversight, etc., etc., etc. All those things add up and contribute to one’s evaluation of confidence.

You might have noticed Saul has a six star scale which pretty much equates to his level of confidence. And he too applies his confidence rating to his allocation decisions. I think confidence is a useful, even important investment tool, even though it is ultimately a subjective one.

But, feel free to enjoy confidence in all your investment decisions even if eschew a sense of confidence in the company too subjective to be useful. Once you get it down to the perfect set of objective measures you can write an algorithm and put your investments on autopilot.


I wavered but held fast. That I had made plan not to sell, long before the swoon, helped a lot.

This community is life saver.



Thank you Saul for explaining! I’m still dipping my toes into the waters of your type of investing, and I’m wondering, for someone just starting to invest in this way, is now a good time to start buying SaaS stocks? I guess now is better than later? Since we know we can’t time the market and the financials for these SaaS companies you have reviewed meet the requirements set forth in your article (having a niche/unique offering; high growth rate; high gross profit margin; low debt; being “undervalued”) and by waiting, one is missing out on that growth which makes these companies so profitable to invest in?