The main reason companies aren’t allowed to capitalize their research and development costs is …
Edison had the same problem developing the light bulb! Some 900 trials before getting it right? ALL THAT MONEY WAS the cost of developing the light bulb! There would be nothing wrong with amortizing that cost over a ten year period either as Cost of Product Sold if they had sales, or as a loss if not. The EggHeads that manage GAAP should try running a real business!
Denny Schlesinger
1) Total Addressable Market
2) Terminal Growth Rate
Right on! Those two determine the top of the “S” growth curve!
Denny Schlesinger
The problem with estimating TAM for many of the products discussed here is that many are essentially new products creating new markets that didn’t exist before or, at the least, significantly changing a market so that it becomes something different.
This is a great discussion and shows why TMF is such a great resource. I have just a few points to add.
Aside from R&D, most of these companies spend a tremendous amount on SG&A. In looking at a few of them ZS, AYC, and MDB, they spend close to or sometimes more than their entire Gross Profit. Yes, perhaps this spend will taper over time, but how much of this spend is related to maintaining existing business and how much of it is related to garnering new business?
These “high” PS valuations remind me of a quote from Scott McNeely, CEO of Sun Microsystems after the tech bubble:
‘At 10 times revenues, to give you a 10-year payback, I have to pay you 100% of revenues for 10 straight years in dividends. That assumes I can get that by my shareholders. That assumes I have zero cost of goods sold, which is very hard for a computer company. That assumes zero expenses, which is really hard with 39,000 employees. That assumes I pay no taxes, which is very hard. And that assumes you pay no taxes on your dividends, which is kind of illegal. And that assumes with zero R&D for the next 10 years, I can maintain the current revenue run rate. Now, having done that, would any of you like to buy my stock at $64? Do you realize how ridiculous those basic assumptions are? You don’t need any transparency. You don’t need any footnotes. What were you thinking?’
Yes these are software companies and not hardware like Sun. Some of these companies will overcome these hurdles and offer a nice return over the next several years, however, I think it is reasonable to expect also that some of them will re-rate substantially lower as they mature.
-Sameer
Edison had the same problem developing the light bulb! Some 900 trials before getting it right? ALL THAT MONEY WAS the cost of developing the light bulb! There would be nothing wrong with amortizing that cost over a ten year period either as Cost of Product Sold if they had sales, or as a loss if not. The EggHeads that manage GAAP should try running a real business!
I can’t even imagine the crap businesses would be loading onto their balance sheets and amortizing if the accounting rules to do so were more liberal … conservatism is one of the conventions GAAP is built upon and I think in the end, all things considered, it works out in the investors benefit.
There is no magic in capitalizing R&D, just a recognition of the time period over which the expense is relevant, just like revenue in a long term contract comes in over a period of years. It might help net profit in year 1, but what about in year 5 when one is expensing some of year 1, some of year 2, some of year 3, some of year 4, and some of year 5? No guarantee that will be less than the total of year 5 expenditures on R&D, particularly if the R&D is front loaded, as may be the case developing a new product.
‘At 10 times revenues, to give you a 10-year payback, I have to pay you 100% of revenues for 10 straight years in dividends. That assumes I can get that by my shareholders. That assumes I have zero cost of goods sold, which is very hard for a computer company. That assumes zero expenses, which is really hard with 39,000 employees. That assumes I pay no taxes, which is very hard. And that assumes you pay no taxes on your dividends, which is kind of illegal. And that assumes with zero R&D for the next 10 years, I can maintain the current revenue run rate. Now, having done that, would any of you like to buy my stock at $64? Do you realize how ridiculous those basic assumptions are? You don’t need any transparency. You don’t need any footnotes. What were you thinking?’
Scott McNeely was the CEO of a company that went bust! Check out the above for the word “growth.” McNeely is talking about a company with ZERO growth, not our SaaS enterprises for sure. It’s the same nonsense rhetoric my broker used to use when he called it “discounting earnings to Eternity.” Listen to Albert Einstein instead, a man who knew something about mathematics.
Compound Interest - The Most Powerful Force in the Universe
https://www.cbsnews.com/news/compound-interest-the-most-powe…
The problem, of course, is that the SaaS companies will not live for 98 years. Growing revenue at 50% they will saturate the market in a few years UNLESS the market is also growing. Material things have a hard time growing fast but data does not have that limitation. There might be a limit to the number of people on Earth but what is the limit of data that they can produce?
Don’t let the wrong math make you bearish.
Denny Schlesinger
all things considered, it works out in the investors benefit.
It’s a huge tax break for Silicon Valley but it’s still bad accounting.
Denny Schlesinger
There might be a limit to the number of people on Earth but what is the limit of data that they can produce?
Used to do client events with funny old-timer on storage side of house, that would open his pitch with interesting facts around data growth, including things like how much data the weather service or even a single airplane (with all their sensors) produces every single day and how it is always growing.
Then he would ask something like “what is the % of data that has been deleted globally?” or something goofy like that, and the answer was “we don’t know…no one deletes anything!”
Cheesy, but served a point.
Not only are we producing massive increases in data every year, and there is no reason to expect that would ever stop happening, but we also still pile that new data on top of all the old stuff we have gathering since the beginning of data storage.
Data analytics and the insights derived from data will be a perpetual activity as you always have new data to mine/compare. Oracle became a titan because of this. SAP the same. Google/facebook leveraged data with their “free to use” models for us consumers to become titans. Amazon built AWS to house data in a more agile manner for many companies. If we pick the winners here, we should do well. And there should always be more/new winners too.
Dreamer (long AYX, ESTC, TTD, and MDB…all very much “data” companies)
All that data has to be stored somewhere, and that something is hardware, which the people here seem to have a distaste for. So if you really believe that data can grown forever and is never deleted, start buying the hard drive makers, start buying the flash drive makers, start buying magnetic digital tape. Data MUST have a physical component in hardware, in some manner. This will, by the way, limit how much data can be generated and kept. As an extreme example the Large Hadron Collider must throw away a very large amount of generated data because even they cannot store it away fast enough to keep it, let alone analyze it. The irony there is that the collider might have already discovered new physics and they threw away the data because they could not keep it.
So if your premise to valuation is that data can grow without limit you need to re-evaluate your premise. It’s simply wrong.
Hard drives. Data has grown at an exponential rate for decades. Yet the cost of data storage per gigabyte has fallen at least as fast if not faster. It is not true that profits of data companies go up simply because data grows.
Same with computing power. You always get more for less. The difference with computing power is that the economics stay way ahead as the power of computers continually improve enough to offset the drop in prices per computing power.
Software has even better economics. But there economy would soon be overwhelmed if the cost of data analysis, storage, and software went up uniformly with the rise of data volume. A problem Splunk has to deal with.
So no, growing data does not necessarily mean growing riches for those who store data, any more than growing data across the internet enriches the optical companies who build the data highways.
In the end it is all about CAP. Something proprietary that you do or have that others cannot equal so that your returns exceed your cost of capital for long term returns.
Tinker
I’m afraid my comment about data has been taken out of context. Saul has been saying “stay away from hardware” and that includes disks, flash, and optic fiber.
Indeed you need “Something proprietary that you do or have that others cannot equal so that your returns exceed your cost of capital for long term returns.” Poster boy, MongoDB. What do they do? CRUNCH DATA! Endless streams of data much of which is perishable and needs not be stored for long. Not the same kind of data collected by the big collider.
Denny Schlesinger
Hi bjurasz
I get what you are trying to say but I think there are 2 forced at work here.
1 is the Moore’s law regarding the growth in economic computing power - the other is the simple point that I think Gates used to make (who I personally think is under rated for his insights and business wisdom), that software is an appreciating asset whilst hardware is a depreciating asset.
(FWIW I also hold Pure Storage and Micron - and back in the day held EMC and NTAP but not because it is hardware).
A
Those are good points people, in rebuttal to my claim about data needing to reside in something physical. Indeed data MUST reside on something physical. But most data has a shelf life too.
This has been a very good thread. For me, personally, I’m realizing that this investing style has a risk profile that I only want for a portion of my portfolio. Currently its about 36%, under half what it was 2 months ago. Some here want to be all-in out of fear of missing out on big gains while knowingly risking a big downturn. I’m running opposite, happy with the gains I’ve had over the last 10 months and willing to miss out on larger gains in the future in exchange for more safety on the downside.
Best of luck to all, in whatever investing style suits you. 
I guess i am here for the dumb questions.
I think I understood everything Saul mentioned in his original message, if not please enlighten me.
But when is the point reached, that a SAAS company is finally overvalued? If we don’t know that yet, are we just hoping to sell on that final day before everyone else does?
mooo
Novice Stock Investor
It isn’t that we expect it to become overvalued. What we expect is that as the company reaches the upper part of the S curve and has something like 85% of the expected TAM that the growth will slow. As the growth slows, so will the multiple. When we see the growth slowing, we get out … though not all of us at the same point!
Asked offline:
So by “growth” we are talking about revenue growth?
And that’s why Saul got out of shopify already?
And if we are talking about revenue growth, i checked the numbers for e.g. Okta.
For Octa the revenue growth rate is also slowing. Does it mean, this is hitting the upper part of the “S curve”?
If i am seeing it wrong, how can i check if “growth” is slowing or if the company hit the upper part of the S curve?
This is why I said, “not all of us at the same point”. The classic S curve is 15% slowly growing revenue during early adoption, more or less straight line rapid growth during the “tornado”, and then 15% tapering growth as the product reaches saturation. But, things are never that simple. In particular, one expects that the larger a company grows, the more difficult it will be to expand the revenue each year at the same rate. So, some tapering is to be expected. The question is the rate. With Shopify, it appears that Saul’s judgement was that it was falling off more rapidly than he was comfortable with. The danger, of course, is if the fall off suddenly catches the market badly, then the share price will be trimmed dramatically because the multiple the market is willing to give will go way down. But, this is a matter of market perception, not scientific rules. So, one company can continue to have its stock price go up and up even though there is some tempering of the growth rate while another may receive a dramatic haircut because of a shortfall relative to expectations, even through the rate is still high.
I haven’t posted on this thread on explaining the evaluation of our stocks for some time because I have no interest in arguing with people who have already made up their minds. I figure that I have tried to explain, but I also feel that there’s no point in continuing. I’m worn out with trying. I’ll just post what I’m doing and what the results are.
I started this thread early Monday morning:
In the 2.5 trading days since, while all this discussion has been going on, my portfolio year-to-date has gone from +57.7% to +63.4%… (gained 5.7 points).
In those 2.5 trading days, my portfolio since the beginning of 2017 went from +398.9% to +415.9%… (which is 516% of where I started, or over a quintuple).
What else is there to say? Those are the results, and I don’t sell companies because their stock prices have gone up.
Saul
The valuation of high growth SAAS companies present an example of the St Petersburg Paradox.
We are currently in an environment of very low risk free rates of return. At the same time, we have SAAS businesses with extraordinarily high growth rates over unusually long time frames.
One could argue if these circumstances persist, these business are, theoretically, of infinite value.
They are certainly of extraordinary value if you choose well, as articulated in this insightful review by a software VC https://medium.com/@alexfclayton/history-of-public-saas-retu…
The average return for all 67 pure-play SaaS/cloud companies that have traded publicly is 4.7x from IPO price as of 9-July-2019.