Everyone is worried about valuation!

[FlyingCircus]

Whether or not they are fairly valued, overvalued or undervalued is, as Saul says, difficult to judge because of the new paradigm these businesses fall in.

The point is about the risk: they are all valued simplistically with an assumption of continued extreme revenue growth in the short term. (Because, paradoxically, traditional valuation metrics cannot be applied reliably.) And there is pure momentum at play. (Today’s polite version of “the greater fool” model.)

The risk is, the second the news hits the wire that revenue growth has slowed even a little, these stocks will get taken out back and shot on the spot.

Reasonably tight trailing stops might be a good idea for those stocks with the most stretched valuations.

[XMFBreakerTinker]

Because the story can, and likely will, change/evolve, it seems fair to wonder how long the market will value something at 30x or 40x or 60x. I had to check just now…ANET still below Jan 2018 prices, when I first came to the board wondering how/why ANET should be valued so highly.

The story with ANET changed dramatically when its finest moment in the sun, 100gb/sec, turned out to be far less than the grand tornado it was billed to be. At that point in time customer concentration became an issue, and future revenue growth became an issue, and thus I sold in early 2018. Gee, in real time as well.

Not based on market technicals, on generalities, on it being overvalued, undervalued, etc., but based on company specific fundamentals. The narrative did not equate to reality.

With Zscaler, Dreamer is correct I believe. On the Slack board I’ve run through it and mentioned it here. Why does not someone else run the numbers and discover for themselves that what seems overvalued (in an obnoxious sense) may be clearly undervalued.

I say this based upon its long-term CAP (literally not direct competition of any concern), tremendous transformative results, unique product results, growth accelerating and much higher than what your calculations will tell you that a 3 or 4 year CAGR needs to be, extremely high revenues, increasing returns of scale, near zero customer churn (basically going out of business, being bought and for some reason moving back to the old corporate database system of the acquirer), or losing some seats due to layoffs or the like, with recurring revenue that is likely to last a decade or more, until the companies products are disrupted (if ever).

Markets go up and down, but examine Zs for yourself. Let me know, AFTER, you run numbers that are materially below present expectations, going 3 and then 4 years out (hint: $1 billion in revs is quite likely in year 3) and then comment

There is always risk, uncertainty, doubt in any investment. Look at GE for crying out loud, widows and orphan investment my arse.

On a risk/reward, let me know.

As far as Zoom, myself, personally, I agree with Dreamer. Simply on risk/reward. I want investments that will grow back to where they are even if the market crashes here and there. I cannot say Zoom will do that any more than Amazon did (hint: took an awful long time).

Tinker

[captainccs]

and that growth can continue at these rates for near infinity is approaching the Kool Aid level.

bjurasz

My brand of KoolAid is the “S” curve:

That marriage happens in the “S” curve which describes growth in the real world. 100% growth cannot go on forever. The limiting factor is market size. The “S” curve shows a slow initial uptake by technology aficionados followed by early adopters. Once market penetration reaches about 15% the pragmatist buyers come in and create “the curve in the hockey stick” that forms the bottom of the “S.” At this point we see the rapid growth illustrated by GauchoChris. That fantastic run has to end eventually which is PaulWBryant’s concern. The top of the “S” curve starts to set in with around 85% market penetration after which growth switches to a vegetative rate (replacement and new buyers).

Denny Schlesinger

[brittlerock]

Bjurasz,
If the economy slows, for example, you’re going to see demand for these services drop.

Just curious, on what evidence do make this assertion? I know, it seems intuitive, if the economy slows, demand for everything should slow, so these stocks (as a generality) will be hit as well.

I don’t own the market, I own a very few, carefully selected companies. Explain to me why Zscaler for example will take a hit with a slowing economy. What company is going to put internet security on hold, or cancel their Zscaler subscription? Alteryx, Mongo, Okta, Twilio etc., what company is going to discontinue these services in a slowing economy. I could make an argument that just the opposite might be true. Each of these companies offers vital services for a lower price tag than self provisioning. And, you get the service right now (more or less) rather than having to make a big upfront R&D investment.

I will grant that there are “other factors” though. Please name a company that is not subject to some external, unpredictable “other factor”. I designed an information policy system at the company I worked at. I had an entire subsystem devoted to determining whether an external factor (such as new legislation, a court decision etc.) impacted which policies (the relationship is seldom direct). Those policies were red flagged them for review.

[captainccs]

Sorry, forgot to post the link:

https://discussion.fool.com/one-has-to-marry-the-purity-of-mathe…

And while I’m at it, a response to ValueNauts: We all know what intrinsic value is, we just don’t know how to calculate it. Let’s take a pound of gold. It’s a pure metal, noble, it does not rust. It just sits there in a bank vault for decades, inalterable. Has the intrinsic value changed? How could it, it’s the exact same atoms sitting there, minding their own business, inalterable.

It the above is true, how come the price on gold exchanges varies from trade to trade? The only thing that can be influencing that price is Supply and Demand. If Supply and Demand is good enough for gold, it’s good enough for SaaS shares!

I highly recommend Reinventing the Bazaar: A Natural History of Markets by John McMillan. It’s a fun read! No higher mathematics required.

https://www.amazon.com/Reinventing-Bazaar-Natural-History-Ma…

Denny Schlesinger

[LifeOfDreamer]

Explain to me why Zscaler for example will take a hit with a slowing economy. What company is going to put internet security on hold, or cancel their Zscaler subscription? Alteryx, Mongo, Okta, Twilio etc., what company is going to discontinue these services in a slowing economy. I could make an argument that just the opposite might be true. Each of these companies offers vital services for a lower price tag than self provisioning. And, you get the service right now (more or less) rather than having to make a big upfront R&D investment.

---

None of these stocks are priced for the business they already have, and retain, today. Rather, they are expected to keep growing.

I keep seeing how “SaaS” is revolutionary, but it is largely a change in delivery system (software vs physical discs or hardware) and a change in procurement (OpEx vs CapEx).
Their customers, and potential customers, still have budgets.

Just like a consumer can only afford so many streaming services before they realize they are not only NOT saving money from cord-cutting cable, but they are actually paying more than they were before.

In a slowing economy, businesses will look to retain vs focus on growth…not saying that is the right approach, but tends to be reality. You cut costs via employees or delaying/cancelling new projects and maybe downsizing existing spend where possible.

So if ZS was expected to sign X number of new clients in a solid healthy economy, they may likewise be expected to sign X-(Y) number of new clients, with “Y” equaling number of clients that decided to hold off on making a change due to switching costs (perceived or real). They may choose not to move to SSO, and instead say “let’s wait for next year” costing OKTA a few new clients they otherwise would have had. While I believe leveraging data to create competitive business advantages should be mission-critical for most/all companies, the reality is that it isn’t, and AYX could easily not close a few new client deals in a downturn as a result of short-sidedness of a client’s execs on how to navigate a downturn.

I hear all the corporate cliches, and one often thrown around when things seem tight or in a correction is “hey…you can’t CUT your way to GROWTH!” in an attempt to spare oneself or one’s team from the chopping block or to spare one’s project from the chopping block. Sadly we have a corporate structure where CEOs get golden parachutes and BODs worry about bottom-line. So even if our CEO’s are awesome and passionate founders that seek growth at all costs, their clients aren’t necessarily the same unfortunately.

The Everything-as-a-Service model is getting a bit overdone, and Enterprise procurement teams aren’t stupid. HPE, a hardware company by and large, just had their annual conference and one of the main takeaways is that the CEO stated by 2022 they shouldwill sell everything “aaS”.
https://www.datacenterknowledge.com/hewlett-packard-enterpri…

I am not arguing against ZS, just that SaaS by itself is getting too much credit. Look at ZUO stock, and we can see market discriminates, as it should.

There are big macro cycles and changes occurring around IT/software and I think we get to a point where cloud is the new hardware (commodity) and our companies continue to be valuable as long as we are selective.

The old lock-in was you didn’t want Cisco or HPE or DellEMC for all your hardware needs. It won’t be any different, imo, for cloud eventually. You won’t want AWS or Azure or GCP for everything you need. Convenience comes at a price.

What is the saying? I can be fast, cheap, and good…but only 2 at the same time.

Dreamer

[thejusticier]

about GM and valuation… so what do you make of PagerDuty? it has very high GM much higher than Twilio for example and has been growing gangbusters.

tj

[Tamhas]

I keep seeing how “SaaS” is revolutionary, but it is largely a change in delivery system (software vs physical discs or hardware) and a change in procurement (OpEx vs CapEx).

It is rather more than that. Typically the software part of the SaaS subscription is equivalent to the perpetual price spread over 3-4 years. This means that, once one gets past that period, that one keeps taking in the same amount year after year without having to make new sales.

Much software also has a maintenance cost which provides bug fixes and may provide entitlement to new releases. Consumer software often has no or a low maintenance cost, but one has to pay for new releases, usually at a discounted level. Enterprise software is more likely to have a higher maintenance cost and include right to new versions. This is usually on the order of 15-20% of the perpetual license cost per year, so the SaaS billing is going to include this in addition to the amount for the license itself.

It is also the case that many times the perpetual license has no good mechanism for monitoring user count so the user count can increase without additional licenses being bought. By nature, a SaaS system is going to track user count so increased use will naturally lead to increased billings.

So, just the SaaS licensing will lead to higher on-going billings, potentially quite significant.

Moreover, if there is a cloud delivery instead of on prem installation, then the vendor company will save significantly in the cost of bringing a new customer on line and in the on-going difficulty of maintenance and support since everything is happening on the vendor’s machines (whether owned or rented) rather than the customer’s machines. Possibly significantly, this also means no customer mucking about in the system doing strange things that then need to be diagnosed and repaired.

So, with cloud delivered SaaS we have a business model which not only provides for higher levels of on-going revenue, but also simplified and lower cost support for higher margins.

[Kevin2017]

I keep seeing how “SaaS” is revolutionary, but it is largely a change in delivery system (software vs physical discs or hardware) and a change in procurement (OpEx vs CapEx).

This change of delivery system make the revenue looks smaller than it really is. For example, the company make a sale of subscription this year for annual sub fee of $100. So the revenue this year is $100. But the sale also guarantee the subscription income in the future years which is not recognized in the current year revenue. If the sub last 20 years and we discount future fees at 10% rate to current year, the present value of the sale is $936. So the real revenue this year is $936 not $100. This makes EV/S 9.36 times higher than the real EV/S.

[LifeOfDreamer]

sale also guarantee the subscription income in the future years which is not recognized in the current year revenue. If the sub last 20 years and we discount future fees at 10% rate to current year, the present value of the sale is $936. So the real revenue this year is $936 not $100. This makes EV/S 9.36 times higher than the real EV/S.

---

Not sure I get the point of using 20 year as example. No client signs 20 years.
I give up talking about this. If anyone needs me, I will be on my board. Thanks.

Dreamer

[XMFBreakerTinker]

In a real recession practically all growth will slow, except…we actually are in a slow down now, world wide. You can see it in chips, storage, Fed Ex, consumer goods, but Zscaler and Mongo are accelerating growth during this period and Mongo reached a 52 week high in the midst of a “historic” end of year market crash.

True, it is a slow down and not negative growth, but then again pre-Trump, 2015-2016 growth was slower and there may have even been a negative quarter of two and yet, did not impact SHOP at all.

Zuo is simply a bad business, Cloudera a bad business, Nutanix hit a business wall OEM wise when Dell turned on them and the subscription model did not compensate, Arista peaked out on its S curve after fantastic business, as did Nvidia (for now anyways) but SHOP is a great business. Their business results occurred during the same economic circumstances but the results incredibly diverged.

I think we worry too much about macroeconomics in regard.

But true in a real and deep recession, growth will be slower in almost every company than it otherwise would be and perhaps materially so.

Thus why it is best to stick w the best, while on their S curves or that remain dominant w no real competition (say Salesforce or Amazon) as they will come back and perhaps even hold up best in the end. But make no mistake about it, remember Oct and Dec 2018. We will see that again whether or not we have a real recession.

Tinker

[Kevin2017]

If we change to 5 years, the total sale should be $417. So the real EV/S is current EV/S divided by 4.17.

[qazulight]

I think we worry too much about macroeconomics in regard.

I have used Macro Economics. It is why I am not retired.

Saul’s approach is the one that works. Macro Economics is great for choosing a career, or when to build a house, killing time on the internet.

However, if one is going to make money, one must continually investigate companies for possible investments using the methods Saul has outlined.

This being said, one must invest with his comfort level. Currently mine is 40 percent cash.

Cheers
Qazulight

[branmin]

But make no mistake about it, remember Oct and Dec 2018. We will see that again whether or not we have a real recession.

Tinker

I’m sure we will. I’m rather hoping that it happens due to picking up high conviction stocks at an unjustified lower price simply due to market conditions fully knowing where they might reach again in a hopefully relative quick time.
Rinse and Repeat.

Just not worrying about it all anymore, particularly valuation or when a recession will occur. If it happens it happens, but we will be aware of the signs.

[Smorgasbord1]

Revenue that has very high gross margins is worth more per dollar of current revenue…
Do you disagree with any of that? No. Okay, let’s go on!

{meekly raising my hand}

Umm, high gross margins/gross profits are great and all, but all software companies have always had high gross margins for the standard definition of gross margins. Yet, not all software companies are great businesses. It’s true that one difference is that delivery of software as a service over the internet is new, but I don’t think all SaaS companies are great businesses, either.

Remember, calculation of Gross Margin excludes R&D. R&D is typically a large expense for software companies, and increases as the product is sold to more customers. Even in GAAP, typically only things like stock-based compensation are factored in, not the entire R&D effort.

Take Mongodb for example. https://www.prnewswire.com/news-releases/mongodb-inc-announc…

For 2019Q1, revenue was $89.4M. Cost of Revenue was $28.2M, for a Gross Profit of $61.2. But, R&D alone was $30.9M, which cuts the effective Gross Profit in half. Factoring in other Operating Expenses, total loss was $33.2M.

Comparing to 2018Q1, revenue was $46M. Cost of Revenue was $13.8M, for a Gross Profit of $36.4M. R&D alone was $18.6M, which again cuts the effective Gross Profit in half. What’s most interesting is that R&D expenses have pretty much scaled with revenue over the last year, which is what you’d expect from an actual COGS. R&D increased 66% while SG&A only increased 37% - more indication that R&D for software needs to be treated more as a COGS.

Indeed, additional R&D is needed to sell more. Software isn’t a chair you design once then sell - it’s a thing you maintain, fix, tweak, and improve every day.

With all the hand wringing over Zoom, for instance, it’s interesting to note that company only spends 10% of total revenue on R&D (https://www.cnbc.com/2019/03/26/zoom-key-profit-driver-ahead… ). That compares extremely well to Mongodb’s 35% R&D spend, or Atlassian’s 40%. These costs don’t show up in the Gross Profit, but they do show up in Operating Profit. Zoom’s relatively low R&D costs mean it’s doing better than the raw numbers show. But, it means Mongodb and Atlassian are doing less well.

In short, a Gross Margin number for software companies can’t be taken in isolation. Accounting rules are still based on selling physical goods or providing labor services, and aren’t good with the new delivery as a service model.

[Tamhas]

The whole idea of R&D in GM seems off to me. R&D is what you spend to get the product. With software in particular it tends to be a fixed cost whether you end up selling 1 or a million copies of the software. Why would that be in GM?

Now, treating it as a asset and depreciating it … that makes sense.

[Smorgasbord1]

Now, treating it as a asset and depreciating it … that makes sense.

The problem with that, as described here: https://smallbusiness.chron.com/rd-expenses-not-capitalized-… ) is:

The main reason companies aren’t allowed to capitalize their research and development costs is that there’s no way to reliably measure the future economic benefits of those costs. R&D involves trial and error – a lot of error. When it set the rules for R&D spending in the 1970s, the accounting standards board cited data showing that only 2 percent of product ideas become commercially viable, and only 15 percent of products that actually go into development become viable.

When your company spends money on R&D, you have no way of knowing which projects will pan out – that is, produce future benefits – and which won’t. Even if you could identify the projects that will work, you can’t put an objective figure on what their benefits will be. And then there’s the unresolved question of how to treat “failed” projects when later successes are built on the lessons learned from those failures.

Like I said, accounting rules for software are outdated. Back in the 1970s companies programmed in what, Fortran? Cobol?

[12x]

There seem to be two alternatives: take a company like zoom and get some sense of what a reasonable valuation could be, or how Zoom is still undervalued to what it could be a few years from now, or just do not include valuation in your strategy whatsoever (I’m being serious, it may not be a bad idea).

A typical Zoom calculation may look like this.

2020: $660 million (100%)
2021: $1188 million (80%)
2022: $1900 million (60%)

I think anyone would agree that these are extremely optimistic forecasts for such high revenues. No company in the SaaS universe has hit anywhere near $2 billion at such a fast rate.

On top of that, let’s assume zoom is at 30x sales. Let’s assume Teams, which is Zoom’s really only threat at this point, just fails and lets this happen. And the market therefore places A premium on Zoom of 30x sales despite slowing growth, because Eric Yusn was right. Zoom is the Tesla to the automaker industry. Just a whole new software driven product that would have to be built completely from scratch. And Cisco just doesn’t get it done nor does anyone else. Zoom outsourced engineering model just lets them stay ahead of the competition. That puts Zoom at $57 billion market cap at the beginning of 2021. Less than a doubling in 2.5 years.

Some may say that is still impressive and they would be right. But I think there are stocks out there that could do even better so I don’t give zoom a dime. I think anyone would agree they forecast I gave above is extremely optimistic and on the top end of best case scenario.

The alternative is to just ignore valuation and say Teams is the only threat. If it does not adapt zoom is on its own for years so who cares about coming up with pointless inaccurate forecasts. There still has to be a methodology to rank risk/return among the stocks out there though.

[Tamhas]

So, what is the issue about capitalizing R&D. There are many capitalized purchases which one makes, but doesn’t know how they are going to contribute long term. One can do something as mundane as buy a new warehouse or store and then have it turn out to not work out. I am fine with having a fairly short life, like 5 years, and even thinking that one of the advanced depreciation schedules is appropriate, I just think that the development work I do today to create new product is not “consumed” in one year.

[anthonyms]

If the conventional company is trading at an enterprise value of let’s say, four times its revenue, isn’t our SaaS company worth four times THAT! Or six times that, … or who knows, ten times that?

This is a great thread and should nudge our board’s comfort zone to think differently about the investment paradigm we are witnessing and our expectations around valuation.

I absolutely do not know what the right valuation multiple is that should be ascribed to our high growth, high margin, high recurring revenue SaaS companies; however…

In terms of the upper limit then surely this has to be constrained by 2 things that I just don’t see discussed often enough on this board:

1) Total Addressable Market
What is the size of the market a company is addressing and what could it address with an expanded solution set? The fact that the TAM ZS or Crowdstrike faces is ~$25bn or something, should at least put an upper limit to any multiple ascribed. Ok if they find ways to address the rest of the cyber security market and it becomes a TAM of let’s say ~$50bn - then again that should give us a clue as to the upper limits of our expectations. FWIW the Ad industry is a $750bn TAM (TTD/Facebook anyone) and eCommerce/mCommerce is a $3 trillion market (and retail is $25 trillion market - Shopify/Square/MercadoLibre/Alibaba anyone).

2) Terminal Growth Rate
We find it reasonable to expect our 40-75% growth rates to slow over time but what kind of terminal growth rate could we settle at? Cyber Security and eCommerce is growing at 20% so that’s a pretty healthy tail wind to converge towards. On the other hand traditional retail, utilities and even semi-conductors are now in single digit industry growth rates. Again - depending on the sector that should give us a clue to the likely outer limit of our multiple expectations.

How we model this I don’t have a formula for but I would think these are reasonable considerations and a place to start to think about it and in particular considerations to bring to the table as we try to compare one investment target vs another. (It might even perhaps in part explain the continued SP performance behind Shopify in the face of declining historic growth rates and the difference of opinion this produces around these parts - but that’s an hypothesis and another discussion).

Ant