Something I’ve been thinking about a lot is the idea of selling stocks. When I was at the Motley Fool, they were just finishing up a study, analyzing every recommendation they made, both buys and sells. What they found was surprising. Had they never issued a sell recommendation, their overall performance would’ve actually been better. That study stuck with me. I don’t think it’s a secret that long-term investing works. But what about this concept of never selling? Does that actually work in the real world or is it a nice study from a newsletter company?
Well, there are a few differences in a real-world portfolio. For one, if you never sell, allocations are created by the performance of stocks. If you have a 100 bagger, maybe that stock ends up at 50% of your portfolio. If you’re fine with that, so be it. That might even be the right choice, but I know a lot of people would be nervous about that. Plus, you open yourself up to a lot of company-specific risk. Who knows what could happen? Maybe the founder gets in an accident? Or maybe a competitor starts acting irrationally, putting pressure on your company? You never know.
Second, I think concentration is important. Having too many stocks makes it impossible to monitor closely and outperform by a wide margin. If you never sell, you will likely keep adding stocks and it may become too much to follow.
Third, there is an opportunity cost. I’ve talked about this concept before but there are so many stocks to choose from globally. If a business you hold isn’t performing well, why not switch it out? You might be falling prey to the endowment effect where you value something more highly just because you own it. It’s a powerful bias. But why not just sell something to buy something where you think there is more upside? As my friend says, “when in doubt, get out.”
These reasons lead me to believe that never selling is probably not an optimal choice when doing concentrated investing. Now, I’m not saying that never selling or holding 50+ stocks is a bad thing. I just think, probabilistically, it becomes much more difficult to outperform by a wide margin. The more stocks you hold, the higher odds of reverting to an index return.
On the other hand, there are serious benefits to never selling. For example, most of my big investing mistakes have come from errors of selling rather than errors of buying. Selling Shopify at $144, DocuSign at $57, Twilio at $98, etc. What I could counter to this is that it’s important to look at what you bought with those proceeds. If we sold Twilio at $98 and bought Zoom at $70 (which is the case), that actually would’ve been a good decision, even though Twilio has been a multi-bagger. But maybe we also have some big laggards in the portfolio that we should’ve sold instead of Twilio? Rather than only focusing on what we bought when we sold, we also need to compare it against the other stocks we continued to hold at that time.
It’s easy to look at things in isolation but each day we implicitly make a lot of decisions. Holding a stock is essentially choosing to buy it at that price. Yes, there are tax considerations but if you have a non-taxable account, this statement is pretty much a fact. Another problem with knowing whether you right or wrong is timing. Tesla went sideways for five years and then has gone up more than 10x in the last year and a half. Were Tesla shareholders wrong for five years? Yes. Are they wrong now? Emphatically no. Timing is so tricky though. If we knew exactly when stocks were going up, we’d all be on a beach right now, instead of thinking about this concept.
One pattern I’ve noticed is that it’s generally a bad idea to sell a stock just because the stock price hasn’t been going anywhere. The problem with selling a great company when nothing is wrong, just because it hasn’t gone up, is that it will almost inevitably start rising once you sell it. I call this the “Murphy’s Rule of Timing.” Adapted from “Murphy’s Law” – anything that could go wrong, will go wrong – “Murphy’s Rule of Timing” is basically a truth 😅.
But I don’t think there’s necessarily a problem with selling a stock because the story has changed. Now, this is a very important point. I used to think the only valid reasons to sell a stock were:
1) your thesis is broken
2) there are better opportunities.
But I think I’m changing my mind on #1. A thesis is fairly egocentric. You’re essentially implying that your thesis is the truth and if the market doesn’t agree with that, it’s wrong and you aren’t. In some cases, you will be right, but there will be times where the stock keeps falling and you cling to your conviction even tighter. You will look at the valuation and think to yourself, “This bad news must surely be priced in BY NOW!” Yet your broad thesis might still be in-tact. Maybe your thesis was that a certain technology will play out and the company is the leader in that (yes, I’m looking at you Fastly!). But the company has a tough quarter and you hold on, while seeing it go down 50%. Were you wrong to hold on? Again, timing comes into play and we have to think about the opportunity cost. Maybe the story has changed enough where now another stock in your portfolio looks like a better deal?
One thing we have to constantly think about is updating our views based on new information. This way our portfolio isn’t static. It’s changing with our changing convictions as we find out more information. This idea of a “changing story” is something that I think is really key. And yet there are always nuances. If you hold 50 stocks, you can afford to give a company whose story has changed much more wiggle-room. However, if you own just 5 stocks, you almost are forced to be ruthless. You can’t really afford to have a drastic underperformer.
Another error I have made is not selling because I’ll have a big tax bill. That was one reason I didn’t reduce Alteryx earlier. Yet it dropped like a rock even though I was pretty sure earnings were going to be terrible. If we have short-term capital gains taxes, those are 30% for most people. That means the new opportunity has to have at least 30% more upside than the current idea for it to make sense. That’s tough to underwrite. For long-term capital gains at 15%, that’s a little more doable but still difficult. Let’s say we hold a stock that just had a terrible earnings report but we’ve held it over a year. It’s now at $100 and we bought at $20. For each share, we’ll take home $68 in profit ((1 – 15%) * $100-20). But let’s say we don’t sell for fear of the tax man and the shares drift down to $95. On the other hand, the stock we wanted to buy, had we sold, went up from $92.5 to $100. Would our decision not to sell be a good idea?
Scenario 1: Sell at $100 and put the proceeds into the new stock
Our after-tax dollars would be $88 (100-(80-68)) + the $7.5 from the new stock = $95.5
Scenario 2: Keep at $95
Scenario 1 is more favorable, but barely. My point was to show the math. If we have a 5-bagger and the stock we hold drops 5%, it only takes 8% upside for the new decision to be a better idea. The math does actually change a little bit if you have a crazy return like a 100 bagger.
100-bagger scenario: Sell at $100 and put the proceeds into the new stock
Our after-tax dollars would be $85 (100-(99-84)) + the $7.5 from the new stock = $92.5
You would actually need something like an 11% return on the new stock to make it worthwhile. The point is that it’s not actually that crazy. Usually we get hung up on taxes because of the idea rather than the math. I think it’s really as simple as this: does another stock have at least 15% upside versus the one I’m deciding to sell (if I have long-term capital gains taxes to pay)? This framing should make it easier to sell a stock with capital gains rather than getting stuck on the tax bill. It’s about optimal decision-making at the end of the day.
Ok, I’ve rambled long enough. I want to summarize a few things.
Selling is tricky. Selling big winners is a mistake but it also matters what we put the money into and we need to compare it to the other stocks we continue to hold. A few other considerations are concentration and taxes. Along the way, we came up with a few rules of thumb:
1) The more stocks in the portfolio, the more wiggle-room you can give
2) Will the stock you want to sell, underperform another stock by at least 15% (30% for short term gains)?
One last thing I want to mention is the idea of selling and then getting back in. I’ve seen very few people do this successfully. Oftentimes, there are other stocks that people would rather check out than one they’ve sold and has now gone up 3x. At the end of the day, it’s all about the fundamentals of the business, that the business is firing on all cylinders (this is what I mean when I say “the story”). It’s not actually the story the media or other investors tell. It’s the business momentum. If that “story” falters and you have other opportunities where there is strong business momentum, it might make sense to sell given you have thought about the two rules of thumb.
Investing is hard. There is always nuance and oftentimes, a tension between two opposing viewpoints. I try to give our companies some wiggle-room but I also think there are times to be ruthless, especially when there are a lot of other interesting ideas. This was the case with the Wednesday trade alert. I still think Fastly is a good business, but Fiverr just seems to be a better opportunity. This might seem like a short-term view, but I also think long-term investors don’t think about this enough. It’s not just action for action's sake. It’s an educated bet on the future prospects of multiple companies. Each company must deserve to be in the portfolio. And today’s competition to get a spot in the portfolio is high. There are a lot of fast-growing, game-changing companies out there. And hopefully, we’ll refine this process based on every decision we make.
Maybe unsurprisingly, a question I get often is “aren’t your stocks overvalued?” So I want to take this week to talk about valuation and where we’re at.
As a caveat, valuation is very tricky. When companies are growing as fast as ours are, a lot of the value is tied up far out into the future. This means their current values are very dependent on the trajectory of growth. Hence, you can see 30% post-earnings sell-offs if growth targets aren’t met. This is one reason I’m not a fan of traditional DCF analysis; the result can change drastically when inputs are barely changed. Move the cost of capital from 10% to 8% and all of a sudden you have a vastly different price than you were previously willing to pay. Don’t get me wrong, the core principle of the DCF remains important – a company is worth the present value of the future cash it can produce. But when our look-through revenue growth is 98%, thinking we can predict anything with a degree of certainty is most likely a fool’s errand.
Ok, with that caveat out of the way, let’s get into the nitty-gritty. Instead of a DCF, I like to use an earnings model. It’s really nothing fancy, but it allows me to see what are the assumptions that are priced into the stock. More often than not, even this is a fool’s errand but the output is less sensitive to the inputs so the range of outcomes isn’t as wide. For instance, instead of a classic 10-year DCF + terminal value, I only use 5 years. I don’t think there is any way to predict beyond that. Even 5 is probably too long but I’ve found it to be a good middle-ground. With a shorter time horizon, revenue growth growing 45% vs. 50% will certainly have a difference but it won’t be magnitudes.
Let’s do an example with arguably our most “expensive” company – DDOG. I put expensive in air-quotes because there is so much nuance. DDOG is the company with the highest forward price-sales ratio but that might not make it the most expensive in the future; and that’s what is important. For instance, if revenue accelerated it would soon be cheaper than another company whose growth fell off a cliff. So in terms of 12-month forward revenue multiples, DDOG has the highest. Just want to clarify between multiples and value, they aren’t always the same.
DDOG, as of writing this was somewhere around $115/share and had 310 million shares outstanding. That means the market cap almost $36 billion on $481 million in TTM sales. Roughly 72x trailing sales. That’s the headline number everyone sees and thinks to themselves, “Oh my word, who on earth is paying that?”
But let’s make a few adjustments. For starters, DDOG is debt free and has about $800 million in cash on the balance sheet. That brings us down to about $35 billion in enterprise value. But we also have to factor in that DDOG will likely grow faster than 60% for next year. So that $481 million quickly turns into $770 million. Ok, now we have a proper EV/forward sales multiple – 45x. Certainly not a low multiple, but also not 72x.
Now we have to do our earnings model. If the future value of a company is the cash it produces, a proxy for cash earnings is simply earnings. We can debate the shortcomings of GAAP earnings but for a debt-free company, with very low amounts of goodwill and one-time charges, and with high free cash flow conversion, cash tracks roughly with earnings.
So we really have 3 main levers: revenue, margins and the multiple.
Revenue growth and a margin profile leads to an earnings estimate and then we need a reasonable multiple to figure out the future value of our company. Now, remember, this is shorthand. We are technically using a shortcut because we aren’t doing the full DCF. But the tradeoff is worth it to me because the range of outcomes is narrower.
What’s great about this approach is that it’s very simple to see where people differ. The difference isn’t broadly “this stock is expensive” vs. “this stock is cheap.” You can actually drill down and see where you differ from someone with your estimates. It really brings things down to reality.
Right now, I think DDOG can grow 50% for the next 5 years. That might be high for people’s estimates but I think it’s possible. This is another difference. Classic value investors will focus solely on the downside which makes all the sense in the world when you’re trying to not lose money. But the math favors people who focus on the upside. Why? Because the most you can lose is 100% but you can make multiples of that. One TTD 10-bagger is equivalent to 10 insolvent companies. And it’s pretty hard to become insolvent with no debt, an important product in a growing end-market, and a founder-led CEO.
So rather than try to think of the worst-case scenario, I try to think about a scenario that I think it reasonable. Not crazy-optimistic but also not too pessimistic. I’m often wrong but I give it a try. One way to check yourself, is to break down the revenue into its components. If DDOG grows 50% for the next 5 years, it will have nearly $3.7 billion in sales in 2025. Another way to look at that is 27,400 customers and $133,000 in revenue per customer.
Is that reasonable? I think the customer count is very reasonable seeing as that other companies like Nutanix already have roughly 20,000 customers and are still growing the numbers by thousands every year. While an incredibly different price point, DocuSign will have more than 800,000 customers in the next few quarters. The average revenue per customer implies a net retention rate of 120%. DDOG has never grown that under 130%.
So, am I full-of-it? Maybe. But these are my assumptions. People can disagree with them but here is my logic. I am always open to being corrected because my ego isn’t important; returns are.
Then we can move onto the next assumption. What’s a reasonable margin? Non-GAAP margins are already 12% and they jumped up from -1% last year. At the same rate, they would reach 30% in two years. But that isn’t very likely. DDOG will continue re-investing back into its business. One trick I like to use is looking at the incremental margins. This is rough exercise, but last quarter, DDOG added 15 million in non-GAAP EBIT on less than $60 million incremental YoY sales. The official count is 26% incremental margins. That gives a quick way to look at what margins could be. I’m giving away some of my secrets this week :)
One quarter isn’t always a reliable measure, but I think it’s in the realm of possibility for DDOG to reach 30% margins. Both Okta and Alteryx have published 30% long-term margin targets and I think DDOG is definitely in the same realm of possibility. In fact, DDOG is more efficient than Okta and Okta is pretty darn efficient!
So now we have an earnings estimate in 5 years: $3.7 billion in revenue at 30% margins is about $1.1 billion in EBIT. Most of these companies will have net operating losses in the future, so we’ll ignore taxes for now. At some point, this will obviously have an impact. But once again, we are going for reasonable, not exact. We don’t want to have false precision because that will give us confidence when we don’t really deserve it. That’s a dangerous combination.
The last step is getting a reasonable multiple. I typically don’t like to give a multiple that is higher than my revenue growth estimate. The simple reason for this is that if margins are expanding and revenue is growing at a certain rate, that means earnings will be growing faster than revenue. There’s this metric called the PEG, price-to-earnings-to-growth, that basically says something is inexpensive if the PE is less than the earnings growth rate. So, very unscientifically, if we use a 50 multiple, that would imply that our PEG would be less than 1, which would be reasonable. Yep, there’s that word again. It’s about being reasonable, not too optimistic or pessimistic. On the other hand, this works for me. If you have a lower volatility tolerance, feel free to use assumptions that make you more comfortable.
Now we have a future company value of $55 billion. If we discount that back to the present, that’s a 9% annualized return. Nothing crazy but not terrible. Now some people may argue that these are very optimistic assumptions just to get a 9% return. I might be inclined to agree but I’m also using a rule of thumb I often stand by. I try to never sell because of valuation. I’ve been burned many times. It doesn’t necessarily make logical sense, but it actually does make mathematical sense. It goes back to the unlimited upside part. If we lose out on a 4-bagger because our return assumption moved down from 15% to 9%, that’s a poor decision. Now this won’t always work but I’d say the probabilities are on your side.
Another consideration is buying the stock at this price. Notice that I haven’t bought any DDOG lately. It has naturally grown into the third largest position. That doesn’t necessarily mean I think it’s a bad buy but there are certainly companies with higher projected returns. But these forward return projections aren’t everything. We still have to look at the competition, the management, the product, the strategy, etc. All of that is the hard part. Valuing it with this back-of-the-napkin methodology doesn’t actually take much time.
There is so much nuance with investing. I’m not saying 45x forward sales is cheap. It’s clearly not. But I’ve gone through my assumptions to get 9%. When it hit $35 in March, I was getting 40% IRRs. As the stock has rocketed back, those forward returns drop down. Yet I try not to sell because of that math. It’s fine to trim and I’ve honestly ben tempted to do so with DDOG but when a company announces an awesome new partnership with Microsoft and is expected to accelerate growth, it’s hard to sell. I surely don’t have a crystal ball and I make a ton of mistakes but I hope this was helpful when thinking about valuation.
When someone asks me about valuation, it’s so nuanced it’s hard to fully explain. But maybe that’s my overconfident side talking and it’s actually simple and I’ve been getting lucky with expanding multiples. I’ll leave that for you to decide :)
Interested in Infuse Asset Management? Let's talk.