It seems churlish to keep pointing this out, but (assuming the investment is consistent with the typical Silicon Valley practice), it is just not true that this investment, or any participant, has now valued the company at $3.5 billion.
Stripe has sold the investors a new round of senior preferred shares (again, making assumptions). These shares carry various bells and whistles whose terms we don't know -- but probably include some preference at an unknown rate.
The seniority and the bells and whistles make these shares particularly valuable, compared with less-senior preferred shares, not to mention common shares, not to mention something like an authorized stock option that will be issued in-the-money at some future time.
The way you would get $3.5 billion (again, making the same assumptions) is by imputing this per-share value to all of the equity units, as well as things like authorized-but-not-yet-issued stock options.
That is not sensible. The investors in this round have not acted in a way that suggests they believe the less-privileged shares are worth as much as these new shares, and neither has anybody else. There is no basis for imputing the same value to all equity units (and authorized options, irrespective of strike price).
Here is a hypothetical cap table that is consistent with public reports: http://qr.ae/lc0ry
Some of your assumptions are off (e.g. our liquidation preferences are 1x non-participating pari passu; there are no seniority games), but I really enjoyed your Bayesian vs frequentist statistics lecture at 6.945 back in the day!
Thanks, Patrick! (And thanks for posting Mosh on here and giving us a kick in the butt to really release it...) If Stripe just has one class of preferred stock, you're right that my assumptions are off, and I guess then it would be reasonable to impute the value of the newly-issued preferred shares to all the preferred shares. I would still disagree with imputing it to the common, especially since you presumably have your own appraisal of the common at a different price for tax purposes.
Do you think there's a practical path to getting some more realistic numbers in these announcements? Does it matter? It seems like the people possibly being misled by these "TechCrunch valuations" are the press (which is enthusiastic anyway and loves to report the biggest number that can be attributed to somebody) and prospective/current employees (which could be a lot more serious, since people use these valuations to evaluate their own equity packages).
I think this is mostly a pedantic point for a late stage private convertible preferred.
While private convertible preferreds typically have a dividend stream, conversion option (typically at spot) and liquidation preference (1x and presumably no participation feature in this case), they are theoretically more valuable than common. However, since there is no public market for the common, hedging that option wouldn't be efficient (if feasible at all) so I don't think the valuation really reflects that additional value (if at all). Investors develop a valuation view based on the metrics they would use for common (growth, fundamentals, perhaps some real option value, etc.) and invest based on that. All the preferred features are in the doc to provide some protection in the downside, which I don't think is much of a consideration in case of Stripe. You can see this pricing dynamic in S-1 filings of a few precedents (Twitter comes to mind) where founders were able to tender some of their common in the later rounds at the same valuation as preferred[1].
On the other hand, you are spot on when it comes to a public security (i.e. post-IPO convertible preferred, mostly capital/ratings instrument and quite uncommon in tech) - you would definitely bake-in the option value and dividend stream into the security valuation which would result in some conversion premium for the company and say 5-10 points in theoretical value above the par for investors. The difference is that primary buyers of the public convertible preferred security would be hedge funds who can short the common and effectively monetize the option value embedded in the security.
If anything, the way I would look at this is that any credible bid for Stripe would probably have to be in 4-5x of that valuation.
[1] See pgs. 139 (bottom, 2011 Third-Party Tender Offer) and II-3 (top) of Twitter S-1 (http://1.usa.gov/1cEqy0J) for difference of ~1%, likely due to fees, etc.
Technically, things are only worth what people will pay for them. Accordingly, you could also argue that the value of the company cannot be known unless and until it is wholly acquired. But these kinds of arguments are somewhat pedantic. They have built a tremendous amount of value. Whether it's $1 billion or $3.5 billion, it's more value than virtually everyone upvoting your comment has built.
Exactly! I would actually argue that if anyone wanted to buy the whole company tomorrow, they would probably have to pay at least 4-5x premium to today's valuation (purely illustrative/gut figure, I have no inside knowledge here aside from what's publicly known).
Stripe is simply one of the SV's greatest hits and has a ton of upside and real option value. Any investor who got in today understands that and would have no good reason to sell for anything less than a really nice premium to where they got in.
Simply put, in the current market their IPO would literally fly off the shelf, they would get great institutional public investors and have a ton of good growth financing options available to them down the road (follow-ons, converts, heck even straight debt soon). Every investor in the deal understands that very well. Having said that, the investor roster doesn't look like a pre-IPO round (I suspect they would have likes of Templeton otherwise), so they will probably grow the company quite a bit more before they hit the public markets (needless to say - a smart move).
At any rate, I completely agree that pointing out that the preferred is more valuable than common is red herring. Theoretically yes - but in this case and in the case of every great late-stage private company, I would argue that the price of the common >= price of the latest preferred round.
I agree with all of what you've said. It's also worth noting that this is just 2% of the company. One person offering to pay cash for 2% of a company with seniority preferences is very different than someone offering cash for the entire thing.
After reading through the quora article, I agree that different tranches may have different terms. However, if investors are buying 70M shares at $1 each, it implies that they believe a share of the company is worth that much. If the company has 3.5B shares, that's an implied valuation of $3.5B. In that sense, the investment does value the company at $3.5B. Whether that valuation (or a higher or lower one) is realized through a liquidity event is a different matter, and if/ when that happens, the valuation would be specific to that point in time.
> However, if investors are buying 70M shares at $1 each, it implies that they believe a share of the company is worth that much.
Well, you need to unpack what "worth" means in that above sentence. If you're using some kind of probability distribution over expected future outcomes (say, using some present value future discounted cashflow model with a probability distribution of possible future cashflows), you can see that there are all these possible future worlds that Stripe can inhabit. In some of these, Stripe is making lots of money, and so the stock is "worth" a lot. In some of these, Stripe is crashing and burning. In those worlds, common stock is worthless. But the preferred stock isn't as badly off due to the 1x (or higher) liquidation preference, and the other terms. So the expected value of the preferred stock across all these worlds is higher than that of common, and imputing its value over common doesn't quite make sense.
Agreed, that's why I've always thought of it as an implied valuation. Investors are going into it expecting upside, and the terms allow them to have a floor (regardless of what happens with common/ prior rounds), but I don't think they would make the investment if they felt an acquisition/ IPO/ other liquidity event would value each share of the company at less than what they paid for it. I don't think its unreasonable to use this to get to an implied total valuation.
Stripe has sold the investors a new round of senior preferred shares (again, making assumptions). These shares carry various bells and whistles whose terms we don't know -- but probably include some preference at an unknown rate.
The seniority and the bells and whistles make these shares particularly valuable, compared with less-senior preferred shares, not to mention common shares, not to mention something like an authorized stock option that will be issued in-the-money at some future time.
The way you would get $3.5 billion (again, making the same assumptions) is by imputing this per-share value to all of the equity units, as well as things like authorized-but-not-yet-issued stock options.
That is not sensible. The investors in this round have not acted in a way that suggests they believe the less-privileged shares are worth as much as these new shares, and neither has anybody else. There is no basis for imputing the same value to all equity units (and authorized options, irrespective of strike price).
Here is a hypothetical cap table that is consistent with public reports: http://qr.ae/lc0ry
Previously on HN: https://news.ycombinator.com/item?id=5798905