SAFEs are not bad for entrepreneurs
blog.ycombinator.com
blog.ycombinator.com
It would be useful for the founder community (especially outside YC network) to have examples of how different recent startups have done it - offered discount or cap or both, how they determined the cap, the experience at A, experience with SAFE when dealing with angels/micro VCs, etc.
Any founder willing to share that here?
Ultimately the terms are going to depend on investor consensus, but my strong preference is to go uncapped w/ discount.
Having a cap is an incentive for me to "grow until I'm 6 months away from X, then focus on raising instead". For some seed investors an early valuation & equity conversion may be desirable, but I'd rather "grow until I'm 6 months away from needing funds to (grow faster|survive)".
Having no cap, on the other hand, encourages me to stay lean, grow quickly & ask for money only when it has the greatest value. That's behaviour which may not be friendly to opportunistic short-term VCs looking to get a quick valuation bump, but which is strongly correlated with long-term success and eventual home-runs.
I want to give a discount because seed investors are taking a big risk on me, they deserve it.
I am also curious how the reactions differ from Angels to MicroVCs to Seed Funds.
If I secure enough commitment on MFN agreements alone, I plan to default them all into a fairly generous discount.
This was my impression as well. I do understand that the deal has to be equitable for both sides. It also has to align interests and as your example shows, the cap doesn't.
The discount obviously makes sense since the seed funder is taking more risk.
Uncapped notes are generally a bad idea for everyone involved because they misalign incentives. If you raise a seed at a $6m cap, both your goal and your investors' goal is to help you make as much progress as possible for a Series A. For instance, investors will do whatever they can to help you get to a $30m valuation instead of a $20m valuation. An uncapped note means that investors invest at your next round's price. That means they benefit most of the price of your next round is low. I.e. they'll do as little as possible so that your Series A is at a $20m pre instead of a $30m pre. The misalignment creates perverse incentives. For example, if you ask an investor on an uncapped note to make a key customer intro, if they say yes then their reward if you land the customer is that they'll have even less ownership at the Series A. That's not a good incentive structure :)
An analogy: Amazon is $1k/share today. Let's say you think there's a 30% chance it will 3x in the next year. Would you like to invest at $2.5k/share today? The answer is No because that has negative expected value, and you'd rather just wait a year and invest at $3k/share than investing at a small discount to $3k/share today when the stock is worth much less than that.
Also, the time horizon doesn't change the misaligned incentives here. Because I get in at next round's price, I would prefer for that price to be lower.
Would it be more appealing if you the round was closed in the next few months, so you investing early would be well rewarded with the discount?
I agree that 20% discount by itself is not appealing enough incentive for the risk of early stage investment. But what about 40%? ...50%? A discount is just the mechanism that most accurately captures the variable nature of valuation. Finding the right number is where the modeling happens.
As was mentioned elsewhere, smaller discounts tend to apply to shorter horizons, such as when bridging immediately into a round. In those cases, they can be a good way for VCs to get on the cap table with prorata rights without having to muck around in series-A drama.
For a lot of people, I see the difference being the start of the process, and the end of process which can be a few months - in which case, an uncapped with discount seems entirely reasonable.
It seems like you are referring to a case where a substantial time has passed to materially affect the valuation in a substantive way?
The real problem with capped notes is that they serve as a proxy for price. If you're going to price it, just price it. But if a note had a floor + cap that served as a min/max range for what everyone felt comfortable with, and then utilize a discount to do the actual work of allowing for uncertainty in the valuation, that would be the setup that best encapsulates the spirit of convertible notes (of course, the devil is always in the details).
The real problem with capped notes isn't the proxy for price, it's that the cap table is impossible to discern. Capped notes actually do have an advantage in that you can do a rolling close of your round, so you don't have to have everyone invest at the same time on the same day. The downside is that it is hard to explain to your employees exactly how much they own since it depends on how the notes convert.
It's not that simple because you're not factoring for time. Yes, both of those stakes are worth $1m, but the risk levels taken by an investor are different. For example, would you rather own 5% of Amazon when it was worth $20m, or 0.01% when it's worth $10b? Those are both $1m stakes if you ignore time, but the latter would've been worth much less than the former.
Here's a concrete example using https://angelcalc.com/model
Priced round
I invest $1m at a $10m post, owning 10%.
- If your next round is $5m at $20m pre, I'll own 8% (post-dilution) which will be worth $2m.
- If your next round is $10m at $40m pre, I'll own 8% which will be worth $4m.
If your valuation doubles, so does the value of my investment, and I keep an 8% take in either case.
Uncapped note
I invest $1m on an uncapped note, 20% discount.
- If your next round is $5m at a $20m pre, I own $1.25m (5%).
- If your next round is $10m at a $40m pre, I own $1.25m (2.5%).
Note that 1) my upside is capped to $1.25m (that's the 20% discount I got) and 2) the better your next round, the lower my ownership.
Closing thoughts
The reason the uncapped note is even more unattractive is that most companies have some perceived ideal outcome. Let's say it's a max $2b exit for this specific company. For a priced round, I own 8% of a company that could go up to $2b in value. For an uncapped note, I own 5% if you raise at a $20m pre, or 2.5% if you raise at a $40m pre. Note that the better you do, the lower my percentage, and hence the lower my potential upside. If you raise at a $20m pre my stake might be worth $2b x 5% = $100m someday. If you raise at $40m pre my stake might be worth $2b x 2.5% = $50m someday. If you really knock it out of the ball park and raise $20m on an $80m pre for your Series A, my max upside is $25m. Again, this makes no sense because I invested $1m at the exact same time in both the priced round and the uncapped note round, but in the priced round instance I own 8%, while in the uncapped round I probably own much less, and the better you do on your way to a $2b exit, the worse I do. That's a strong misalignment and there's no way for me to make that pie bigger.
If SAFE is a way to start what would be a priced round (a few months), then your numbers would tell a different story. You getting to invest $1M and $10M cap, and then the company raising at $20M valuation later, means that you got an unfairly low priced deal (assuming company's valuation wouldn't naturally double in those couple of months).
In short timeframes (SAFE -> Priced rounds), its hard to imagine something other than discounts being fairer to all parties concerned.
In longer timeframes where the next priced round is a Series A, an year later, your point stands and is well taken.
I agree 20% discount is not an accurate assessment of the risk and potential increase in value from seed to A, which is usually 3x. Also, the original purpose of a note is to solve the problem of inaccurately pricing an early stage company. If that's not a problem then there's no point in not pricing. But if there's great variation in the potential of a company, then it makes sense to use a variable pricing model, which a significant discount (30-60%) better accounts for. Add in a cap and a floor (something like $4m - $40m) to protect founder & investor in extreme situations like you described.
Obviously a fixed price is better for you, the investor. But is it better for the company? Is it the best way to model the uncertainty of valuation? You argue for not dissentivizing the investor. What about the founder who raises $1m seed at $5m valuation, then raises Series A at $30m? They mispriced their seed round, and the likelihood of the A investors finding a way to diminish the value of the seed investors greatly increases.
Just an alternative view of things from an outside.
- are there any other areas where, as a buyer/investor, you buy at a discount to a future price instead of an estimated current price? For example in places where houses appreciate quickly, people still buy houses at a fixed price. No seller ever says "this house might be worth $2m-5m in 10 years, so instead of buying it for $1m today, which don't you buy it for a 20% discount to when you sell it 10 years from now?" Same thing with paintings, stocks, etc.
- the $5m -> $30m mark-up is not a mispricing. For public stocks, pricing is based is based on expected cash flows. For example, if a company is expected to make $10m/year for 30 years, it might be worth $300m today, minus an adjustment for inflation (so maybe it's only worth $200m today). For startups, the valuation is based on "% chance of a huge outcome." So when a company goes from $5m to $30m in valuation, that doesn't mean its revenues jumped 6x. What it really means is investors think the company made enough progress so that instead of a 1% chance at a $1b exit, there's now a 6% chance at a $1b exit. In that regard, the company is worth $30m today, but it was also not worth that at the seed round.
2) Complex pricing is hard to implement well (both for valuations and for product prices). It's friction for what could be a straightforward transaction.
3) It's not clear to me why the investor would want their upside capped. E.g. if I buy stock in your company and the company becomes more 3x more valuable, then why would I be okay with my stock being 2x more valuable? Why not 3x just like you? (And likewise, the founder wouldn't want my stock to be 4x more valuable if the company is 3x more valuable).
Investors will do uncapped notes for bridges because there is usually a very short time period between the note and the next round, so they actually make a better return due to the discount than they would in equity appreciation over that short time.
Serious question - how is this still the case, or make any sense? Wouldn't it be in the investor's interest that the company doesn't spend $60K out of their raise on this, and instead on hires, product, etc?
And given how standard a process this must be for every VC firm, I imagine they would have a well-negotiated rate, which for them is an incremental cost of investing?
I'd like to believe that there are firms out there that don't do this, and that this is turns out to be some sort of advantage for them (a form of founder-friendly/company-friendly, if you will).
But I'd also like to know how standard/non-standard this is.. and why is it even the case in the first place?
Given how Fred Wilson (and others) talk about not liking SAFE, they really should then make it easier to do priced rounds without putting the burden on founders.
1) Because VCs were working on 'fixed income' 2) because they could. The standard deal historically is "2 and 20", i.e. the VCs get paid 2% of the fund raised per year, and the first 20% of the profit. That 2% is used to pay for salary and office space, and doesn't increase until you raise a new fund, and the 20% isn't realized until the fund ends in 10 years. So pushing the cost onto the startup saves them a decent amount of money when you're doing dozens of deals a year.
Leaving aside bloat in lawyer rates because the VC doesn't care about price, the total cost to do the deal is basically unchanged though.
No, it would not be in the investors best interest to make that change. Under the current system all of the money spent is used to buy shares in the company. Under a system where the VC firm was responsible for their own legal fees then $25k (or whatever) would go directly to the lawyers instead of buying shares so they would end up with a bit less ownership.
EDIT: To be clear I'm not saying that I am in favor of the status quo. I am not. I'm just saying that it is understandable given the incentives of VC firms.
Say that a company needs $X to have "enough runway." The VC firm can either invest $X and pay for the firms legal fees themselves or invest $X+legal_cost and have the company pay the fees. In the latter scenario the VC firm will probably end up with a greater ownership percentage.
It is worth noting, however, that it might not really make any difference. Much like tax incidence who directly pays for the lawyer really might not matter much in terms of where things end up in the end. The cost will always, to some degree, be shared by the firm and the company.
Even then, it is interesting that the optics of this doesn't bother investors.
But I suppose it's really kind of a matter of accounting. As long as the outlay of the investors is equal in both scenarios, the post-money situation should be approximately the same.
That said, in reality, the VCs probably have a better understanding of the real costs of the fundraise, so I can see the argument that they'd be best served in the long-run by minimizing surprises for their portfolio companies.
There are 3 parties, the startup, the VCs and the LPs. The LPs are generally pension funds, college endowments, and sovereign wealth funds, i.e. institutions who can write $10M-$100M checks. The VCs are their agent, like a real estate agent helping you buying a house.
While VCs put their own money into the fund, generally the vast majority of the fund money comes from the LPs. The VCs get paid 2 and 20, 2% of the fund per year for things like salary and rent, and the first 20% of profits from the fund when it ends, in 10 years. Because of this structure, they don't get to treat the fund money as their own personal piggybank.
No matter who pays, the cost of a deal is going to be <money startup receives> + <startup legal costs> + <VC legal costs>.
Raising the cost for the VCs, without changing 2 and 20 just results in increased costs for the VC. Paying their own legal costs means they'll do fewer deals, with larger check sizes.
if it paid its own legal fees it would get that much less equity.
imagine I said "I'll invest 1M at a 2M valuation but you have to burn 800K of it in a bonfire." vs if I said "I'll invest 200K at 2M valuation but I'll celebrate by burning 800K in a bonfire."
The same thing happens to the money (it burns in a bonfire / goes to the lawyers) but in the first one I have 50% of your company and in the second one I have 10% of it.
Because in the first one I added it to the equity investment before making you pay it.
The best you can do is limit the legal fees in the TS.
It'd be in a VC's best interest to invest $60k more for the same equity, and make the startup pay that expense, vs. paying for it directly. It's a win-win for startups and VCs. LPs get a bit screwed, however.
> The safe (simple agreement for future equity) is intended to replace convertible notes in most cases, and we think it addresses many of the problems with convertible notes while preserving their flexibility.
(I almost said improper noun but apparently that's like land seahorse.)
For what it's worth, I think this causes more confusion than it's worth.
Safes are great. They're easy to understand, and any semi quantitative founders should be able to build dilution spreadsheets without the help of a lawyer.
I've never encountered any compelling rationale as to why/how that standard emerged.
Market terms for a seed round are either 1) a priced round, or 2) a SAFE or note with a cap and a discount. 20% is a typical discount. A typical cap, in Silicon Valley, would be $4m-$5m on the lower end (idea or prototype phase, little or no market validation) up to $8m-$10m on the higher end (product is live in the market, has decent usage, and perhaps $25k-$75k MRR if you're b2b). There are also outliers that are >$10m cap -- usually those companies have exceptionally strong growth or founders.
In practice, SAFE/note discounts don't come into play that often but caps do. That is, next rounds are usually significantly higher than the cap (and the discount doesn't apply).
I paid closing costs when I bought, and as my realtor said, it's like free cable. It's not a massive savings but it's free cable every month.
The rationale, from the VC's perspective for the company paying for a priced transaction is the same. Really this is a way of making the LPs of the fund pay for the transaction (i.e. their money given to the company to pay the transaction) rather than the VC having to pay transaction expenses out of their fees.
I have met many companies where the first few employees think they own 1% of the company, and after a Series A where 25% is sold they find out they only own 0.5% because the SAFE conversions took up another 25%.
In some cases founders don't understand what is happening or how to include SAFEs in their cap table, in other cases they are purposefully obscuring the cap table. Whatever the reason it's very bad.
You are right, though, employees are better off when they understand conversion mechanics. And when they don't work for dillholes that mislead them with complexity.
To the extent that founders suffer too much dilution while raising via a SAFE, it's more likely the case that there was something not working with the business.
Sure, if the founders could have raised a priced equity round from the get-go, they probably should have done that over a SAFE, but more likely the legal expenses would have been too onerous for that to have been an option...
I won't make one anymore.
I've done two deals that involved a SAFE, and it's been almost 2 years and the companies are still looking to raise a round. If they do, I'm looking at a 10-20% return.
It's not worth it for the risk.
Indeed, at least convertible notes are debt and can be seen you're in a liquidation. A SAFE doesn't even give you that.
Also, if there is a liquidation event, then the SAFE should convert, right? And not sure if you act as an angel or represent a fund, would you rather use convertible notes or price?
I actually tried to get some lawyers here in Australia to convert over the YC SAFE agreements to Australian law and I could not find one. Apparently the big blocking point is none of the law firms wanted to take responsibility for the legal liability. This is one area where our "innovation" government could get involved to sort out.
The short answer: lawyer time.
When you do a priced round, you have to go in and change many aspects of a company's legal structure. For example, the documents drafted in a priced round may include amendments to the charter, a voting agreement, a stock purchase agreement, a right of first refusal agreement, a shareholder agreement. Drafting the legal docs for a convertible note or SAFE is less intensive and time consuming.
In general, the problem is that most benefits that investors enjoy are properties of their shares rather than the money that they invested. For an equity round this is one and the same. Not so much for convertible notes. A simple example:
An entrepreneur raised a $1M convertible note with a $5M cap. Ignore discount, interest and other factors for now. She then raises a $5M round at a valuation of $20M. That yields a dilution of 20% for the round plus a "hidden" dilution of ~17% for the note conversion (1/6). That's the blurry issue that both authors discuss. But if anything the share rights are even blurrier. Let's say that the equity round came with what is commonly referred to as a 1x liquidation preference (non-participating). So they would get $5M back before other shareholders get anything. Even though I just worded that as matching the money that they put in, it is generally a property of the share class that the investors hold. For example, their $5M might have bought 5M shares at $1/share that each says "redeemable for $1 or convertible to common shares". Our note investors also hold those shares now. But instead holding one per dollar, they now hold four per dollar (since they pay 1/4 the price for such a share). Suddenly, they have effectively a 4x liquidation preference benefit and the company has to return a full $9M before common shareholders/founders see a penny of payout (despite only having $6M in the bank).
Interest rates, pre-round ESOP increases, and many other factors in convertible notes make this problem worse. And it affects just about all aspects of the cap table including voting rights, protective provisions, redemption rights, etc.. Basically, the bigger the gap between the cap and the eventual round, the bigger the privilege the note investors pick up. Not just in economic benefit where you would expect it, but also in power/insurance/protections/etc. where it isn't obvious at all. Nowhere in your term sheet for the note or equity round will it mention 4x liquidation preference. Doing so would cause instant rejection of the deal by even the most inexperienced founder! But that's exactly would is going to happen once all the conversion mechanics are executed. And that can catch even seasoned entrepreneurs off guard (and seasoned investors, including plenty of note holders who never understood that they would get these benefits).
Convertible notes - SAFE or otherwise - have a role to play in venture financing. But they are complex instruments and should be use carefully. Anything else is just a recipe for pain in the long run.
The simplest one is that there are multiple sub classes of preferred stock ("shadow series") - eg for a Series A, there are Series A-1, Series A-2 shares that represent each cap. These classes each have their own liquidation preferences matched to the dollars put in originally. The YC-standard safe also contemplates this by referring to "safe preferred shares".
Another option is that the "extra" shares that the converting safeholders receive as a result of the difference between the conversion price and round price are given as common shares (which have no liquidation preference).
Really, no one has said it, but one of the few remaining competitive advantages of Silicon Valley is the legal talent available here. While I read Venture Hacks and Brad Feld as much as anyone, I'll get a good lawyer when/before I get to A.
Similarly, LLC's can be a simple stepping stone to Delaware C. Also, provisional patents are a stepping stone to a full application. Along with Safes, these allow people to move forward without the complexity and cost of completeness.
few->many
Of course the other concern is that you actually have to be cognisant of this issue - or have a lawyer who is - to catch it. Maybe the SAFE could just mandate the pref+common conversion?
So, in your example, for each dollar invested, seed investors would have 1 preferred share with 1x liquidation preference and a price of $1, and 3 shares of common stock.
A SAFE _is_ a convertible note, with standardized language.
A convertible note is structured as debt - there is interest earned until it converts and a maturity date. There are also terms about repaying the debt.
A safe is not debt - it does not have a maturity date and interest does not accrue.
So referring to a "Safe Note" is non-sensical because there is no such thing.
> - Carolynn Levy (inventor of the SAFE)
edit ... You could argue it's no longer a convertible note with those provisions taken out, but I'm not sure how much mileage you'd get with that argument.
You may be right, but what makes you think that?
The common law and some statutory definitions require only that promissory note be, upon breach, reducible to what are called "liquid damages" i.e. "a sum certain in money". This does not mean monetary debt or future payment of money with or without interest. It just means reducible to a certain amount.
The purpose of having a promise with "a sum certain in money" is on the enforcement end. The monetary compensation upon breach (damages) of a promissory note can be unambiguously calculated by a registrar (as opposed to assessment open to interpretation by a judge). This expedites processes such as obtaining summary or default judgment.
However, the promise needn't be in money itself, as long as there is an unambiguous objective calculation of the amount of damages.
I'm not sure how it would be calculable in the case of a note without debt or interest; I'd need to read the note language. If there was nothing specific, one could argue it is portion of the pre-money to the the valuation cap, but that'd be a very loose guess.
http://abovethelaw.com/2015/01/innovation-more-than-another-...