Growth as a false signal in Y Combinator startups | Hacker News Reader

Growth as a false signal in Y Combinator startups

techcrunch.com

149 pointshappy-go-lucky44 comments

As a VC, the premise of this article makes little sense to me.

1) Investors aren't idiots when it comes to growth. No one expects today's growth rates to stay constant. Growth decelerates over time.

2) You can't take averages of revenue and growth across lots of startups and use that as your model. Here's an example with two startups:

* Startup 1: $4.99m annual revenue, grew 5% last month.

* Startup 2: $10k annual revenue, grew 145% last month.

"Hey look, if you average these growth rates out, you have a combined $5m in annual revenue growing 75% monthly. That'll be >$4b in annualized revenue in a year!!1!"

No, it won't.

The reason investors value high growth in young companies is that growth rates are an okay proxy for a lot of valuable things: product-market fit, execution ability, go to market strategy, etc.

You just forgot to divide the revenue by 2, but your point stands.

Fixed, thanks! The article switched between "combined revenue" and "revenue per company," and so did I. I should've made that explicit in the first place.

One thing that confuses me about YC is that they're an accelerator rather an incubator. In particular, this article talks about revenue growth which is an acceleration stage metric. Seed stage needs seed money to prove a hypothesis. Acceleration stage needs A money to accelerate. Stressing revenue growth at seed stage seems like it could be dangerous to a startup's health. (Yeah, ideally you would want the seed/acceleration transition to be c2 continuous.)

YC offers seed money for seed percentage, $120k in return for 7%, and attaches accelerator metrics. This transaction just doesn't make sense to me at all. Clearly it makes sense to others, given the number of YC applications.

I think the history of YC explains the confusion. YC was originally a seed-stage incubator (although they disliked that term, as they were very different from the 1990s "incubator" stereotype). As it became more famous, it focused more and more on later-stage companies, and the investment sizes also grew. From 2005 to now, the fraction of very early stage companies (pre-incorporation) has gone from 100% to ~12%, and the investment amount has increased from ~$20K to $120K. However, YC has (AFAIK) never explicitly acknowledged its transition to accelerator, and much of the writing on YC is years old, which creates some misunderstandings.

> As it became more famous, it focused more and more on later-stage companies, and the investment sizes also grew.

Something I've wondered: was this focus change deliberate? Or is it just a result of becoming more famous, attracting many more applicants, and accepting more mature companies since their success to date looks better?

Good question, I don't know. If anyone knows anything about this, I'd be very curious to hear.

Another explanation is they may have realized that the pre-incorporation "super early" startups were a riskier bet, and moved their sweet spot preference just a little bit past that based on the success metrics of their previous classes.

(Recent YC founder)

This looks confusing if you consider YC to be an accelerator or incubator, labels which they hate. If you look at them as a seed stage investor, their shifting focus makes sense.

They give $120k for 7% of the company. Since the deal is set, the only variable, to maximize shareholder return, is the companies they accept. Since they are an investor, they're going to pick the best investments.

How do they know the "best" investment? Past history helps, so (relatively) later stage companies will have an advantage. Revenue, team, traction, etc. are all good indicators.

If you look at YC as just a seed stage VC, their actions are pretty consistent and logical.

If you view them as just a seed stage VC then is 7% for $120,000 reasonable? Seems that if someone thinks $120,000 is a lot of money (cuz 7% is a hefty percentage) they're desperate. If they think the YC Rolodex will make them, they're desperate. I'm not getting the value.

Yes, YC has a brand and they can command this and their actions are indeed pretty consistent and logical. But it doesn't make sense from the other side of the transaction at least to me.

Won't YC get diluted along with the founders? That 7% will end up being less than 5% after a few rounds of VC when the founders' collective share of the company is 40-60%.

7% for $120k is a pre-money valuation of ~$1.6M. The value of a startup is not just its present value but also includes its expected future value. If you believe that your startup is worth less than that before YC, it makes sense to sell as much as possible at that valuation.

On the flip side, that is a post-money (and post-YC) valuation of over $1.7M. I think any reasonable founder would expect their startup to be worth at least $2M after going through YC - for the Rolodex and for the 3 months of heads down focus that it forces you to do.

> But it doesn't make sense from the other side of the transaction at least to me.

It does make sense, because YC is effectively help price a round. What doesn't make sense is that most companies that go through YC are getting convertible notes from non-YC seed stage investors, while YC itself as an equivocal seed stage investor is getting equity. This is where the waters get a bit muddy.

Empirically, selling 7% of your company to join YC improves your average outcome so much that the remaining 93% is worth more than the entire company before to the transaction. If anything, YC is too cheap.

What empirical data is there about this? Selection biases would seem to prevent a simple analysis.

> This transaction just doesn't make sense to me at all.

As a YC applicant I feel I am in a position to offer an explanation.

Money is very hard to raise for many first time founders; 120k is a fortune at this early stage. YC offers access to its rolodex and being accepted (whether the startup succeeds or not) is a badge of validation that opens many doors internationally.

The trade-off is worth it.

Also, the YC badge allows founders to raise a round at Demo Day at a MUCH higher valuation than even the 2nd best accelerator (and this has been shown to be true for almost every single company in each batch) -- the difference is staggering.

Source/data?

I've attended dozens of YC and non-YC demo days as an investor. I can vouch for the grandparent's comment. YC companies have valuations that are 1.5-2.5x higher than elsewhere, but (in my opinion) have outcomes that are 3x+ larger on average. So they are expensive investments on the surface, but still relatively great deals.

As a complete outsider who has never raised money or set foot in the valley my feeling is that without data this sounds like some sort of buddy-buddy system. Where's the data to back this up?

Basically the way I read this is if I have a business model that is likely to require a series A eventually (i.e. not profit oriented) if I won't apply to YC I'm setting myself up for a massively worse deal later?

Well, it could be the case that companies that get into YC are already "good enough" to already have higher valuations in the first place.

I suspect there's a bit of groupthink involved though, creating a feedback loop. (Well, of course there is, VC investment decisions and valuations are based on signaling from other investors!)

So, building your company, maybe not massively worse, but yes, probably at least somewhat worse. It's a tough Valley out there.

1) I have anecdata about returns: my first fund wrapped up earlier this year. About 30% of our investments were in YC companies. Based on investment multiples, 3 of the top 6 companies in the portfolio were YC. This is based on multiples, not absolute valuations.

2) Investors are strongly incentivized to be greedy, so buddy-buddy systems are not in their self-interest. Most funds take a 20% cut of the profits they generate, so they are typically investing in YC companies only if they think those returns will be as good as lower-valued non-YC companies

3) FWIW, the valuations are only substantially higher for YC, and not for other accelerators (many of which are quite good). I think that is also evidence against a buddy-buddy conspiracy.

As to your question about Series As, I don't think YC has a huge effect on Series A valuations, although I haven't looked at the data there. The effect is mostly on seed valuations. I think of the YC badge as similar to a going to Harvard. When it comes to getting your first job (seed round), having a Harvard diploma is a great signal of your potential and might get you a higher starting salary. When it comes to future jobs (Series A and later), people will look more at what you've done since graduating than where you went to school.

This is the kind of thing that isn't public.

https://angel.co/valuations

So maybe the metric should be traction rather than acceleration. To me, personal opinion, if anyone is talking about acceleration at seed stage then my bullshit detector is firing. If someone is measuring acceleration at seed stage, I'm out.

Yes, I'm an engineer. How could you tell?

From the point of view from someone else: the consumer side of things, and the way I start to approach and recommend companies like this. Basically, if the company still seems to be in growth first phase, I avoid recommending or using the product. This is generally because at that point in time, the product still doesn't have a solid financial plan.

B2B offerings are less worrisome, and I imagine with the the SFO culture, easier to sell. After all, if your investors can introduce you to other companies that could use your product, well, it comes across as organic "growth." But at least it has a real business plan. Whether it sticks around, well, again, that really depends on the company and what they are offering.

In the end though, most often I see this working out poorly for the consumers in the long run.

I don't think the author's math is wrong here. Looking at this sentence, he/she is not taking the average for both like you described, but applying growth rate to revenue for each individual company:

If we applied the companies’ monthly growth rates to their reported revenue, then after just one year the 22 companies would be generating about $21 billion in combined monthly revenue, or $963 million monthly revenue per company.

I didn't read in the article that he was taking the other investors for idiots. I think the point of the article was much more directed towards early entrepreneurs. 1) Growth is good but 2) ... don't fake the growth numbers with tricks 3) ... pay attention to other metrics like retention

Agree completely Leo. Growth rates are a snapshot and no replacement for understanding the market, founders, and product.

Further, after having worked with 700 startups at YC, it is obvious to me now that if you can't grow from small numbers, there is no way a company can truly be a good startup. If that's true, then the best advice you can give starting out is to grow. It's just table stakes.

The article is an absurd straw man argument.

Garry, it might be an exaggerated premise, but how is it an absurd straw man? Every demo day (YC or otherwise) has a large number of startups claiming these high growth rates. What percentage work out, even one year out from demo day? Maybe you can say that the high growth rates from small numbers aren't long-term predictors, but isn't that just what the article is saying? That's what I read.

It's obvious that, as you say, growth is "table stakes" for a good startup. But it's also true that competitive metrics are gamed, and that many startups claiming these numbers are, in fact, gaming their metrics.

Does this make a difference? Probably not to good investors. But it does change herd dynamics, and makes it harder for teams that are unwilling to do bad things in the name of short-term growth.

> That'll be >$4b in annualized revenue in a year!!1!"

Sadly, I've seen way too many instances where one month of growth in an early stage means a ridiculously published ARR number to the press. If anything this article is a reflection on TC itself and how they don't properly investigate in their reporting.

This data analysis is cheating. You can't provide an earnest data visualization and obfuscate the data (the corresponding YC companies) at the same time, as it makes it impossible to validate. True, VCs can't disclose the metrics revealed in a Demo Day, which makes the existence of this article problematic.

Additionally:

> If we annualized revenue for each company on the twelfth month after demo day, then annualized revenue per company would vary from $1 million to $159 billion.

What the hell?

Tangentially related to this article, a few growth number tricks I've seen from founders recently:

-Flying all over the country speaking at conferences giving away tons of schwag while claiming zero dollars spent on advertising. "It's all organic growth!"

-Hiring a PR firm to get you mentions on industry blogs while again claiming zero dollars spent on advertising. "We get 1000 new visitors to our website every day and we've never spent any money on advertising."

-Un-launching their product when they don't see the growth numbers they want, "Right now we have about 50 users in a pre-launch focus group." AKA "Our first launch failed, so we are going to pretend it didn't happen." Followed by a second "launch" so that growth numbers look good "since launch".

It's popular to pick fights with YC and its ideas these days. Easy press. Other examples: Gab, CB Insights, various reporters. The recipe goes: Find someone or something famous. Find a bone to pick. Leverage famous something's name to generate headlines. People are going to click.

I don't think that growth is a false signal. One major thing that kills startups is that people don't want what the company is building. The second derivative of revenue is a good way (i.e. signal) to measure product-market fit. The other major cause of death is cofounder breakups, and that's hard to predict.

Regardless, it's tough to predict the future.

> Nick Mayberry, director of content at Tandem Capital, contributed to this article.

This might explain why the article seems to try so hard to bash YC.

Agree, appears there's a pattern of Tandem Capital referencing YC in a less than positive way.

The article itself is written by a general partner at Tandem. It is pretty unusual in the business to have that kind of agenda frankly. Seed investing is the least zero sum stage in finance that I can think of. When founders win, seed investors win. I just can't see how this is a good tactic.

Garry, I wrote the article and by no means have any agenda against YC. Tandem owes much of its success to YC companies, and we don't intend to stop backing them now. The main point of the article (as some other commenters pointed out) is simply that not every company needs to be growing through the roof at demo day to be on track for success. Perhaps some don't think that warranted mentioning; I happen to think the reminder is healthy.

Two points. One, my understanding of Paul's basic premise is that it's not the class quality that matters, it's the unicorns that matter. As in, in this game, you're not chasing good classes. You want the one or two superstars, even if they're in a crap class. Secondly, the analysis I'd find useful is one that looks at unicorns using hindsight. How early can you tell them apart? They are the ones we came here for, remember? It's a very hard task, like telling the gender of day-old chicks.

> Paul's basic premise is that it's not the class quality that matters, it's the unicorns that matter... You want the one or two superstars, even if they're in a crap class....

I used to wonder at this thinking and so apparently have many other observers until I hit upon the realization that the narrative from YC was misleading. It doesn't make sense to rest your entire strategy on seeking one or two superstars from a crap class. What is really going on, is that they organize the most promising applicants into a class then nurture and wait for unicorns to emerge. The midlings make returns too

The real point of this article isn't to be right, it's to generate more eyeballs and provoke discussions. More TC clickbait.

TL;DR

Not every company that has a revenue chart shaped like a hockey stick is going to become a multi billion dollar business. Consider looking at metrics other than growth when investing.

Is this actually something that needs to be said?

Am I reading the second chart correctly? $44m in MRR across the 22 companies, ie an average of $2m MRR after demo day?

That sounds like Series B metrics? Why did they go through YC?

Deleted

Reply on news.ycombinator.com