Hacker News new | past | comments | ask | show | jobs | submit login
Y Combinator’s latest batch of startups is too big for one Demo Day stage (techcrunch.com)
156 points by laurex 30 days ago | hide | past | web | favorite | 81 comments

Hi, I work at YC. I think this article is a bit misleading, so I wanted to take the opportunity to give you some of the data. I'm happy to answer any questions on this topic.

1) YC's acceptance rate for applications has remained basically constant since 2015 (< 2%). This batch is not bigger because we accepted a higher percentage, but because we got way more applications. The primary reason we got more applications is that over 15K companies participated in our startupschool.org program and many of them applied to YC. Our general policy at YC is to grow the batch size proportional with the number of good applications we get, and this larger batch is just a natural consequence of that policy.

2) We have been wanting to move demo day to a new larger venue for a long time now. The reason is not that there are too many startups, but that there are too many investors who want to come. We have enough demand from investors to fill the previous demo day auditorium several times over, and this new space means we can allow many more investors to come, which is better for the companies.

3) As the batch at YC has grown, we've grown the number of partners proportionally and plan to continue to do that. The ratio of companies to partners is actually lower now than it was a few years ago, and we've increased the amount of time we spend per company.

4) I think the ultimate measure founders should care about - and certainly the one we do - is the success rate of YC companies on average. If we grew the YC batch and the success rate went down, that would be bad. So this is something we track extremely carefully to make sure that does not happen. And so far, it hasn't. Based on the last 1-2 years of companies, the success rate of YC companies is the highest it's ever been.

Also for general background, I'd really recommend reading a great short statement Paul Graham wrote about why we think growing YC is better for both us and the founders: https://www.ycombinator.com/atyc/#size

Hi Jared,

Thank you for taking the time to write this. I appreciate the info, and while this is not meant as a criticism I am curious how you measure for success.

> 4) I think the ultimate measure founders should care about - and certainly the one we do - is the success rate of YC companies on average. If we grew the YC batch and the success rate went down, that would be bad. So this is something we track extremely carefully to make sure that does not happen. And so far, it hasn't. Based on the last 1-2 years of companies, the success rate of YC companies is the highest it's ever been.

If fundraising at a higher valuation is one of the components used to determine success, I wonder how much the frothier market could contribute to being misleading here? Obviously a smaller cohort wouldn't necessarily solve this and I realize we can't really control for macroeconomic factors, but I could see it skewing the data (depending on how you measure it). Of course, if success is measured simply by liquid ROI my point is moot.

You guys have all the data and know way more than I do, but it seems YC has moved away from funding scrappier startups that might survive an economic setback whereas the much more growth-at-any-cost companies may not survive a hiccup in funding quite so easily.

Again, this is just a thought and you have all the data, so I don't pretend to be as informed as you.

I do think that's possible, and that is one of the confounding factors we spend a lot of time worrying about and trying to control for when we do the data science on this. One of the ways we do this is by licensing data about all the startups in the world and then normalizing our own data against the market overall.

Getting a true apples-to-apples comparison is difficult because the world doesn't stay the same. The funding market has changed dramatically since YC started, and the companies we're funding have changed a lot too - they're much more diverse now. We've tried to control for this but you are right that when building a complicated model, you can never be 100% certain and need to stay humble about your ability to correctly control for all these factors. We won't know for sure for many years; for now all I can say is that we've looked at this from many angles and the early indicators look good.

>>> licensing data about all the startups in the world ...

Can you share where you get such data? It sounds incredibly interesting on one hand and academically useful on the other?

Probably some combination of Pitchbook, Crunchbase, and other similar aggregators.

You talk about the “success rate” of YC companies? What is success in this context?

That's a great question. We don't look at a single success rate; we look at a wide variety of success rate metrics.

It's important to point out (thanks @ferrelty) that there are many forms of success itself. A company that takes an early acquisition for $2M, or a company that becomes a profitable life-style business, won't make YC much money. But it may be life changing for the founders and a big success for them. We like to consider those companies successes, and tons of YC founders choose paths like that and we always support them.

There are other kinds of success too - YC funds nonprofits, and when those are successful, we don't get any return at all but we think it's very valuable and important.

Still, from the standpoint of YC's business model, we need at least some companies to get big in order to make our organization sustainable. In this case, the "success rate" metrics I was referring to are more that narrow financial kind - the question of whether the same percentage of companies are on track to become big successes.

We measure that by comparing each batch to prior batches, adjusted for time. So for example, if we're looking at the W2016 batch now, we want to compare it to the W2011 batch as it looked in February 2014. We compare batches in lots of ways: fundraising and valuation, revenue, profitability, growth rate, exits.

> Based on the last 1-2 years of companies, the success rate of YC companies is the highest it's ever been.

But what constitutes a “successful” outcome? I assume it is not merely a monetary measure, because as we know, just because something is financially successful doesn’t mean it’s good for us or for the planet. These days it’s even debatable if Facebook should be considered a success (for anyone except Mark Zuckerberg).

So what is the measure of success? And can you really make an estimation of that 1-2 years after founding, as you have done?

Facebook is one of the most successful companies in the last 10 years by any possible metric. The fact that you like it or not doesn't change whether it's successful or not.

It's pretty successful at decreasing the happiness of people who use it. And pretty successful as a targeted propaganda delivery network. Also pretty successful at pushing people towards briefer, shittier, and less-nuanced discussions of things that matter, though not as successful at that as Twitter. And pretty successful at wedging ads into every possible corner of the portion of the public discussion commons that it has taken over.

All of these things have made Facebook a lot of money but there is definitely an argument to be made that Facebook has decreased the Gross National Happiness[1] of both the US and the world in general.

1: https://en.wikipedia.org/wiki/Gross_National_Happiness

I'm sure the happiness of people decreased over the last 10 years, but I believe that it's because of the bigger division between rich and poor people.

Personally I'm not a fan of Facebook, as I saw my home city (Budapest) becoming from a nice livable city to a tourist destination / loud party / get drunk fast city between 2008 and 2010 (and it hasn't gone back), and I believe Facebook took part in that (of course later AirBnB helped, but AirBnB was much smaller in that point).

As for other effects, like making people less connected in real life, and more through computers, it has good and bad side effects, I'm torn whether things got better or worse. (Also I don't post too often to Facebook, so I'm not a good person to judge it).

For what it’s worth I visited Budapest after your indicated timeframe and we still thought that it would be a very nice place to live.

I agree with everything else, but have people shown causation for

>It's pretty successful at decreasing the happiness of people who use it.

this? Because unhappiness and facebook use is obviously going to correlate for reasons that have nothing to do with facebook being bad, and I think Facebook is a valuable enough tool for maintaining relationships that I'm skeptical it would outright decrease happiness (on the individual level. Society as a whole I find much more plausible).

I know several very early employees (first 20) who, other than the financial windfall they received, regret that the company became what it is now.

Other than that, Mrs. Lincoln.

>by any possible metric

So..you imagined a few metrics and wrote "any possible metric"? Ok, so you were exaggerating..

I will try to imagine some metrics FB may not be the most successful in: companies that did the least evil, percentage of users that unreservedly loved the product, percentage of users that felt exploited, percentage of users that felt the product was better than 10 years previously, percentage of users that trusted the product, etc.

I have no opinion or interest in the size of YC batches. But just wanted to point out that this:

>The primary reason we got more applications is that over 15K companies participated in our startupschool.org program

Comes across as insincere and hurts the credibility of the rest of your response. It’s just not plausible that a brief online course made a major difference in the number of YC worthy companies out there. Just go with the real reason. (Eg. We have more money and partners, we happened to get a strong application batch, etc.)

It actually did. It surprised us too!

We can track this directly so there's no uncertainty over the result. We're not entirely sure why it worked so well, though.

Brand Indoctrination. They went to your school, then applied to the next step after a few months of getting your message.

Alright, I take back my comment after looking at the Startup School site a bit more. It seems to offer somewhat more than the basic course I assumed. Looked interesting enough that I joined the waitlist!

I did Startup School and they definitely tried to upsell you to YC.

It’s more than likely true.

Related/meta question: despite the increase in YC cohort sizes, has anyone else noticed a sharp drop in the number of Launch HNs / YC Startup Launch Announcements on HN?

EDIT: Checked the data and yep, there's a drop: https://docs.google.com/spreadsheets/d/1EBcI3Jm2I9Kj_JMGyRnN...


    SELECT TIMESTAMP_TRUNC(timestamp, MONTH) as month_posted,
    COUNT(*) as num_posts_gte_5
    FROM `bigquery-public-data.hacker_news.full`
    WHERE REGEXP_CONTAINS(title, 'YC [S|W][0-9]{2}')
    AND score >= 5
    AND timestamp >= '2015-01-01'
    GROUP BY 1
    ORDER BY 1

I think that's because there are a lot more bio/hard science startups now, and fewer pure software plays. Pure software plays are great for a Launch HN, but the bio and hard science ones probably don't see the benefit, and also have longer cycles to launch.


I’m doing a life science startup and our medical validation will likely take 18 months.

For what it's worth, in my experience 18 months is ambitious.

What's more, it's also probably because companies at YC are generally later stage now.

It's not unusual now to see companies well past launch or even product market fit prior to YC.


It was even talked about a lot from YC about how they are seeing more and more people who reapply year after year and how that is seen as a "positive signal".

And I saw this first hand from our StartupSchool batch. The only people who made it through were ones who had already had significant numbers of customers.

That's actually untrue - almost half of the Startup School companies that we ended up funding for the W2019 batch were unlaunched.

I never said that all Startup School companies had customers.

My point was that it seems like the needle has shifted towards companies who do have some level of product market fit. Has the percentage of companies who were unlaunched changed over time ?

While it's true that we have some companies in the batch that are well past launch, in S18 only 44% of the companies we funded had revenue and roughly 40% were not launched prior to YC.

I also added a tab tabulating YC Job posts on HN; surprisingly they're downtrending too, although not as seasonal as the YC Launches (which makes sense).

I appreciate that you took the time to do the math.

Regarding class size, the incentives are not aligned between the incubator and the startups. For the incubator, the marginal cost of adding one more startup is almost zero, but the chance that you might miss out on the next unicorn is big. So you grow your class size.

For the entrepreneur, the larger class size dilutes you. Yes, the incubator can scale mentoring reasonably well, but the visibility and prestige of pitching at Demo Day is going down.

If YC is graduating 400 startups per year, it's fair to say hundreds of them will not lead to anything. I don't have a good solution for this, each actor is trying to maximize their self-interest.

There are new funding models that have popped up just in the last month or so e.g. Earnest Capital and TinySeed. I suspect for many of these companies they would be better served by those models, which vary but are essentially designed with no expectation that the company must march toward "Unicorn" status to be considered a success.

This is only true to the extent that YC is a monopoly.

If other accelerators compete with YC, entrepreneurs may choose to do business with an accelerator that selects smaller classes.

Are there any other accelerators that have the same cachet as Y Combinator however? The only other one I've really heard of is TechStars.

Erik Torenberg's Village Global [1] was started fairly recently and is already extremely highly regarded in the start-up community. Much more so than Techstars or any other non-YC accelerator in my impression.

[1] https://www.villageglobal.vc/ https://www.villageglobal.vc/network-catalyst/

500 Startups and AngelPad also seem fairly well-known. I'd love to hear from someone more involved in the VC community about how those two accelerators are generally regarded.

The marginal cost of an additional startup through YC is definitely not zero. Just look at how many more partners there are now compared to when batch sizes were smaller.

You’re right that it might be lower than average cost, otherwise batching wouldnt make sense.

> For the entrepreneur, the larger class size dilutes you. Yes, the incubator can scale mentoring reasonably well, but the visibility and prestige of pitching at Demo Day is going down.

I don't think that's true. For one, as an entrepreneur, you get access to a bunch of people going through the same things you are, but in totally different areas, so you get a greater diversity of opinion. And as for dilution, I don't think that's the case either. They have a lot more investors now, more than they can accommodate, so there is no shortage of attention.

The value YC sells for $120k and 7% (the standard YC deal) is “exposure”. It’s been YC’s justification for this staggeringly expensive proposition for over a decade. Having more companies in the cohort, means less exposure, and thus even less upside.

It’s a market. YC isn’t the only accelerator out there, and if the proposition is returning less than before, then the correct solution is to not accept the YC deal.

Startups get a better experience, and Sam Altman learns the lesson about overextending. Eventually, the market will correct.

cue the problem solved song

but the chance that you might miss out on the next unicorn is big

If selection skill counts for nothing, sure, but by your logic why not fund all comers? Where do you draw the FOMO line?

As someone who has sat through a bunch of demo days: No I don't think it's too big. It's a lot more diverse now. Back in the day, everything was a software startup. Now they have a bunch of bio and hard science and bigger startups that have longer cycles.

Not every investor will invest in every space they cover, so having two stages probably won't be that big a deal. Even if a firm invests in multiple categories, they probably have partners that specialize in different areas and will go to their respective stages.

From what it looks like from the outside, they are basically running multiple parallel tracks that get the added benefit of sharing a timeline and having founders in totally different areas to talk to as well as founders in their own area.

Seems like an advantage for everyone.

Maybe this is partially a personality thing, but my biggest fear with starting a VC-backed co is that you're effectively trapped as soon as you accept money even if your company goes nowhere. If your burn rate is low, you could just exist for years without accomplishing much of anything.

For most startup founders, there is an alternative cost of well over $100K salary as an employee, especially for technical founders in the Bay Area.

So lingering for many years is not so great of an option.

Liquidation preferences can dramatically change the math on how big an acquisition needs to be to be profitable for the founders.

Say you've got 3 founders who each want to make at least $2M for a startup they've been toiling away at for 5 years which has built something kinda-sorta useful but not a huge hit (that's roughly equalling what they would've made at Google/Facebook). If they took a $5M Series A at $15M post with 1x participating preferred, they need to sell for ~$17M to hit their target (the VC gets their $5M back, then $4M of the remaining $12M for their for their 33%, then 20% goes to the employees from the option pool, then the founders split the remaining $6M). If they took no funding, they need to sell for $6M. If they took just angel/seed funding for 25% of the company, $8M.

It's generally a lot easier to get acquihired for $1-2M/head than for $5M/head.

I'm looking at https://yclist.com/ (no idea how updated or accurate it is) and seeing a lot of companies are have been around for years and not dead or acquired that I've never heard of. I'm mostly referring to cos that have raised another seed or series A for a few mil and then just sort of lingered. I do not know of every 100 cos how many successfully raise another seed or series A beyond what YC provides (I would assume many do)

Would love to see something like a "YC ETF" where an investor could invest, say, $50k, in a portfolio of YC startups.

I actually have essentially such a thing.. I started it last class, I completely randomly select 10 YC companies and invest $50K in each at their current demo day terms. All accepted the investment last time so it should actually be a valid random sampling. If you’re an accredited investor and possibly interested in joining this “lucky fund” for W19 (I’m doing it again), PM me!

I run one of the YC Demo Day funds on Angel List (not sure if it's an SEC violation to link to it) and therefore spend weeks assessing founders and startups in each batch to make about 15 investments. It'd be very very interesting to compare our returns and outcomes to see if your approach destroys mine or not. Will send you a note!

HN doesn’t have PMs, sent you a LinkedIn request or email is in my bio.

Can't find your email. Mine is on my bio. Thank you.

While a pure YC ETF doesn't exist, there are several YC demo day funds that approximate it. Both FundersClub and AngelList run funds you can invest in every batch that invest in many of the YC companies.

I run one of these funds. Happy to chat with anyone who has questions about them! Email is in my profile.

Series a funding is flat last 5 years... just funding more failures.

You'd need to see if there's a rise in acquhire, that your expensive capital intensive long term biotech project got acquired by a big company is usually not a failure for anyone. Not the investors, founder, or even the project itself.

FWIW I remember this question being posed about our batch (Summer 2012) which had ~80 companies in it.

That was only ~3 years after the economy came out of a deep free fall caused by bad investments. It was a reasonable concern in context.

Too big for what? Without some qualification, the question doesn't make sense. Is it too big for it to be worthwhile to YC itself? YC obviously doesn't think so. Is too big for startups to feel like they are getting a good deal? You'd have to poll the startups themselves and find out. Is it too big for investors? No, investors like having a big funnel.

As an investor, having 200 startups to comb through is similar to being an employer with 200 resumes to choose from. 200 resumes is better than 100. It's your job to filter and choose the ones you're interested in.

Too big to be interesting. As the cohorts grow in size, the returns begin to mirror the returns of the entire startup market.

It's the difference between having a large index fund, and having a portfolio of 5-10 good companies.

Basically it's like they're not even trying to filter much anymore.

“Basically it's like they're not even trying to filter much anymore.”

I believe their acceptance rate this batch was still <<2%. That’s still a pretty fine filter IMHO.

That's not unexpected, with the popularity more companies apply - even some who would have never done it in past. So, I'd assume applications will continue to grow and acceptance rate should be going down. If the acceptance rate stays the same, I guess YC is lowering the bar. This likely happened already given the top ranking post from a YC member saying that acceptance is same still even though many more apply.

I've always been curious about what YC demo days are like but since it's been limited to VCs/investors, figured I wouldn't be able to attend.

With this change, would it be possible to get a live stream or even a post-event Youtube video for the rest of the community?

If not, would appreciate any insights into if/why this is a bad idea.

Unfortunately we can't open the video stream to the public because many of the companies share confidential information that they don't want to be publicly accessible (in particular, visible to their competitors).

If you want to see what it's like, this TC video does an excellent job of that: https://techcrunch.com/video/behind-the-scenes-at-yc-demo-da...

Typo in the first sentence: that’d be pole position

Yeah, that's a motorsports (NASCAR) reference that is more familiar to us HN readers from the midwest. Surprised this one was missed because track racing is big in the VC world.

Just change the name to "Demo Week." Problem solved.

What's the yc deal $100k for 10% or something like that?

$150K for 7%.

What's going on with Techcrunch's format here?

  * Headline
  * Hero image
  * Sales pitch for TechCrunch premium
  * 98 words (598 characters) for this article
  * seemingly non-sequitur into next article

This particular article link is a TC weekly summary (not unusual for tech blogs).

A better URL for this submission is: https://techcrunch.com/2019/02/11/y-combinators-latest-batch...

Are the mods able to change the link? This one is much better.

They got a click. Mission accomplished.

This is what happens when you over-optimize for metrics (premium conversions, article clicks)

All five news topics are one article.

I'm guessing the title was changed after many of these comments were posted? Can anyone fill me in on what it was?

Right now the title is "Y Combinator’s latest batch of startups is too big for one Demo Day stage".

The original article's title was “Is Y Combinator’s latest cohort too big?”: https://news.ycombinator.com/item?id=19186238.

The age of prestigious incubators and accelerators is coming to an end IMO.

As we wrap up the 2010s, I feel that a lot of the entrepreneurial wave in this decade was catalyzed by the founding of mobile app stores, the network effects of expanding social medias and cloud services making software deployments easier than ever.

Incubators seemed well poised to help this flood of new entrepreneurs navigate the changing landscapes and get exposure to investors, but by now people have become more savvy and this is mostly a solved problem.

Investors are also well aware by now that incubators do not magically produce unicorns or even great companies, and are looking for better returns elsewhere. Demo Days are mostly a lazy way to sit and listen to a bunch of pitches without really giving a fuck. I know at one demo day a couple of investors spent most of their time talking in the hallway while founders pitched on. Investors know all your tricks to make your company seem more appetizing: fake appointments in your calendars, strategically chosen metrics, name dropping, paper trails, etc...


if you just look at a sampling of these startups, i think you're wrong - entrepreneurship has always been about applying technologies in disruptive ways and the companies are following the trends - IoT, alternative reality and disrupting traditionally non-tech spaces like assisted living.

you may be right about the rest but then again entrepreneurship is becoming a more mature market and investors actually have more people on their team to "guide" them about what a "hot" startup looks like(hence them seeming bored at pitches that aren't going to predictively skyrocket imo)

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact