Hacker News new | past | comments | ask | show | jobs | submit login
Investor Day (ycombinator.com)
257 points by dshankar on July 7, 2016 | hide | past | web | favorite | 88 comments

As an investor, I think this will accomplish a few things:

1) A lot of scheduling friction will disappear. The week just after demo day is usually crazy because hundreds of founders and investors are all trying to schedule meetings with each other, and there are inevitable race conditions that lead to a lot of rescheduling and wasted time. (E.g. I email three founders with 5 possible time slots, and they all reply and ask for the same time slot.)

2) I think this will be great for investors who act quickly and go by their gut. There are plenty of investors out there -- especially those who write smaller checks -- for whom 20 minutes will be enough time to make a quick decision.

3) I'm not sure if this will be great for investors like me that approach investing more methodically rather than with their gut.* I love 1-hour meetings because that's plenty of time for both sides to dig in and learn a lot about each other. Twenty minutes feels very short to me, and I'm not sure if a 20-minute meeting is more likely to save me and the founder from an unnecessary 1-hour meeting, or if I end up having just as many 1-hour meeting -- but now with an extra 20 minutes tacked on.

That said, I don't want to judge this process before I try it at least once, and I'm looking forward to trying it out in August.

* FWIW, there are great gut-based investors and great methodical investors, and I'm not implying either approach is better.

I am an investor and my experience is, if you have done enough homework about the startup you want to meet, twenty minutes shall be long enough to clarify a few key things so you can make an informed decision.

I personally like asking a lot of questions, so I'd still prefer >20 minutes. I 100% agree with you though that preparation can save a lot of meeting time. One catch with YC companies in my experience is that they rarely have detailed pitch decks or significant web presences, so there's often not that much to research before a meeting.

Did you check out The Macro - http://themacro.com/ ? Yes the info disclosed by YC for each startup is still limited but better than nothing in the before.

Definitely helpful, but AFAIK only about a dozen startups from the 100+ startup batch have been covered so far (https://www.google.com/#q=%22meet+the+batch+s16%22+inurl:the...). I can imagine the coverage will double or triple by demo day, but I'm skeptical that all startups will be covered. Btw, I love the /data subsection of your website!

How long would you want?

I typically do 1-hour meetings and that works pretty well. 45 min would be okay. I haven't tried 20 min before, but that feels a little short. To apply the over-applied analogy that fundraising is like dating, 20 minutes would be enough time to tell if a date is terrible, but I'm not sure if it's enough time to distinguish between a B date and an A date.

Two ideas related to Investor Day that I'd be interested in:

- let investors ask for one or two consecutive 20-minute slots. Founders can just allocate one slot to someone asking for two, but asking for two slots would be a good signal of investor interest and/or investing style.

- don't automate scheduling, but instead host all founders at a single location for a few days, and make it easy for investors to book slots on founders' calendars once those founders opt-in to meeting up. Maybe this could use something like https://calendly.com/

Love the "open to experiments!" attitude. I think you nailed it here - 20 minutes is enough to rule someone out.

If our original scenario is:

Scenario 0: 25 1-hour meetings over the course of a few days. Investments = 25-n

You seemed to be suggesting in an earlier comment you might end up with:

Scenario 1: 25 20-minute meetings followed by 25 1-hour meetings. Investments = 25-n

I think you'll actually end up with:

Scenario 2: 25 20-minute meetings, eliminate x companies, (25-x) 1-hour meetings. Investments = (25-x)-n.

Saves you and the founders some unnecessary hour long meetings.

Roughly speaking, our investing process is: 1 hour one-on-one meeting, then 1 hour full partner meeting, then decision.

The way I think about the math: Current scenario: 25 1-hour one-on-one meetings -> 10 full-partner meetings -> 2-3 investments.

Worst case for Investor day: 25 20-minute meetings -> still don't have enough info -> 25 1-hour one-on-one meetings -> 10 full partner meetings -> 2-3 investments.

Good case for Investor day: 25 20-minute meetings -> 10 companies don't seem like a fit -> 15 1-hour one-on-one meetings -> 10 full partner meetings -> 2-3 investments.

Best case for Investor day: 25 20-minute meetings -> I get all the info I need -> 10 full partner meetings -> 2-3 investments.

These four cases represent 25, ~33, ~23, and ~8 hours of one-on-one meeting time, respectively.

lpolovets - shouldn't the follow up one-on-one meetings be only 40 minutes?

That's a good point. I guess they could be!

40-min meetings might be a little awkward because usually the first 5-10 minutes of a meeting are social / chit-chat, so it might feel a little weird to immediately start a 2nd meeting exactly where you left off on the first meeting. Also, most of the time a 40-min meeting still takes up a full 1-hr time-slot since most people schedule meetings on the hour.

This is probably sufficient for many angel investors, a few seed investors, and very few series A investors. No matter how much homework you've done on a company you can't responsibly deploy $5M after 20 minutes of diligence on the team.

Since this is YC, "many angel investors, a few seed investors, and very few series A investors" are the target audience.

Out of curiosity what constitutes that homework for you and do you feel like you can complete that homework prior to demo day?

>>what constitutes that homework for you

This is the homework we do:


>>do you feel like you can complete that homework prior to demo day

Yes the time is very tight so we have built an app to automate that.

I think you'll still be able to be methodical afterwards via direct communication over emails/calls to the startups, just as you used to before. Unless this Tinder-style app removes all the contact info you'd normally receive, you should be able to take it slow if you want.

I agree that I can still be methodical after Investor Day. I'm just saying that before, if I was interested in 25 companies after demo day, then I'd do 25 1-hour meetings spread over a few days. With Investor Day -- and without a chance to do much prep/research before each short meeting -- I'm not sure if I'm going to have 25 20-minute meetings and then still have 25 1-hour meetings to follow up, or if it's more like 25 20-minute meetings and then 5-10 1-hour meetings.

In what percentage of 1-hour meetings do you eliminate the company within the first 20 minutes?

As a CEO, I've eliminated at least 25% investors in that time period. Given that most meetings start with the company pitch, my guess is 20 minutes would be enough to winnow down the list of hour meetings.

It's hard to say. If it's literally my first exposure to a company, then probably 80% or 90%. But demo days are more complex. I might see 125 companies pitch on stage and try to follow up with 25 of them. So that's 80% eliminated from the demos. For the other 25 companies, I already know a little bit from their pitch and find them interesting, so I'm not sure how many of them I'd be able to eliminate within 20 minutes. My guess is 25%-50% (of the remaining 25), but it's a wild guess.

> there are inevitable race conditions

I've found https://calendly.com/ super useful for eliminating race conditions in scheduling, especially across time zones.

20 minutes is a good amount of time to determine if you like the Founders or not, but I agree that you probably want to do more digging. Would love to see YC post a "post mortem" about how all of this went.

> ...if Investing/General Partner is not present, the slot will be cancelled.

These seems to be an attempt to reinforce FOMO to force key people to attend demo day. Resorting to these tactics implies a pretty big perceived power imbalance... Eg. is this a reaction to senior partners sending their underlings because of DDay burnout? If so, this could backfire. The senior partners might just not show up (still), and force founders to attend offsite meetings later anyway (in addition to the new DDay meetings). In other words: This might just be a net increase in founder effort without changing the investor-founder DDay dynamic. After all: What do the senior partners gain / lose by this new situation? Not much (aside from FOMO), I'd guess. But the best investors will always be in demand anyway -- whether they attend DDay or not.

Actually, for a long time now we've only allowed partners to come to demo day.

We are relaxing this restriction and allowing partners to bring associates to Investor Day, but only if they come themselves.

It seems to me that this rule is beneficial to the companies, which I think is the right side for YC to err on.

Why the downvotes?

Seems like a well-reasoned hypothesis and contributes to the overall discussion. Has HN become that critical of different opinions that downvotes are used to punish dissent rather than using upvotes to indicate agreement?

Sometimes downvotes mean people just disagree not that the comment was malformed.

I actually UPVOTED when I saw "implies a pretty big perceived power imbalance". It's possible that the downvotes took that as a negative where I was in agreement.

Downvotes (for whatever reason) mean that more people vote down than vote up (duh). As a result they can change simply because of the people that are reading a particular comment stream. Same comment on a different post could have entirely different results vote wise (at least when it's an intelligent comment as yours was).

Who cares about downvotes! just enter the "Aspergian mindset" and write what you think is right.

If you are looking for social feedback you will never get far on startups.

Well, users are social feedback . . .

Are investors weighted by their fund's size? Or are founders told the ranking they were given by each interested investor?

I can imagine many firms go into demo day with the resources and willingness to fund 10-20 ventures, whereas others are looking for only one or two.

I would much rather be the fifth choice of the former than third choice of the latter. If I knew my position in the stack I could figure it out for myself, or it could be calculated for me if YC is trying to avoid the negative consequences of making that transparent.

Investors are not weighted by their fund's size - they are ranked by the founders individually. Founders would be able to apply your stated preference for meeting with larger funds or with known high frequency investors, if they were so inclined.

Without knowing how one ranks, though, it's substantially harder to prioritize.

This doesn't hurt the big funds so much as it does the small ones. If I were an eCommerce or Logistics company, for example, there are a subset of niche funds which would be great strategic investors; but knowing they're only going to invest in one or two ventures would leave me disinclined to prioritize them when high-frequency investors are in the mix.

I don't want to kvetch here, it's a much better system than the current free-for-all. But since investors already get the first-mover advantage, it would be nice if entrepreneurs had more complete information.

To carry the theme of relationship analogues in this thread: knowing who's interested in me is great, but I'd rather be someone's dream girl than a fallback or fling.

We do have a lot of context on investors that can help founders when choosing who to meet with. In addition, we're going to let investors specify a reason why they want to meet with each company (optionally). This should give less well known investors who are highly suited to particular companies a chance to get noticed.

> we're going to let investors specify a reason why they want to meet with each company (optionally).

Even better! You guys have really thought this through. :)

I really like that they cancel slots when no partners of the VC firm are present.

I assume this is using the same or similar algorithm as the NRMP algorithm uses for U.S. residency match day?

Yeah, the Stable Marriage problem provides a solution for matching these ordered preferences between two sets of people/elements: https://en.wikipedia.org/wiki/Stable_marriage_problem

Very close. It's many:many with multiple iterations until a stable solution is found.

Could you write up the differences? I'd really enjoy reading that.

We plan to in the near future.

Almost definitely. Stable matching is a very well understood problem, and as far as I can tell investor day is a textbook example of it.

I assume so as well.

This algorithm is outlined by the stable marriage problem for those that are interested.

It's not quite the stable marriage problem, because in the stable marriage problem you're trying to find an optimal 1:1 pairing. It's also not the NRMP problem because in that case, residents are only matched with one hospital in a year. Here companies are matched with multiple investors, and investors are matched with multiple companies.

But regardless, I believe the algorithm is roughly the same as the NRMP algorithm: you run through multiple rounds where you look at each combination of company and investor. If both sides would prefer that they meet with each other, rather than the people they're already paired with, then you reassign them. Once you've done that several times you arrive at a stable state.

Yes you are on the money. There are also some additional constraints, e.g., we don't want to match startups with multiple investors from the same fund.

In the NRMP there is a N:1 matching between residents and hospitals, where N <= the number of open positions in the hospital.

If the meetings are time sliced, let's say by hour, then run the algorithm for each time slice.

Well, with the constraint that no startup wants to have multiple visits with the same investor (and vice versa) - so you need to rerun the algorithm with the earlier matches removed from a preference list.

Depending on how this is organized, this can be a spectacularly BAD idea.

Investors would love it because they now have visual confirmation of who all are the "hot" matches. I bet everyone other than the Sequoias would be straining to look at who the hot startups are...and completely ignore the ones in front of them.

For the long tail of a YC batch - the ones that are not hotly contested - this could be a disaster. Previously, investors would be forced to actually look at a startup and decide in isolation. Now they can simply look at the Big VC.

I can completely see why investors would love this.

One might as well make public the interest match list and rank it by order of "likes" received.

People are perfectly capable of driving up and down the Bay Area. Guess what - we get a few free meals and coffees out of it.

EDIT: guess what, you can bring associates to tailgate Sequoia & A16Z investors.

EDIT2: what you guys might be trying to do is be helpful. For example, this lets you force-schedule investor meetings for startups that had no investor interest and term it as "the AI did it!". But I'm not sure if that will be really helpful ... at the cost of drastically reducing FOMO factor for most other startups.

I don't see anything in the description that allows investors to see an aggregated ranking of YC startups.

I said - it might as well be that. They can physically do that now (using their associates,etc) by checking out the marquee VC activity in the room.

Even if these were individual rooms, you would have significant tailgating behavior.

People don't realize this, but the most popular question of demo day is "who else is investing in you"

While earlier every deal was always evaluated in isolation.

Is there a signaling issue introduced at all?

1) can investors see that a startup is not at the venue or at a table all morning and get a feel for demand?

2) can investors gather any information from the matching results to get a feel for demand?

3) is there a chance for investors to send false signals by showing their interest in other startups artificially? Vice versa for startups at all (would require collusion so unlikely)

2 and 3 not so much but regarding number one, will physical observation of the meeting space introduce any signaling oppurtinities be it genuine or fraudulent by either party?

This is exactly what will happen (re: my comment in this thread). This is a highly investor friendly move and will benefit the top 10 startups of a batch. This will worsen the situation for the long tail.

I don't think so.

1) There are way more investors than startups and we anticipate a busy day for everyone based on past data. 2) Not the demand for a company. 3) No.

Seems like a solid idea. I honestly thought something like this was already in place so it makes sense. Still though the first thing I thought of is what it might feel like to be part of the statistic few who get zero investors clicking the button (I mean statistically this should happen at least a couple of times due to the batch size, right? Or am I severely underestimating the amount of investors that attempts do demo day?)

If you can't get a single investor to give you 20 minutes after Demo Day you've got much bigger problems.

A few years ago sure, but 2016 has been pretty brutal for seed and series A. I imagine a small percentage of great companies will lose out as a result of a cold funding climate.

Seed is roughly the same as it's always been. Even if a lot of VC firms are retrenching, there are still tons of angel investors and seed funds out there casting wide nets.

Series A, definitely.

20 minutes? I thought they got an order of magnitude less time. Regardless yeah it's probably a sign of an issue but with the current climate I don't think that necessarily has to be true.

The linked article states Investor Day appointments are 20 minutes.

Yeah I read your comment wrong and that it was more like "if you can't get an investor to talk to you after your 20 minute demo...". I was probably lacking coffee that day.

There are a lot more investors attending demo day than startups so this is not likely to be an issue for the companies.

Not every YC company will successfully raise money, but I suspect all of them will at least have a few people interested enough to take a 20 minute meeting.

Yes if investors are not forced to meet with X number of startups

This is extremely valuable data for YC to capture for themselves. I wonder how they might use it other than for a match-making algorithm..

This is great, it would have saved us (GitLab) a lot of driving around. It is one of those idea's that are very obvious in hindsight but for sure I didn't think of it.

The stable marriage problem has two natural algorithms that produce a stable matching: one algorithm give priority to women's choices and one gives priority to men's choices.

If the algorithm used by yc to produce the matching for each slot is based on these, which algorithm is used? The one that gives priority to women's choices or to men's choices? A variation that doesn't give priority to any gender choices?

The stable marriage problem has two natural algorithms that produce a stable matching: one algorithm give priority to women's choices and one gives priority to men's choices.

This is true under two conditions: 1. Marriages are heterosexual; and 2. Marriages are monogamous.

While startup/VC relationships are heterosexual (err, heterofinancial?), they are absolutely not monogamous, so the analogy to the stable matching algorithm already fails.

I don't think those algorithms are applicable here, because it's normal (even good) for investors to invest in multiple companies.


The logical optimization target for YC is the total valuation of their companies, which probably means whatever approach creates the most competition among investors.

Edit: more active discussion on this topic at https://news.ycombinator.com/item?id=12051218

For clarity, the "women" : "men" are "investors" : "founders" in this context

Reminds me a lot of Waterloo's co-op matching system, Jobmine.

Which is an absolutely awful system, for a multitude of reasons.

One of the biggest faults that looks like it might be replicated here (the article does not specify) is that once employers (VCs) rank students (startups), the students can't see the numerical ranking they were given.

For example, if I'm a student who was ranked by several companies, I cannot see the numerical ranking they gave me. This introduces a significant bias against students for no apparent reason; employers do not lose out if we can see how they ranked us.

This is one of the reasons that UWaterloo as a school is hugely overrated (although the students are generally as good as they are hyped up to be).

Why is sharing the rating with students so critical? I can see why people would be curious, but I'm not sure why it matters so much?

Because students don't know how much a company wants them, and so cannot set their ranking with that knowledge in mind. For example, Apple could rank a student a 9 and Microsoft a 2, but to the student it looks like their level of interest is the same. The student could personally prefer Microsoft, but have a low chance of getting matched, whereas if they knew Apple was closer to a sure-shot of having a job, they might be more risk-averse and rank Apple as their most desired over Microsoft.

I still don't see how on average it's worse for students. It might be worse for student A who would have applied to MS knowing a higher chance of getting in, but that's then probably worse for student B who genuinely wants to work at MS and doesn't get the slot that A took.

Isn't it better if everyone expresses their true preferences and then we do the best matching we can on that basis, rather than having students attempt to game the system in response to company ratings? Your description makes the system sound close to optimal to me.

> Using both these ranked inputs, our software will create “Investor Day” schedules for each investor and each startup, which will take place at the Computer History Museum on Wednesday 8/24.

Sounds like the backend of a dating site for startups and investors.

This is almost exactly how some law schools run their on-campus interviewing process (except the bit about actual decision-makers being required to attend). From what I've observed, it works really well. Pre-screening for at least some level of mutual interest can save everyone a lot of time.

From the way they've set this up, it appears to be optimal for investors and pessimal for founders. Anyone else notice the same or disagree? It's in the proof for a particular theorem.

I disagree and believe this should be better for founders. Demo Day works so well because it's a forcing function for getting a lot of investors to see you at once and make a decision quickly, because so many other investors are deciding at the same time. Typically, investors want a long time to consider a company, and they only move fast if someone is / might bid against them. Scheduling meetings post-Demo Day is tough, and most people do it sub-optimally (by scheduling them too far apart). This new system makes meetings happen faster and more per day, increasing a chance of getting a bite quickly, which should lead to a higher chance of closing the round.

The downside I see is that if you are bad at meeting investors / need to iterate on your pitch, these meetings won't allow you to do that, as they'll be over before you can rethink it. But I think it's generally worth more / faster meetings.

*As another investor pointed out, 20 minute meetings may also be too short. I am not sure what the optimal meeting length should be. 30 minutes might be better. 1 hour is probably too long (you can always schedule a follow up meeting).

The way the algorithm is written actually balances out the preferences of both groups. Investors and founders will each get to meet with their counterparts based on a ranking they supply.

What makes you say that? This seems to me like a great idea that gives startup founders a quick pitch and an instant response from potential investors. It's almost like a match.com or Tinder for startups/investors.

It is optimal for the more scarce and selective group, whoever they happen to be in a given batch.

"all the companies will be set up at tables to meet with you face to face."

A study on speed dating showed an increased opinion from the person that approaches the other. Intentional?

If you're interested, please swipe to the right...

This is cool. Good to see YC improving.

This is so romantic!

I don't know, your answer very wrong to me. What do you even know about math.

Did you even win the Putnam, if not then please don't be bolder than the parent poster.

We detached this subthread from https://news.ycombinator.com/item?id=12051967 and marked it off-topic.

That joke ceased to be funny about 8.5 years ago.

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