5 things we learned from YC W23

5 things we learned from YC W23

Jack Fischer and I recently graduated from YCombinator's W23 batch, launching Credal.ai, landing some pretty amazing customers (including multi billion dollar logos) and closing our Seed round led by truly world class investors (more to announce there soon). It was an incredible program, where we met many remarkable founders, and learned an unbelievable amount, especially as first time Founders. With many of these sorts of lessons, in some ways, the lesson is obvious, but really learning it can still be hard. I thought I'd share some of it, for folks either starting their own startup journey or perhaps even applying to YC's next batch...

  1. Make something people want =/= make something that solves a problem

Many things can solve a problem. But the bar for building something people actually want is way higher. There are a few reasons for that: sometimes using your tool requires new behaviors that users aren’t willing to build habits around, or other times even if the problem is real, it’s just not any particular person’s primary problem to solve. Before Jack and I really understood our customers’ needs, we struggled with this - we were solving a problem that was real, but it was a distributed problem that for nobody actually felt like their personal “hair on fire” problem. To build a business from scratch, you have to be solving something that for at least one or two people at an organisation, feels to them like the single most important thing they need to do. When you’re building features as part of an existing product, you can kid yourself that you know how to build stuff people want because you shipped some feature and a million people chose to use it. But this just doesn’t translate to 0→1.

Building something people want at a startup requires them to simultaneously 1. Take a leap of faith that the software you built can actually do the thing you say it does, 2. Change their behaviour in some material way to integrate your product into it and 3. Put their reputation on the line at their company for something that is highly unproven. Unless they *need* to solve that problem urgently, the bar to do that is not in the same ball-park as building even the most loved feature of some existing product. One thing I've struggled with is how easy it is to kid yourself into thinking that you’re solving someone’s hair on fire problem. Maybe you find some people that get really excited seeing your demo, or even someone explicitly mentions this as a problem they urgently need to solve. Sometimes those can correlate with a good problem space, but they’re not strong indicators. The strong indicator is when they are urgent about setting up the next meeting, when they're going to fight any of their colleagues who are just lukewarm, and when they’re much more worried about how quickly can we get this going than they are about how much is this going to cost, and of course ultimately, whether they actually spend a legitimate amount of money on it in the end.

2. At the beginning, you are the product

On day one of your startup, you have no product, no moat, no credibility. And yet you can still try and sell to customers. That means that who you are is critical to the customer’s buying decision. To sell in that context, you need to persuade your customer that your unique life experiences can be a source of advantage for their business! Of course, you’ll need to marry that insight with a *lot* of customer conversations, but at the end of the day, the life experience of the founders is the literal ‘Foundation’ of what gives your startup an advantage. Lean into that, and really focus on building the product that *only you* should build, not the product that seems hot at the moment in time you are building it. For founders with real experience building stuff already, you may as well use it! The first thing Jack and I set out to build was tied to a real problem we had experienced, but it wasn’t tied to the unique aspects of our experiences, and so we found ourselves building something analogous to things several other founders in our batch were building. We weren’t particularly expert in the lives of the operational users we were selling to, and so we felt like we were playing characters in an unfamiliar video game, suddenly trying to learn all the moves from scratch. The knowledge and intuition that made us effective before was no longer relevant. However, the product we were building was differentiated in one important way: Jack and I had obsessed about the security story for our product, because we had spent the last 10 years building software on top of highly sensitive data from the DoD, HHS, pharma companies and more. In the end, it was the security features we built that really resonated with customers. We leaned into that, and went back to selling directly to the people that we knew the best: enterprise IT and infosec. When we leaned into what made us unique as founders, everything changed: we’d gone back to our favorite, familiar characters, with fully upgraded weapons and armor, and although everything was still hard, at least now we weren’t having to relearn the basics from scratch. Sales conversations were suddenly much easier! If you're working on a problem that makes sense for you, your customers will notice that, investors will definitely notice that, and even the people you hire will recognize it too, but ultimately the most important thing is that you will find you are personally driven to make the startup succeed because you know deep down in your heart: nobody else will.

3. If people don’t want your product, no amount of clever sales tactics will compensate. But if you have a product people want, deals still close even when you make mistakes.

In the first product we built, we had this really cool demo, showing off lots of flashy tech and many integrations, with a powerful narrative storyline etc. For all that, it was super hard to sell the product. Now, I find that even when I give demos that feel super lame and take 5 minutes - yet the customer is often immediately like, OK this is exactly what I need, how do I buy it? Of course, I still really want to get a slicker demo in place, etc, but I know I need to be laser focussed on growing the business, and the truth is spending a few hours nailing the narrative, and demo data, click path etc is all well and good but isn’t what’s going to move the needle the most on additional revenue. Sometimes I make mistakes in my rapport building, or do a mediocre job asking enough questions at the beginning - but for all that, many of these calls still end in “OK, this is what I need, how do I buy it.” Before we landed on the right problem, we had the exact opposite issue - users would be like “this is so cool, I’d love to try it out” and then when we were like, sure, it costs this much it would immediately become “ok yeah I’d rather not have to pay you for it,” or they’d just go cold.

4. There are a lot of investors. That means that any heuristic that applies to a large group of people, will apply to investors: some are great, some less so. Humility and engagement let you know which are the smart ones.

Like any diverse group, investors can vary greatly. Some can provide substantive insight, ask thought-provoking questions, and even actually be fun to talk to! Others, with limited to no experience building companies, act as if they know the secrets to success, assert misguided advice as fact and overall just clearly have an IQ of 110 which they think is 140.

I learned to be especially skeptical when investors assert their opinions as facts, and a huge red flag would be when they did this regarding our fundraising process, valuation, amount being raised etc. The most effective investors recognize that their best investments emerge when their goals align with yours. You do want them to be able to share their perspectives on deal terms honestly, even when it doesn’t perfectly align with yours, but the ones trying to bully you into accepting the shitty terms they need for their arbitrary fund goals are not the ones you want on your cap table. (One tactical thing I learned from this process is that even the best institutional investors have lawyers who put silly terms into their deals because it's on some template somewhere. All the investors we worked with were really reasonable at recognizing and accommodating when we pointed out that certain terms shouldn’t really apply to us).

Remember, when it comes to your industry and your business, a good investor knows that a founder worth investing in will always have a deeper understanding than they do. Although you may encounter some pointless meetings where an investor's mind seems only half-engaged, or they attempt to pressure you into unfavorable terms, there are numerous outstanding investors who ask insightful questions and engage in meaningful discussions.

The best ones you'll encounter will be extremely smart and legitimately knowledgeable but also aware of the limits of their knowledge. As a result, you can learn stuff from them and, hopefully, share your own wisdom and insight with them at the same time.

5. Go faster

This is something that I really felt like I learned from the other founders in our batch. Michael Seibel would remind us frequently - the vast majority of companies fail. If you are moving at the same speed as the vast majority of startups, you are going to fail. The vast majority of startup ideas ultimately are invalidated. If you aren’t able to invalidate the dead ends fast, you can waste all of your runway on one or two doomed ideas. Since almost all startups fail, the bar for a startup to succeed is extremely high, which means conversely the bar for invalidating a startup idea is not that high (caveat - sometimes invalidation doesn’t mean throwing the idea out completely, but rather, approaching it from a different angle or framing).

YC is, of course, a hotbed of successful companies, but even there, the majority of startups don’t succeed. But seeing some of our batchmates move at blistering speeds was a wake up call for us that we also needed to do the same or more. Many of the most successful companies had changed their idea mid-batch; seeing that helped us internalize that taking many bets and learning quickly was actually much more likely to yield success than simply placing one big bet and putting all our chips into it. Maybe if we were Steve Jobs that would be fine - but we are data driven people, we do have a vision for where we want to take the product but we know we need to marry that vision with what our customers actually need. We drew a needle on our whiteboard telling us the probability of success for our startup as things stood - it started out at 2%; roughly the hit rate for median YC startups. We would try to collect evidence that would either push that needle up or down: how did customers react when we quoted pricing? How quickly did customers onboard? Did users engage and become sticky? Each answer would move the needle up or down. Was this in the 98th percentile for what we were hearing from the batch? If we could not move the needle up each week, we needed to rethink our framing/approach. To keep ourselves honest, we actually made predictions over things like user growth and engagement each week. If our predictions were optimistic relative to what happened - the needle went down. If they were below - the needle went up!

All this obviously means a lot of hours, but more than that, it means only doing maximally leveraged things with those hours. Some examples: we started our outbound campaign selling our PII / PHI redaction product before we even had the product fully built out! By the time our larger customers had asked for a demo, we had patched one together. By the time they bought it, they could just log in and it would work! Everything was evaluated in terms of what would have the biggest impact on the needle.

Building blocks towards secure AI apps

Credal gives you everything you need to supercharge your business using generative AI, securely.

Ready to dive in?

Get Started