We want it to be very crazy and memorable because people don't ever do this in enterprise. And I jump over the roof of the stall, and I'm like, wait. Stop. You know, don't flush. And he's like, I don't have an option to use a small flush.

And I'm like, well, you know, with Graybeam, we give you the option to use a small flush.

Kyle is one of the cofounders of Grey Beam, a tool that helps you reduce your snowflake bill. I met Kyle at the on deck program, and him and his cofounder, Arsham, are just two of the funnest people I've ever met, and they've just been doing this really wacky marketing. So I wanted to get him on and hear his side of events and what they've been up to. Enjoy. To give background, so we know each other because we did the on deck program and you guys, I were like some of the funniest, most fun, cool people that I met.

And it's been really cool to see you guys just I like a doing well, but also just having a lot of fun. And I see this just like you're posting some your your message is coming through, but you're also posting it in like ways that are like quite ridiculous. So would say I've seen you like, it looks like a pile of coats in the corner, and I just see Kyle emerge and start talking about the benefits of of Greybaby and stuff like that. So it's been brilliant to watch, and I wanted to kind of invite you on and hear a bit about the journey. How long has it been since you founded Graybeam now?

Yeah, Jack. No. Thank thanks for having me here. And I would say the current iteration of the product has been since February, so it's been about nine months, maybe less. Okay.

Yeah. Arshem and I have been working together since 2023, maybe May. So it's been a little bit closer to two and a half years. We pivoted a number of times, I'd say, like, six or seven times to to get here. And when we first met last year, we were still solving the same issue of snowflake cost, but we did it very differently.

Ended up pivoting in February, and then we've been building that solution since.

Yeah. And maybe just before we even get into, like, where you're at, I think, a, it seem seem like you guys were best friends before, had known each other a long time. Right? It's fair to say.

Yeah. Let's go with that. Yep. We yeah. We've known each other since high school.

Separate friend groups then. We went to university together. We had some mutual friends, for the first year, maybe first, like, first term actually in university. We were hanging out quite often, like, going to a bunch of parties, you know, just doing your typical first year stuff. And and then I think Arshan dropped out in the second or third year, and then we reconnected several years later through mutual acquaintances as well.

And then we started hanging out much more often than that.

Okay. Yeah. But it's it's it's a very good relationship, it seemed. From from everything I see, it's fun to watch you guys hang out.

Yeah. Yeah. Definitely. Yeah.

And this was like, you kinda mentioned pivots, but it seems like, to me, I actually didn't realize it was a pivot because it feels like very much on the same thread that you guys have been pulling the same thread for a long time. What like, what's the what is the problem you're solving, and how did you kinda come to the conclusion that it should be solved?

Yeah. The problem that we're solving today is just the burden of managing Snowflake costs. So for our customers now, the immediate value proposition is that we're able to reduce their Snowflake costs by very high rates in the orders of magnitude of 70 to as high as 95%. And, you know, we stumbled across this through just talking to a bunch of users. You know, I I used to lead data teams at several different start ups, and many of the different pivots that we that we had were a result of my hypothesis in the pains of managing a data team, this being one of them.

So before we actually landed on trying to trying to solve Snowflake cost, the first solution we tried to sorry. The first problem we tried to target was AI data analyst. You know? When I was leading a data team, I had a lot of the same questions come up over and over again. Even though we deployed self-service business intelligence tools like Looker, people would always come through the ticketing funnel.

And I was just know if we had a AI data analyst able to to resolve some of these simple questions, you know, it wouldn't have to burden our team to answer them. So the the first solution was to build build that when GPT three point five first came out. Ended up taking it to market. Didn't really work out. And then the next problem that I faced was managing my data catalog, and that was a huge pain.

And this was because, you know, a data catalog is only as good as the content within it. And before LLMs were a thing, we actually had to manually go in and complete all the different catalog documentation to find the assets, to find the column tables, etcetera.

Sorry. What is a data catalog just for those noobs like me that are not familiar?

Yeah. You can think of a data catalog as a repository of descriptions of the different assets that, you know, your database has. So for example, maybe, like, you have a database that contains all of your podcast information. So you would have a definition. No.

This is the database that contains all of the podcast information. You might have different schemas in there. Maybe there's a a table or a schema of guests, and then you have a definition in there. And then within the schema, you have different tables. I'm not sure what, it would pop you up be populate.

So your data catalog would contain information describing all these assets and potentially how they all relate to each other.

And this is, I guess, something that because you don't hear this talked about much in, like, people just using, like, Postgres and stuff, but it becomes much bigger deal for data analysts and data teams.

Yeah. Exactly. So, typically, teams that have even a small number of analysts. And then if you compound that with a number of business users, maybe, like, 50 to 60 or plus, you know, all of them might have very similar questions like, hey. You know, does this type of data live in our database?

You know? Do we have a list of podcast guests in our database? So they would be able to go into the catalog and identify, oh, we do have this, and this is in this table, and these tables are queried in this dashboard. So data catalog can contain information like that.

Okay. Super cool. And sorry. I interrupted you.

Oh, yeah. Yeah. So, you know, one of the pain points of using a data catalog is keeping it up to date. So when I was leading a data team, you know, we were a small, call it, like, five person team. And outside of just maintaining the infrastructure itself, you had to, obviously, support end users.

And then now you have all these questions coming in. And sometimes it's like, yeah. Does this data live inside our BI tool? And often oftentimes, that can be found in the data catalog, but then we also had to update the data catalog to keep it you know, being able to keep up with these questions that come in. So that was quite annoying, and this was maybe two summers ago, Arsham and I or actually, it was last summer.

Arsham and I went to the market trying to understand whether or not, this was a huge pain point for some of our other users. So our hypothesis was, you know, what if we could create a data catalog that just scrapes information from PRs, from tickets that people submit, from Notion docs to be able to have a self populating data catalog? Ended up turning out that people didn't care. You know? People really either didn't have the budget just to bring on a data catalog because there's intrinsic value.

Right? It doesn't really, reduce costs or increase revenue. And the second piece was just people were cost saving overall. So many of the our interviewees were responding with the same answers, so we decided, hey. Actually, maybe maybe we should tackle costs.

And then lo and behold, Snowflake cost was always the first, if not the second highest line item in almost everyone that we talked to.

Wow. How how much do people spend on Snowflake? I mean, I know that's like a very, very broad question. But

Yeah. It it can really vary depending on the profile. But, you know, typically, when we're talking to folks, on average, people are spending between 50 to a 100 k. I would say if you are a fairly mature company, the average you're probably spending is, like, 300 k plus or minus, like, 200. And then in into the late startup early enterprise, you can be, you know, 1,000,000 to as high as, like, 25,000,000 I've seen.

Scaling DevTools is sponsored by WorkOS. At some point, you'll land a customer who needs enterprise features like audit trails, SSO, role based access control. You could spend ages tearing your hair out, building these things yourself, or you could use WorkOS. Let's hear from Uplal from digger.dev.

I can speak for open source companies because I think that's where I have the most experience personally. If you're open source and you're doing enterprise first, the minute you think about monetization is when you should think about Work OS. How it's designed is that you can start as early as day zero. But for us, it wasn't day zero. It was closer to when we first started monetizing because we didn't have sign up at all.

People could just anonymously use our tool. To be honest, if we do that again, I think I would think about that on day zero, to be honest. Because, like, we should have done it on day zero ideally. Anonymous usage should be permitted, but you should know who's using your tool. It should be opt in a 100%.

But it'd be great to have auth from day zero. So, yeah, that's what I think.

Thanks, WorkOS. Back to the episode. What are they storing in Snowflake typically? It's like usage data or like

The first things that people would store is going to be the application data. So, you know, if you're hosting Facebook, for example, you're gonna be loading up all the different, like, user analytics and pages and logins, etcetera. And then as companies evolve, they'll start bringing others other data sources. So, typically, those would be either customer support data or their CRM data, so Salesforce, maybe some event based data, like clicks from coming from, like, a heap or post hoc, some of these other things. And then, of course, as a company continues to grow and grow, you're just bringing in more and more different data sources in.

Okay. Super cool. And okay. So what so people were talking about cost. And then what was the first kinda step that you guys made?

So we were taking a look in the market. Previously, where I used to work at Wrapbook and Clearco or not Wrapbook. Sorry. At Clearco, they deployed a cost optimization tool called Select dot dev. So we were looking at companies like that and adjacent to them, and we just noticed that almost every single tool optimized Snowflake in the exact same way.

And without going into too much detail, Snowflake works by turning on what's called a warehouse, and warehouse uses compute to process your queries. And what they did was just turn it off when it's not being used. Very simple. Mhmm. And every different optimization tool did that in some JSON flavor.

And what we realized was that there was no one trying to actually optimize the other half of the equation, which is actually all of the queries being sent to Snowflake. So with LLMs being very hot at the time, we thought that, hey. What if we could have some framework that the LLM could use to actually optimize a SQL query? Because if a query runs faster, that means you're using the Snowflake warehouse for less, which in turn translates to, cost savings. So we went that route, which is when we first met in OnDeck last year, and we tried that for about a month or two.

We had, like, three design partners, and we just quickly came to realize that doing pure SQL optimization was not gonna save as much as a hit we had wanted unless we had the perfect customer. You know? Maybe they were using some BI tool that wrote SQL queries in this canned way that was very inefficient. And then you pair that with a team that perhaps was didn't have a data engineering background by trade, and they wrote just fairly inefficient SQL queries. Then you have a perfect storm where, okay, SQL optimization can work, but, obviously, that's very few and far between.

So we ended up going back to the drawing board in December, and we noticed in the industry there was a huge trend, of a technology called Iceberg, and also of a second technology called DuckDB. And, what we realized was that, you know, if people are moving on to Iceberg and what Iceberg means for some of the folks who don't understand is Iceberg is an OpenTable format, and you can think of that as a universal data format. And more and more data warehouses started adopting support for Iceberg. And what that means for companies is that if I store my data in this Iceberg format, I can actually access it using other forms of compute. So for example, if now I'm storing my data in Snowflake, instead I move it over to Iceberg, this means I'm keeping the data in my own personal blob storage, whether it's s three or, Google Cloud Storage or or Azure.

And Snowflake can now actually query my data where it sits instead of having first moving that data into Snowflake. So if I store it in Iceberg, I can now access it using BigQuery or Databricks or any of the other plethora of open source compute engines. So with this movement, Arshan might have thought, hey. What if we actually help customers, you know, maybe secretly move them on the iceberg, and then we can access their data using other forms of compute that could be much, much cheaper than Stuff Lake? So this comes the other piece of the equation, which is DuckDB, which has been taking the data world by storm last year.

And DuckDB is an incredibly lightweight query engine that gives you the performance of Snowflake except it can run on your laptop. You know, you can be querying gigabytes of data, within seconds using this very, very lightweight query engine. So we combined the two, and we decided to see whether or that would work. And in our test cases, in most cases, it is as performing as Snowflake. So we thought, hey.

If you bring this to the enterprise and we help them move their data into this universal format, can we just intercept these queries and process from using DocDB into the Snowflake? It turns out you can, which is why we're able to deliver these very, very aggressive savings between 71 to 95%.

Wow. And so we does most of it comes from DuckDB or most of it comes from using Iceberg, or is it you have to use them together?

So for us, Iceberg is kind of like an implementation detail. We actually haven't even implemented Iceberg for a few of our customers yet. And, really, what we're doing is unloading our customers' data from Snowflake into the customer's s three bucket as Parquet files, and then we just interact with those files using DuckTB. Where Iceberg really needs to come in is if we actually need to bring in a a third query engine, because now we actually need this universal access layer, so to speak.

Interesting. Okay. So they're saving money because they don't need the Snowflake data warehouse. They don't need to put the data into Snowflake data warehouse before they query it. It's like sitting in s three and then use DocTV to query it.

Yeah. You can you can mostly think of it like that.

Okay.

The reason is because the way that Snowflake bills, obviously, they have a 10 to 20 x premium on the compute that they're charging. And turns out that most of the queries that we're able to support, and today, it is just read only workloads. So you can think, like, business intelligence or, ad hoc analytical queries. Most of these actually don't need to use Snowflake. And, obviously, if you're not querying with Snowflake, you don't have to pay Snowflake.

So if we're able to even route a 100% of these queries to execute on DocDB, then theoretically, the upper limit for us is a 100% savings. Whereas, you know, if you're using some of our our competitors' tools, they're just fine tuning and optimizing within Snowflake. So you really can't

Mhmm.

Optimize past that base floor that you're paying Snowflake. So for us, what's really important is how many queries we can actually execute on Duck TV, and we're trying to optimize for as many as possible.

Okay. Interesting. And so do they need to know how it works, or can they just come in and use Graybeam exactly like they were using Snowflake, and then they just it's cheaper, and they they don't need to know.

Yeah. We we try our best to make the onboarding as frictionless as possible. So we're actually publishing our first customer success story. And, effectively, the only thing you have to do is just change the connection string. All you have to do is go into your settings.

If you're using Looker or Tableau Metabase or any of these other tools, just change that from to from Snowflake to Graybeam, and the and then you're done. You don't have to understand how to write DuckDB flavored syntax. You don't have to even know what's happening behind the scenes. But suddenly your queries are just as fast, if not faster, and you're saving 90%.

Wow. That's very cool. And that that does bring me to your, your h one messaging. I think it's really good. I was talking to you about this before.

But if people go to the graybeam.ai website, the h one is just real snowflake BI queries to DuckDB and save 86%. And then the subtext is Graybeam provides fully managed DugDB clusters, easily plug in your existing Snowflake work flow workloads. No migration required. I I think it's like a really like one of the simplest messages that I've seen. Yeah.

Was it always simple, or did she

It was always simple. I think, like, how we approach that is just if I'm someone who has this problem, the moment I land on the page, you have to just read it and say, oh, this solves my problem. So we went through a number of different iterations, and we had to keep the h you know, the hero messaging as simple as and concise as possible. Previously, it was like, route your Snowflake queries to DuckDB and save 70%. Although it wasn't specific enough, it was it was too broad.

So we started getting people coming in and booking demos who didn't really have the pain of Snowflake cost on the BI side. Today, we only support these read only workloads, as I've mentioned before, and we're building up to support transformations. And that means, like, being able to write data back into the warehouse. And we don't support that just yet. So I wanted to make the updated messaging as specific as possible to the problems that we do solve, which are these BI queries.

So, you know, for customers who have products built as embedded analytics or maybe they use BI as some kind of operational workflow, I I want them to just click, which is why, you know, the thing is is how you see it now. And we did update it from 70% to 86 because this is now our new average.

Okay. Nice. That's great. That's a it's a big number. Yeah.

And so who are the ideal customers then? It's like BI analysts or it's Yeah.

The ideal customers that we have are of course, we're still trying to figure out, but the typically, where we're seeing success is one of two profiles. The first would be, you know, if your company actually serves data as some kind of product. So to give you an example, one of our earlier customers is a cannabis analytics platform. So they aggregate data a cannabis analytics platform. Okay.

So they aggregate data from the point of sale systems and, you cannabis producers and all that, dump it into Snowflake. They shape it and model it, and then they serve that as dashboards to, you know, all the brick and mortar stores across The US and Canada. So you can imagine they have so many users just interacting with their their Snowflake data through these dashboards. So that's a very good kind of IC for us because, you know, Snowflake at that point becomes a cost of goods sold for them. The other flavor that we have is if you're using data as some kind of operational tool.

So where I usually work at Rapple, for example, the ops team would be using our dashboards as, like, their ticketing mechanism to understand, you know, when new tickets come in, when to solve it, and then they would go in to do more digging through these dashboards. So those are the the be the other half that also serve as a good customer, and you usually only see those in larger startups or maybe early enterprise.

Okay. Yeah. That makes sense. So I guess it's just that it's like a key part of the business Yeah. That's right.

Rather than yeah. And do you typically reach out to is there, like, someone responsible that's like the like, if you see this job role, or is it it's just more about the company at this point?

It's a little little bit of both. I would say for the former where it's you know, they're serving data as a product. Usually, the CTO is pretty good or some kind of, like, senior leadership in engineering. And then for the latter where they're actually using data internally instead of serving it to to customers, it would be, yeah, VP of data, director of data engineering, some kind of, like, data esque leadership role.

Okay. Super cool. Alright. I just wanted to establish that since now I wanna get on to, like, hearing about all the fun stuff you've been doing. So in my LinkedIn feed and people should just go see Kyle's, LinkedIn.

Like, you wanna tell us you wanna tell us, like, I've seen this, like, microphone you've got. Like, you wanna tell us about all the videos you've been making?

So a lot of the marketing or even content that we publish on LinkedIn, we try to keep it very fun. I think first ever piece of content that we did that was really fun. It actually might have been the ice bath one that we were talking about earlier.

Oh, yeah.

But the earliest one to my memory is our hiring video. And, you know, we just did some very silly random out of the blue hiring video. You know, I'm working, and I'm just this erratic person doing, you know, erratic movements. And I'm like, I'm working, and then the camera's looking at me this way. And I'm like, hello.

I'm Kyle. Very robotic, and, you know, the the video is cut so poorly on purpose. Know? And there's another one where I'm just, like, turned away, and the camera's like, actually, I'm like, you know, we're hiring for, x y z roles, etcetera, etcetera. And, you know, we've kept that kind of non nonsensical content throughout.

And I think it really helps because, you know, you you go on LinkedIn every day, some folks, and you just see the same boring thing over and over again. Right? It's like someone talked about, oh, what I learned about BDB from getting married or just, like, someone's face. And, like, if we're a small startup, you need someone to stop and watch and read your content. And, you know, they only do that if you have some kind of pattern interruption.

And I think for us, keeping it lighthearted, and, nonsensical, it it it's just it brings, like, entertainment to a platform that's just so bland and boring, and, hopefully, we keep it up. We have we have had some other ones. I'm not sure if you saw the whole toilet gag. But

Oh, wait. Remind me. Maybe I did. Well

So our very first, like, call it ad about, you know, what we do to the world was this video of Arsham, you know, about to use the bathroom, and, you know, he goes to number one. He's about to flush the toilet, and I jump over the I jump over the roof of the stall, and I'm like, wait. Stop. You know, don't flush. And he's like, well, like, what are you doing?

And I'm like, hey. You only, you know, did a number one, so there's no need to use a big flush. And he's like, I don't have an option to use a small flush. And I'm like, well, you know, with Graybeam, we give you the option to use a small flush. In parallel, we're talking about as a company where, you know, you should use DuckDB for all of your small queries and, you know, keep Snowflake for the big queries.

So we had that gag going for about a a couple weeks. And, you know, for folks who aren't really familiar with Snowflake, DuckDB, or the databases, it really helped them quickly understand, like, oh, okay. I get it. Yeah. You're right.

No. I don't need to use the heavy flush all the time if I'm just, you know, doing a number one.

That's amazing. That's amazing. And then I I saw you you did a lot of conference, the conference videos. You're doing like these I my so here's actually, maybe this is interesting. This is why I remember from like seeing it was just like you just running around, like, with I think you had, like, a duck was it a duck microphone?

I think that's how

I remember. A duck in a banana with a cup.

And you're just, like, you're just running around, and you just stick the microphone in people's face. And you're like, I can't remember what you asked, but it's like, what is what do you read about, like like, what I what were you all the questions you're asking was like Yeah.

Yeah. The the the so there there there is a few conferences, but I think the one that you recall closely was whether or not people use trailing or leading commas.

Oh, yeah. Yeah. Yeah. That was that that was that.

Yep. Yeah. Surprisingly, a lot of leading commas, which which which I love.

Yeah. I saw that was seemed like a lot to a lot of people there, it seemed quite a contentious point. I guess it's like the kind of classic, like, semicolons and stuff like all these

Yeah. Yeah.

Kinda yeah. Taps.

Yeah. Taps spaces. We did have a did have a lot of different microphones in there and using using the different props, just a way to hide the mic, but also brings that kind of insane comedic relief. And I don't know if you caught all the different mics that we had, but, yeah, we had a duck. We had a mini ping pong paddle.

We had a chair, which I really want to get more in, but that was only in for, like, a one second clip. We had, like, a massive capybara stuffed animal. Yeah. I wanna bring more crazy mics into the into the mix.

And and are these actual mics, or are they just, like like, are they sold as mics? Or are they that you're just, like, taking a stuffed animal and just just inserting a microphone inside?

Yeah. It's why we have we have a stuffed animal and then I I hold the microphone behind like the ear or in a way that you can't see them.

That's the trick. Okay. Okay. But I do think that it's it it does just seem like you are basically just being yourselves on the camera. It doesn't seem like I mean, obviously, you're being a bit ridiculous, but it's it's not that much of a stretch to how you guys are in person.

Yeah. Yeah. No. We do have to hold back a a little bit. You know?

But, yeah. We we just wanna make sure everyone has a good time. You know? Like, building a company is a very long, arduous, and painful journey. And if if there are some ways that we could have fun with it, we we definitely try to.

And I think it pay I think it's it is paying off. I I get a lot of people who recognize us for, oh, you know, I love your content. It's very, really refreshing. Yeah. People talking about our toilet gag a lot and in good and bad ways.

So, like, I'll I'll meet some other folks and like, oh, why'd you go with that toilet stuff? You know, it's it's not good. But, you know, they saw it, and they saw it, and they remembered it. Right? So I'm not sure.

For a while, we were deciding to just drop the whole toilet theme, but now that I think back to it, it's I kinda wanna keep it going because it's so memorable. You know? It it like, the messaging that clicks, and everyone always remembers us for it, you know, for better or for worse. I just can't think of another way to that's as creative as the whole toilet flushing analogy. Analogy.

So, I mean yeah. Yeah.

TBD. What's your process for making this?

It's usually just on the spot. You know, I'll we're we're we're there, and I'm like, hey. What kind of crazy thing can we do to have this video? So there's one where we were at a conference doing another set of interviews, and I forgot to film an introduction for us. So we're in our office, and I was like, okay.

What can I do? That's just ridiculous. And the that one, I just rolled I rolled out from under the desk, you know, just like a I rolled out, and I was like, what's up, everybody? This is Kyle from Graybeam, and we are it's it's just a what's the most ridiculous thing I can do at this moment?

It's great, though. There's there's this guy, Gonto. He was, like, the the he was, like, the VP of marketing of zero, and he's got this phrase that I love. That's like, it's better to be different than be better in marketing.

And I

felt like you guys are like the perfect illustration of that because the stuff you're doing is not expensive presumably. It's just you're just literally pulling out your phone and filming something, get a mic maybe, but like Yeah. Basically zero cost. It's just like your creative. It's different.

Yeah. Yeah. Yeah.

And and this is the kind of mindset we wanna do with all of the content we that we push. And, you know, we're trying to be very intentional with our brand without spending a lot of money. You know, going through branding exercises can be, like, tens of thousands of dollars. And, you know, to give you that example, the first thing that we were super intentional with was with our swag. You know, like, you go to conferences or you get company swag, and it's always just like, okay.

This this is a T shirt. So when we were making our swag, we wanted to make it not just a T shirt. Like, it it should just be a shirt that people would wear on a regular day. So we put a lot of thought into it, and this goes with with everything, you know, with all the marketing. We just did our first customer success story as I was mentioning earlier, and, you know, we don't want it to be a vanilla, oh, we saved 90%.

This is the problem he solved. This is what it was like. We wanted it to be both a memorable case study. So to do that, we had to, like, first film it. And, you know, the the ideas that we had was, okay.

Let's fly to Seattle. Let's go get hot wings, and that's progressively spicier and spicier, and then ask, you know, our customer our interview questions. So he would be like, oh, yeah. I saved and with you know? So, like, we want it to be very crazy and memorable because people don't ever do this in in enterprise.

Yeah. We ended up not doing that because he he doesn't like you new wings. So Oh, nice. Yeah. Our customer is a is a pilot, as a hobby, so he actually has his own private plane.

So we're like, okay. Well, why don't we just fly the plane and do the interview on the plane? So we we did that instead. No. Yeah.

And then to actually publish the blog itself, again, like, I don't wanna do a vanilla case study. It it should add value to people's lives as opposed to just, like, having a badge on our website. Oh, you know, we we we serve this customer. So for the case study itself, it's it's a playbook. You know?

This is the exact situation this customer is in. They are using embedded analytics. They use these tools, and this is where the breakdown of their costs are. These are all the different things he did to reduce his costs, and, also, this is how he's using GreatBeat. So it's, like, something that's actually actionable as opposed to just something very, very vanilla and and boring.

So, yeah, we try to be very intentional with everything, and this is, you know, one of those few examples.

Okay. So here's a I'm curious on this. I think you're gonna have opinion. So why is doing, like, silly things, like, okay in, like, enterprise? Because I I don't know why people I guess, some people probably feel like, oh, I don't wanna I need to be really serious.

I need to what's your Yeah. Take on this?

Yeah. You know, we just haven't gotten there haven't gone there yet. Right? There there might be there might be some wall where being too silly is not gonna help us close deals. But I think at the end of the day, everyone that we're talking to, you know, they're just another person.

Right? They put on, what's the phrase, like, one sock at a time in the morning. And if, like, this level of silliness means we can't win this customer, that's obviously gonna be very bad. So when we hit that wall, we're definitely gonna have to face it then. But I think in this very, very early stage of the business, you know, we need a way to stand out and to get people to remember us and and to to hear us and to see us and to get our content liked so there's more engagement.

So it hits the LinkedIn algorithm. And and being silly helps us get there. But the the question becomes, like, how silly can we be? There's probably some wall in the future, and I think I think we'll have to tone it down at some point.

Well, I think Suburbase and Posthog are now both like multi billion. Well, Suburbase is like 5,000,000,000 valuation. I think Posthog's like at least a billion. And they're both I I I think they're both still very silly in all their marketing. Yeah.

So you probably got still, like, some headroom on selling this.

How deep can I go? Right?

Yeah. Yeah. Yeah.

Oh, Toggle's a really good example. Like, they changed their landing page to just the most outrageous thing. I absolutely love it. Although, I think the one challenging thing with their new landing page is it's just so hard to navigate. I love how different it is.

It's just super memorable.

Yeah. Yeah. Agreed. Agreed. James is just like I just find it fascinating that he just writes, like, ridiculous tweets, like, all day.

And, like, most of them get, like, zero, like, you know, like, two likes or something. And then just, like, the tenth one will get, like, a 100,000 likes or something just like and it's just him just it's just always just ridiculous stuff.

Yeah. Yeah. I I still need to figure out how to use Twitter or X. I don't I don't even have it, kind of.

You don't have it?

Okay.

Yeah. I like, I made it from when I was in, I don't know, high school. I I revived the account. I just don't understand how to use it. Are you supposed to just say things?

Yeah. I don't know. I'd I'd like to see you just experiment now. Just see how you I bet I bet you'll get the hang of that. Yeah.

I don't wanna say the wrong things. You know?

Yeah. Maybe just bring your LinkedIn behavior to Twitter and just see what happens.

I don't know. To your point, I think the the, like, the thought of just being out there, gets a lot of people going. Because, you know, a majority of folks play it too safe. Right? And, you know, saying or posting just bland content is in itself by nature boring.

Right? So if you just go the outrageous route, then you you have a very high success rate of potentially or not higher success rate of of getting traction than just, you know, being a vanilla data infrastructure company.

Yeah. It's it's just fun. I think it's it's it's it's just so fun. I think RevenueCat that this, like, mobile app subscription tool, they which they just, like, completely dominate that market. And I was speaking to the marketing guy, Rick, and he very cool guy.

And he was saying they just like their whole thing is like, they wanna they wanna help mobile developers, app developers make more money, and they wanna have fun. And they'll do like anything that's like kind of marketing that kind of pushes either of those masters. And so they're, like, when they go to when they went to a conference recently, they, like, created, like, a human like, what they called, like, the grabber machines. So it's like you kind of you get strapped in on a harness and then you kind of like go and grab like it's like a human human claw kind of thing. Yeah.

And just all ridiculous stuff. Yeah. It's yeah. Love to love to see this stuff, and very few people do Yeah.

So the the next Snowflake conference, you know, if you ever get a booth there back back in the day when we were still running with the toilet gag, I was like, we should literally just have a toilet stall in the booth, and there would be, like, a small flush and a big flush. Yeah. You know, if if if we ever get there, it's something that's still on on the wall. And it would be like, oh, you know, you you run your single queries. Like, you would, like, go to the small flush with a big flush, and, you know, how much money are you wasting or how much water are you saving?

Right? It'll be a cool gag. Well Yeah.

That'll be great. Yeah. Yeah. You can send pictures of Asham sitting on it. So fully in a in a decent way, of course.

Yeah. Yeah. And if he if it's the big flush, it shoots water up. You don't wanna use it.

True. You win a prize or you get hit by water, maybe. Yeah. Yeah. Anyway, this is awesome.

Okay. So I'll hear you one more question. Since you started, since since we were hanging out in OnDeck shout out OnDeck, by the way. Great program. Everyone should I should realize I should say that.

It's like, it's it's really cool.

Mhmm.

What is is there anything that you believe now that you kind of know now or you would say to friends now that you wouldn't have thought back then about startups?

I would say that or having a successful fundraise is not does not really describe how successful of a company or, like, a person you are, I would say. And I feel like back then, you know, everyone's goal was to build some kind of prototype to raise money. And if they don't, then, you know, they're a failure. I would say fundraising is just how well you can tell a story. You know, when we when we raised our first fund of capital earlier this year, basically, didn't have like, we were in the same position as when I first met you in Nondek, except we just had a a more compelling story, and someone believed us.

And I would say if you're going out to fundraise, you have to talk to hundreds of of investors because it really is a numbers game. You have to you have to find those one to five people who resonate with your story because half of them aren't. You know? All these investors, they don't really know better. They're just trying to believe and see and and mitigate the risk of, hey.

Is this a a person that I should invest money in and, like, can they solve problems? And, like, how you're able to articulate your answers to some of their questions and then also paint a compelling story of your journey, why you're the person to do it, and how this company scales is really just storytelling. So, yeah, if I were to tell myself back then, like, yeah, fundraising is really not all that that it's, you know, jazzed up to be.

That's awesome. And where can people learn more about you and about Graybeam?

They can go to our landing page, graybeam.ai. That's spelled g r e y, And it's beam, like light beam, not the the food bean. My dad was like, why are you guys named after, like, food? You guys are doing data. And I'm like, it's great beam.

Dad.

Yeah. Exactly. Otherwise, yeah, can find us on LinkedIn. Feel free to reach out. We're always happy to chat, you know, whether you're in the data space or you're also just an early stage founder who who needs someone

else

to bounce ideas off of. We're always happy to help. And that'll be it. I guess, yeah, I gotta start my Twitter now, so you might be able to find me there at some point.

Hell yeah.

And then that's it. Yeah. LinkedIn, website, Twitter.

Awesome. Well, thank you so much, Carl, and, thanks everyone for listening.

Thanks, Jack.