The Longest View with Dez Fleming

[Desmond Fleming] (0:00 - 0:21)
Well, Cyrus, thanks for coming on the pod. For sure, yeah, great to be here. We are into episode six.

We're going to talk about a bunch of stuff. Entrepreneurship, haven journey thus far. But maybe, and I actually don't know this about you, but what was your kind of first, do you have a first memory or entrepreneurial instinct?

Totally. Let's start there and we'll work backward.

[Cyrus Shirazi] (0:21 - 0:25)
Yeah, so my dad worked in private equity for ballpark, let's call it like 20, 25 years.

[Desmond Fleming] (0:26 - 0:37)
Quit his job in PE to become a full-time entrepreneur around the age of like 12. And ever since that moment onwards, I knew I was going to build my own company. Yeah, you were like, whatever you're doing, I like it.

[Cyrus Shirazi] (0:37 - 0:48)
100%. I would literally run home from school every day and then ask him how I could help. And he would even let me move paper around, open envelopes and do mundane, boring office work.

[Desmond Fleming] (0:48 - 3:24)
From that point on, I knew I was going to build my own company. From my college dorm room, I started a digital marketing agency, would literally skip class, like make cold calls to small businesses. And yeah, I always knew I was going to build my own shop.

And I thought the best place to go and learn how to build one would be firsthand at an early stage company, fail as much as humanly possible and just learn from every instance of failure. And then basically gain the chops, knowledge, know-how, experience to go and build my own company one day. So what was the first startup you joined?

Yeah, so I joined a company called Intelips right out of college. Actually, basically before I graduated, I graduated during COVID. And so didn't get the second half of my senior year, just jumped right into it.

And we were basically an SMS marketing company, similar to the likes of Klaviyo, Attentive, where if you're an e-commerce brand, you want to capture leads, you want to be able to follow with them across multiple channels, not just email, but also SMS. So those little pop-ups who steal your phone number and then harass you a ton via text. Had a blast doing it, learned an absolute enormous amount. And basically, while I was there, we went from maybe like 30k in revenue to like a few hundred thousand in revenue.

And were you employee number one? No, no, I was employee number seven. I was the second business employee.

And I got to do everything from like SDR work to marketing stuff, to sales, like direct sales, i.e. like being an accounting executive, even design, even though I'm a terrible fucking designer. And yeah, dude, I literally just rolled around in the mud. Yep.

Whatever came at me. Yep. Just ate it all and kept going.

Okay, so from there, what was the next part of your journey? Thereafter, I joined Mainstreet, which was an R&D tax credit company. They were a hot Silicon Valley darling at one point.

Raised like 60 million from SignalFire. And that was a great experience. Got to work with a bunch of amazing people.

We basically, from a GTM standpoint, were solely reliant on paid acquisition. And as everyone knows, paid acquisition is kind of a well to some degree. It tends to dry up after like a long period of time.

And unfortunately, those channels of growth started to dry up. We launched a few new products, failed to execute on those like new products. Before building the products, we went and pre-sold and upsold our customer base, charged them for those products, then weren't able to deliver.

That right there is like terrible customer experience.

[Cyrus Shirazi] (3:25 - 3:25)
Yeah.

[Desmond Fleming] (3:25 - 4:28)
We gave all the money back. Okay, got it. We refunded them.

But regardless, you've, in that moment, lost a ton of trust. Yeah. And...

It's brand dilutive. Exactly, exactly. And ultimately, we launched like four products, four new products.

They all ended up failing. And it was kind of death by like a thousand cuts to some degree. Yeah.

And I would say the biggest observation that I had while I was there is kind of indicative of a mantra going on in the early 2020s within like the venture context, which is to raise a fuck ton of money at a crazy valuation, hire a ton of people, and hope it all goes right. Yeah. Well, man, that's not really how it works.

Yeah. You gotta build a durable business. 100%.

And by the time I left, I think Mainstream was close to 200 employees. Yeah. Six months later, after I left, they fired half the company.

Yeah, so you feel like you saw in that moment, that experience, a lot of things not to do. 100%. But I also learned a lot of things to do, right?

[Cyrus Shirazi] (4:28 - 4:28)
To do, yeah.

[Desmond Fleming] (4:28 - 5:07)
Even with all Mainstreet's shortcomings and pitfalls, they kicked ass, man. Yeah. This is pretty much my time, but we went from like zero to $10 million in revenue in, I want to say, a year.

Yep. And then while I was there, we probably went from somewhere in the realm of like 10 to 20. Yep.

But the biggest, call it dagger to the heart, was all of those failed product launches. Yes. Yeah.

And all the degraded trust. Yeah, and the inability to expand out from the core. And so...

And it was also a non-recurring line of business. Yeah, it's transactional, right? So you can only claim the R&D credit for the first five years of your existence as a startup.

Yeah. So after those five years are up... What do you do?

How does Mainstreet make money?

[Cyrus Shirazi] (5:07 - 5:07)
You know?

[Desmond Fleming] (5:07 - 7:48)
So then from Mainstreet is you went to Meow. Correct. Yeah.

So I joined Meow from Mainstreet, first business hire, joined a crypto company. Yep. And...

And this is 2022-ish timeframe? Correct. Yeah.

Yeah. Correct. This is around January of 2022.

Okay. And joined a crypto company. We got a lot of traction really quickly.

We're at the tail ends of kind of the Zerp bubble. Yep. We're able to go and raise a Series A from Tiger, QBD, and a few other awesome folks.

And then crypto blew up. Yep. And we needed a new line of business.

And we needed to basically go to zero to one again. And luckily, interest rates were rising at that point in time. And we had already proven as a team an ability to launch financial services products and successfully take them to market.

And so we began to build boring old, like regular financial products. Yep. And...

Like business bank account, cards... Starting with like treasury bills. Yeah.

And we executed on that like motherfuckers and absolutely kicked ass. Yep. And got to like hundreds of millions in AUM really quickly.

And then we continued to launch like financial products after I had left. Yep. But yeah, that was a hell of an experience.

Okay. Earned an absolute ton. Did the whole zero to one thing super successfully.

Went through like complete 360 pivots from one world to another. Which I feel like if I think about your story, it's like three different startups, three different tours of duty, plus or minus three or four years in that experience. You got to see kind of what it meant to build a core software product in the kind of marketing automation space.

You got to see what it meant to build a wedge product in kind of the financial services space and see both good and bad of why it's valuable to have recurring revenue. And also with MainStreet kind of figure out again how to like just adapt to a dynamic and evolving environment, but also just figure out a way to survive. How do those things influence how you think about building Haven?

Totally. So one thing I would also be remiss to mention is the way that Meow and MainStreet built their businesses were fundamentally different. So Meow was all hand-to-hand combat, no paid acquisition, no crazy marketing budget on LinkedIn ads and Twitter ads and whatever else.

It was complete guerrilla warfare. Yeah. Well, meaning what?

Like let's get into it because one of the things I care about the podcast is helping people understand the tactics like in real time.

[Cyrus Shirazi] (7:48 - 7:49)
Yeah.

[Desmond Fleming] (7:49 - 8:11)
So I'd be curious what that guerrilla warfare was. Totally. So I'd say one of the biggest levers that we pulled was we really tried to diversify our cap table.

Yeah. And we wanted to align ourselves with as many firms and angels and founders and individuals within the ecosystem who could then help us when it comes to like go to market.

[Cyrus Shirazi] (8:11 - 8:11)
Yeah.

[Desmond Fleming] (8:12 - 8:53)
So we literally made a point of going and raising from let's call them Twitter influencers. Mm-hmm. Sawhill Bloom, Paki McCormick, Austin Reef, etc.

And then they would help amplify our presence across Twitter and social media. That would lead to a ton of inbound. Incentives are aligned.

Exactly. And then, sorry to all the founders who I harassed, but my crazy ass was reaching out to literally every single company that was backed by those guys and all of our other investors. And did you, sorry to get really into the tactics, did you go and ask them like, hey, give me your full portfolio list or did you like...

Sometimes. Sometimes, yeah. More often than not.

Yeah. Yeah. And then we would like also bother them a ton.

[Cyrus Shirazi] (8:53 - 8:53)
Yeah.

[Desmond Fleming] (8:53 - 9:46)
All of our investors for intros to the portfolio. As you should. Yeah.

They probably... Make us earn our keep. Totally.

I mean, if I'm being honest, like if I was on the other side of the table with how we did it and how aggressive we were, I would be annoyed. Yeah. Which is like totally fair.

But never going to change how we did that. Yeah. Haven't changed anything from that approach when it comes to like Haven either.

Yep. And then like when it came to Main Street, man, they grew so fucking fast. Yeah.

Through like paid acquisition. There was a marketing genius named Rohan Vora who after his time at Main Street, I'll go into like what he did at Main Street, but he started DGods and FTs, which did, I don't know, billions in transaction volume in like late 2021 and like early 2022. And he grew like an anonymous like social media presence on like Twitter from like zero to like 300,000 followers.

[Cyrus Shirazi] (9:46 - 9:46)
Yeah.

[Desmond Fleming] (9:46 - 9:54)
Five months, had like a cult-like following where people were literally drinking beer out of shoes at DGods like events.

[Cyrus Shirazi] (9:55 - 9:55)
What is DGods?

[Desmond Fleming] (9:56 - 10:04)
So it was an NFT project. Yeah, yeah, yeah. And there's probably not much utility to it outside of that.

It's like, where's Bored Ape Yacht Club now?

[Cyrus Shirazi] (10:05 - 10:05)
I have no idea.

[Desmond Fleming] (10:05 - 10:32)
Exactly. But long story short is Rohan had this knack, and I actually saw him yesterday after not seeing him for like a few years. Yeah.

Had this knack to basically engineer virality. Sure. If you look back at a timeline of Main Street, when that kid joined the company.

Yeah. Things began to go vertical. Yeah.

The guy literally cracked LinkedIn ads better than anyone else ever has. I'm sure the budget was like somewhere in the realm of like $20,000 a day.

[Cyrus Shirazi] (10:33 - 10:33)
Yeah.

[Desmond Fleming] (10:33 - 10:45)
A lot of money. A lot of fucking money. But he was optimizing their paid acquisition channels like no one has ever done before.

I won't say sells itself.

[Cyrus Shirazi] (10:45 - 10:46)
For sure. But it's free money. Exactly.

[Desmond Fleming] (10:47 - 13:21)
Like people say, oh yeah, I can claim this R&D tax credit. Bang, let's do it. Yeah.

And so their approach was, let's just pour as much gas on the fucking fire as we possibly can. And I'm not going to say hope it works, but... Well, act two didn't work.

Act two didn't work. Yeah. Yeah.

I would say that act one was great. But when it comes to GTM and go to market and growth, you need like a foundation. Yes.

You need a foundation of organic compounding channels where you don't have to torch money. Yes. And then you pour gas on the fire once you have those to accelerate GTM.

Yep. Mainstream kind of took the opposite approach of going all in on paid acquisition from the get-go. Yeah.

And then once that dried up, they didn't really have any other channels and had already kind of tried to sell all their customers on new products. And ultimately, those didn't work out. But long story short is, in my opinion, you want to take a hybrid approach, right?

You don't want to just do the guerrilla warfare, and you don't want to just do the paid acquisition. Yeah. So coming full circle to my company today, Haven, we've taken the approach...

And tell us what Haven is too as well. So we're building a unified financial operating system for startups and small businesses. What that looks like in practice is one platform does everything from accounting, tax, banking, payments, invoicing, bill pay, etc.

And coming full circle, our approach thus far, we're 20 months old, has been pure guerrilla warfare, pure hand-to-hand combat, community building, reaching out to every company in our investor's portfolio, etc, etc. And then we're now in the realm of like $5.5 million in revenue. But a month ago, we began to turn on and begin to experiment with paid channels after laying a great foundation and building a lot of brand equity.

And a month before that, we hired our first salesperson. So it was founder-led sales up until somewhere in the realm of 4.5, 4.7, me and one of my co-founders. One of your co-founders.

Who led all of sales. Then we brought our first sales hire on and have now begun to experiment when it comes to other channels. And so I think when it comes to GTM, you want to come up with as many experiments as possible, test a finite amount of them in parallel, double down on what works, and get rid of everything else.

[Cyrus Shirazi] (13:21 - 13:22)
Mm hmm.

[Desmond Fleming] (13:22 - 14:45)
But after you've laid a foundation of organic compounding customer-led growth. Yeah. Make sure your core product works, is needed, and is loved.

Correct. And then start to experiment. What are you testing?

Because I'm almost getting back to like my like mini operator chief of staff hat. What are you testing as your conversion metric? Like what is it?

Is it when you think about your marketing channels, obviously, you can pull things through the funnel. It can be top of funnel, middle of funnel, bottom of the funnel. Are you optimizing for bottom of the funnel?

Are you just optimizing to get people into the top of funnel? How do you think about those experiments you're running? Yeah, purely top of funnel.

I still close deals, less so. My co-founder still closes deals. And then Tyler, our first GTM hire, closes a ton of deals.

But you're running experiments to where, let's say you go from closing 10 clients or meeting 10 clients in a month. How do I get that to 100 or 1000 clients in a month? Yeah, exactly.

And so everything we do when it comes to paid acquisition is predicated around filling our calendars. Yeah. And then putting the onus on us to build that great relationship.

Yeah. Like our long-term vision for Haven is to become the family office for businesses. Yeah.

So think of the one firm that every business works with that handles all the shit that they don't want to think about. Yeah. And so...

Do you do bookkeeping too? Yeah, yeah.

[Cyrus Shirazi] (14:45 - 14:45)
Okay.

[Desmond Fleming] (14:45 - 15:20)
The platform today is like bookkeeping, taxes, tax credits. Yeah. It's been a hell of a ride, my man.

We're firing on all cylinders. We've got an awesome team, building a bunch of amazing things. And yeah, we're having a lot of fun.

And there's nothing else I'd rather be doing in the world. I get to work with a bunch of amazing entrepreneurs. And I think the one thing that I do differently at Haven that I previously didn't do in other prior roles, which has worked really, really, really fucking well for us.

[Cyrus Shirazi] (15:20 - 15:20)
Yeah.

[Desmond Fleming] (15:20 - 15:48)
Is I do shit for founders and prospective customers that has nothing to fucking do with Haven. Yeah. Meaning what's an example of that?

Help them fundraise. Yep. Help them with intros to set up their infrastructure across their full stack.

Yep. Help them acquire their early customers. Are you doing that out of the kindness of your heart?

Are you making it known that like, hey, this is not part of what we do? No, no. I do it out of the kindness of my heart.

[Cyrus Shirazi] (15:48 - 15:48)
Yeah.

[Desmond Fleming] (15:48 - 16:16)
I don't expect anything in return. Yeah. But when you build goodwill with people.

Yes. Those relationships transcend the test of time. Yeah, yeah, yeah.

And it all ultimately comes back around. Yeah, yeah, yeah. Yeah.

Other things that we basically think about as a really, really core component of what we do is community. Every founder has gotten to a place, in the majority of instances, they don't really resonate with everyone else that's in their circles anymore that they grew up with.

[Cyrus Shirazi] (16:16 - 16:17)
Like their social groups. Yeah.

[Desmond Fleming] (16:18 - 22:02)
You're going on a lonely and unique journey. 100%. You lose touch with a lot of the people that you grew up with.

You lose touch with a lot of people that you went to college with, et cetera. Yeah. I remember I have a buddy who's a very good friend, shout out Holt.

He's a friend from college. And he's like, look, man. And he's an entrepreneur in the music industry.

He's like, look, man, I've missed countless birthdays, weddings, marriages, bachelor parties. I had to do it. So, yeah.

Yeah. And our approach to that is basically kind of evolved out of serendipity. And I somewhat realized it after the fact, which is I spent a lot of time trying to help and connect people.

Just something I've always loved doing. And it gives me a lot of fulfillment. And I basically got to a place where I was like, I meet all these amazing people.

How do I help provide value to them in a one-to-many form? And so I basically was like, all right, I'm going to create a WhatsApp group and have everyone that I meet that I think is an awesome person in the early stages of company building get added to the group and then give a quick intro to who they are, what they're doing, what they're working on. This WhatsApp group is now almost 500 early-stage VCs and founders.

And once everyone gives that intro, all the other members in the group have context that it's encouraged to reach out to meet, greet, help, connect, etc. So I've gotten selfies from people in all parts of the country hanging out together, helping each other, becoming customers of one another, investing in one another, all other sorts of things. But the agenda of it originally was just how do I help facilitate connection in a one-to-many form?

But then it's turned into this thing where we built a bunch of amazing friends in the New York early-stage startup ecosystem. And then we got this place and me packing to be able to bring them together. And so we will frequently...

The poker nights, right? Have poker nights or happy hours or dinners. We had a chess night the other night.

All sorts of random shit just to bring all these people together and get to interact and hang out with like-minded individuals who are dealing with the same problems, whether it's growth or fundraising or employees or not having a social life or whatever it is, and gives them a place to... It's like like finds like in that environment. Take a step back, put things on pause for a little bit and complain to others who have the same complaints.

Who get it. Have you heard about this idea? Obviously, you don't have a board, right?

No, just me. Yeah. But have you heard about this idea of your own personal board?

Has that ever come up? Like an advisory council to some degree? Yeah, but for you personally, right?

So obviously, there could be an advisory council around Haven as a company, as an entity. But for you, Cyrus, have you ever thought about that? Because it came up for me in a conversation a while ago around the context of also developing mentors, people who've maybe seen that, been there, done that, seen the things you're going through, but also can help open up those doors in a very strategic manner.

For sure. I haven't heard about it in that like specific context, but I've heard about the general like sort of theme. Yeah.

And I definitely have people like that in my corner who I always try to go to before I make like a critical decision for the company to get their advice, input, thoughts, feedback around whether or not this big decision is going to be the right one. Yeah. A few of those people are Harry Raghavan, Chris Quinn, my old man, and a few others.

Yeah, always good to have pops in the corner. 100%. So let's talk a little bit more about Haven.

What is your vision for the company? Obviously, bookkeeping, payments, et cetera, these financial services for SMBs and startups, it's been done before. It's a crowded category.

Yeah. How do you think about separating from that noise early on, and then where do you want to take this company over time? The only reason I built Haven, the business specifically, is because, man, I fucking love working with entrepreneurs.

I believe that having a court side seat to watch them do all this epic shit that they're doing inspires not only myself, but also like my team. Yeah. Without naming names, what's the coolest company that you support?

There's a bunch, man. There's a digital cloning company. Yeah.

Which is doing some epic, super, super cool shit. Yeah. There's some AI CS aging companies, which are doing some really, really cool shit.

There's a lot of others. Yeah. Yeah.

I wouldn't say there's like one that stands above the rest. Yeah. But they're all awesome in their own right.

Yeah. I started it for my love of entrepreneurs and my love of entrepreneurship. And so the end goal is the same goal that Haven started with, to help entrepreneurs as much as we possibly can.

And what I basically envision is, going back to this theme of a family office for your business, if you think about an individual who works with Goldman Sachs private wealth management, you call them up and you say, hey, I want to get my kid into Stanford. They're going to make calls for you and try to make that happen. I ultimately want to be that call for business owners.

I want to be the go-to person to help them solve their problems. And so I think of Haven long-term, not just as a fintech or a financial services company, but as a fixer for business owners. A doer.

[Cyrus Shirazi] (22:02 - 22:03)
Leverage. Exactly.

[Desmond Fleming] (22:04 - 22:45)
And so that can come in the form of accounting, bookkeeping, tax services, banking services, payments, whatever else. At the end of the day, it's tied to one ethos, which I love, around just getting to work with the best of the fucking best, doing amazing things and trying to chart a new frontier in their own right. I don't care if it's a venture-backed startup building fucking time travel or an agency helping businesses acquire customers through paid acquisition.

They're doing their life's work. And being able to help them in any shape or form, it might be cliche or boring or whatever, but I don't know.

[Cyrus Shirazi] (22:46 - 22:47)
That's what we're excited about.

[Desmond Fleming] (22:47 - 23:53)
And how do you think about building that business over time? So one way to say it's like to start, you're building a financial services company, a fintech. But when you start to think about being that doer, being the person who provides leverage for these business owners, that's a services company, right?

That makes me think of wealth managers and RIAs. Totally. But we're at this really unique moment in time where LLMs have emerged into the ecosystem.

100%. And now some people believe like, hey, you can start to productize those services. So is that part of where you want to be going in the future?

How do you think about... Yeah, I mean, it already is, is the short answer. The longer answer there is I am of the belief that as long as humans are on this planet, they want to be able to have other humans they can trust and rely on.

Yeah. We like to connect social beings. Exactly, right?

Otherwise, we're going to end up just living in a simulation. Yeah, what are we doing? Not have any friends.

And that's just not a world I want to live in.

[Cyrus Shirazi] (23:53 - 23:54)
Yeah.

[Desmond Fleming] (23:54 - 25:26)
So our goal is to be the family office for businesses, manage the relationship with the best human touch and the best experience anyone has ever seen, and then automate everything else that we possibly can. So everything from transaction categorization when it comes to accounting, when it comes to tax, auto-populating 1120s, auto-populating 1040s, whatever else. When it comes to sending invoices and getting paid by your customers, helping to automate that process, OCR a contract, then auto-populate all of the fields onto an invoice that are relevant, programmatically send it out, programmatically follow up to make sure that you don't need to have a collections agency.

And then if your payment terms are net 30, net 45, whatever it is, that invoice sits on your balance sheet as AR. Then once it's paid, it zeroes out. People think that this can happen today, but it unfortunately has to happen in a manual form.

I'll give you the funniest example of all time, which I still don't get to this day, which is the biggest accounting software on the planet and probably the biggest payments platform on the planet. Accounting software, QuickBooks. Yeah.

Payments, Stripe. Yep. They don't integrate with each other.

That sounds wild. Like in my head, in my head. Check that.

It's weird. Yeah. Right?

I don't get it. And open banking is great. Yeah.

To help all these platforms integrate with each other, et cetera. But man, oh man, it's not anywhere close to where it should be.

[Cyrus Shirazi] (25:26 - 25:26)
Yeah.

[Desmond Fleming] (25:27 - 27:11)
The data feeds always break. There's erroneous data that flows through. It's not as much of a source of truth as people believe it is.

Correct. So there's a lot of scaffolding you have to build on top and around it in order for these processes to truly work. Exactly.

And you get to this place where, so Parker Conrad talks a lot about atomic units. The atomic unit of Rippling is identity. That is something that's connected to each and everything that they do.

The atomic unit of Haven is financials. Every other component of the financial stack flows back to your general ledger. And unfortunately, open banking, integrations, et cetera, aren't truly at a place today where you can, in real time, handle all of your accounting, monitor all of your spend, see what your aging AR is in one unified holistic platform.

There's a lot of pieces of information that live within the financials, like the general ledger of a business that you also need to know about when you're looking at your revenue, when you're looking at your expenses, et cetera. Because they all connect. Yep.

Right? And if you are using bill.com to pay your bills, or you're using even RAMP, for example, you get a bill from someone, you don't have to pay it for 30 days, 45 days, whatever it is. You have to wait until the end of the month to go and sync RAMP with your GL.

Yeah. It's not continuous? It's not continuous.

Okay. Yeah. We built a direct integration with RAMP, so it can be continuous.

Yeah. But let's say you forget about that fucking invoice.

[Cyrus Shirazi] (27:11 - 27:11)
Yeah.

[Desmond Fleming] (27:11 - 27:26)
That you got from a vendor. You're not going to pay it. They got an unhappy vendor.

Let's say you're trying to manage and project out cash flow. Yep. You forget that the invoice is in RAMP, it needs to be paid in 30 days, 45 days.

Yep. But when you're looking at your financials. It doesn't show up.

It doesn't show up.

[Cyrus Shirazi] (27:26 - 27:26)
Yeah.

[Desmond Fleming] (27:26 - 30:43)
All the systems are disconnected, and also I think about it as on different timelines, right? They're not speaking to one another, and they also don't have embedded directly into each system. Some systems might have it, but they don't necessarily have the kind of automations or the business logic of, hey, if X, then do Y, right?

If three days before invoice needs to be paid and still unpaid, do Y, right? Exactly. Yeah.

Exactly. We unfortunately live in this context where everyone's competitive, right? So they all have their own competing agendas, which leads to things like QuickBooks and Stripe not integrating.

Yes. Because Stripe is of the mantra and all the power to them. They're a great fucking company.

Everyone's got to integrate with us. QuickBooks, same thing. Everyone's got to integrate with us.

When you got two people like that, it's like, well, the customer is the one who gets fucked over and forgotten about. And there's just so many disparate platforms that exist when it comes to managing your financial operations. And long-term, that's basically one of the things that we want to solve for.

Yeah, you're building out this horizontal layer. Exactly. Right, to simplify everything underneath.

But presumably, you're not creating a payback from scratch. You're sitting on top of... Talk to me about how you think about building product.

And I think about, again, when you're building that horizontal kind of platform, there's some things that you have built, and there's some things that you have bought. How do you go about that process and decision-making? And a lot of that question comes from the framing of like, if you're a first-time founder, that's actually probably one of the harder things to decide on.

For sure. Right? Because you can say, oh, I'm supposed to...

I want to have control and full understanding and all this and that. But like, hey, if... I don't know if Rippling has a white-labeled offering or not.

But like, don't reinvent the wheel with payroll. Just use Rippling or sit on top of Rippling. Yeah.

So our approach is long-term, build everything we can. With one caveat, which is do everything we can to make the customer's life as easy as possible. Don't get into the situation of Stripe and QuickBooks where we're not going to integrate with each other because we're these two big behemoths and then forget about the customer.

Our approach is basically, we'll integrate with everything. We'll be the ones to build the integrations to consolidate as much of your financial data as we possibly can, while also having the bells and whistles, which are home-built from scratch, to give you even more efficiency and give you even more autonomy and insight and like alpha into your own financial operations. The Build vs.

Buy discussion is probably one of the most interesting ones out there. Revolut, for example, built their own payroll system to pay all of their employees. They don't use Rippling or ADP or whatever.

You ask a normal founder or an investor about that, they're going to be like, you're fucking nuts. Right? Most investors would be like, please, God, no, don't build that out.

[Cyrus Shirazi] (30:43 - 30:44)
Literally.

[Desmond Fleming] (30:44 - 30:49)
Focus, focus, focus. And I think you want to take like a blended approach.

[Cyrus Shirazi] (30:49 - 30:49)
Yeah.

[Desmond Fleming] (30:49 - 34:01)
Where you want to partner and you want to partner for the long term, but you also want to build things that drastically improve the customer experience. And the experience I kind of articulated when it comes to bill pay, when it comes to invoicing, when it comes to giving businesses real-time, like a real-time understanding of their financials. I'm not talking like a first day after the month or the second day or the third day or the fifth day or the 10th or the 15th.

Are you talking next day or real time? Like daily. Yeah.

Right. If you want to be able to provide that sort of like insight, you need everything fully unified. Yeah.

Because if you think about banks, they are required to reconcile daily. But all of these different financial technology platforms that sit on top of banks, while they reconcile against the bank core and the bank geo daily, they're not providing that information or that reconciliation to the end customer. It's a lot of fucking work to do it.

And it's not as relevant to their businesses because they're not trying to build like a fully unified sort of experience. Yep. But in order to be able to do that, because the other players don't do what I'm talking about, you have to build it all in one.

Yep. And the end goal is to be able to say, say, hey, while you're using all these different, like disparate platforms, sure, we're connecting them for you. You're still not getting your financials until like the 5th, the 10th, the 15th.

Yeah, because it still sits elsewhere. Exactly. Yeah.

Yeah. Yeah. What would you say is the culture that you're building at Haven?

A culture of winning, period. I think that we want to build a team of people who are the oddballs, are the misfits, are the ones who are overlooked, the ones who everyone forgot about. And I think those types of people, for better or worse, and I definitely embody this, live their life in an ever-increasing state of never being satisfied.

It could be because they were wronged as a kid. Because they were cut from the basketball team. Whatever fucking reason.

But I think that our approach to company building is predicated around finding undiscovered talent that all has a chip on their shoulder. Yeah. What are your tactics for finding undiscovered talent?

Right? Like the M&T program at Penn. Not undiscovered.

Very, very well known. Totally. But what are your tactics for finding people?

Yeah. I mean, recruiting is fucking tough, dude. But the game of recruiting is pretty honestly similar to the game of GTM and sales.

Gotta stay on top of people. Gotta stay on top of people. Gotta get into every nook and cranny, every ecosystem.

Doesn't matter if they went to Southern Connecticut State University or if they went to Stanford. The best people come from all different types of backgrounds.

[Cyrus Shirazi] (34:01 - 34:01)
Yeah.

[Desmond Fleming] (34:01 - 34:36)
And they live all over the fucking world. So our approach is, one, don't limit ourselves based on geography. So you're distributed.

For sure, yeah. The core founding team is in New York, like me and my co-founders. Don't think about resumes on paper.

If you were selling milk to convenience stores before, but you have the gift of gab, man, I'll turn you into a fucking monster. Yeah. Hard job.

Hard job, dude.

[Cyrus Shirazi] (34:36 - 34:37)
Hard job.

[Desmond Fleming] (34:37 - 35:10)
Selling milk to convenience stores, hard job. And I think my approach to talent is predicated around potential and hunger. And some of the ways that I gauge for this stuff are asking like crazy fucking questions.

And so one of the questions I asked, I've asked a bunch of the prospective people who we've hired at Haven is, do you want to walk on water? And I don't give a shit about what their answer is.

[Cyrus Shirazi] (35:11 - 35:11)
Yeah.

[Desmond Fleming] (35:11 - 41:03)
That doesn't matter. The reason I ask the question is, I want to know if you're weird enough, crazy enough to be able to answer that question with a straight face. Engage in it in a thoughtful way.

Correct. Whether it's someone saying, no, that's crazy, I would never want to walk on water. Why would you ever want to do that?

That could be a positive for you. Yeah, absolutely. And here's why.

Yeah, part of the art of sales is just... Because at the end of the day, life, professional work, business, whatever it is, they're constantly going to get punched in the face, get curveballs thrown at you, shit's going to go wrong. And you need to know how to adapt.

And you can't always throughout an interview process see how people are going to respond to failure. But what you can do is basically ask them things which are completely out of left field and observe how they handle that. Yeah.

Again, part of life, right? You're going to go into some customer meeting and they're going to say, oh, by the way, I need to have this done by tomorrow in order for your RFP to be considered. Yeah.

Okay. That blows up the next 12 hours. 100%.

Yeah. So from a product perspective, correct me if I'm wrong, you, like for payments, you partner with someone, like you sit on top of Stripe? No.

So like we haven't launched like all that stuff yet. Like our main product just went live. Okay.

Okay. Got it. Got it.

It's not even officially live. It's like Haven opened its own Haven account. Yeah.

But then like if I think about the core thing, like bookkeeping, like ledger, you've got some sort of ledger built out today. So our approach, we have our own internal ledgering system. Yes.

But our approach is to integrate with all the different platforms where people are currently operating. Yeah. And running like their financial operations.

Yep. Why? Because people fucking hate to switch.

And then going back to my ethos of, we don't, we want to prioritize the customer above all else. We want to make sure that their experience is like the best you've ever fucking had. We don't want to be the big player or the player is going to take the stance of, oh, we're not going to integrate with people.

People need to integrate with us. Because at the end of the day, the only person who's getting hurt by that is the customer. What is your customer base today?

Is it mostly like coastal startups or is people all over? There's people all over. I would say mostly coastal.

But our customer base is 70% venture backed. And then about 30% are SMBs. And when I say SMBs, I mean agencies, recruiters, etc.

Yeah. Yeah. What do you think is going to happen over the next, you know, let's call it relatively tight timeline, next two years within the startup ecosystem?

If you just take a step back and think about your best effort guess like, hey, are we going to have 20 more versions of Cursor? You know, do you, yeah, I would love to get your opinion. Totally.

One thing I believe is that AI companies, LLMs, OpenAI, Anthropic, etc. are going to wipe out a lot of software companies. And I think we're going to begin to see a lot more forward deployed style consulting firms.

Yeah. Which come in and help businesses use AI in various different forms to then solve various different problems for businesses. Similar to like Palantir.

And once they've built that shit, they're going to then replicate it and sell to someone else. I met someone recently who is helping private equity firms and family offices source and buy cash flowing businesses. And his approach is exactly that.

So he built software for the private equity company to source deals. So they build software, which the private equity companies then use to source deals and have another offering where they come in and build proprietary things, which help those firms do a variety of different things when it comes to diligence, Yeah, managing their internal stuff. Yeah.

So not just sourcing, right? But helping with the sourcing process, like more broadly on like a future cadence, like evaluating markets, etc. Is it like an ex-Palantir team that's doing that?

No. Okay. They will then develop all different forms of like technology.

Which they can go and resell and sell to other firms that want to do similar things. And so the reason I believe that this is going to become pretty commonplace is we're seeing an enormous amount of roll-ups. I'd say more than we probably have in like prior years.

Yeah. I mean, for sure in the venture community, as far as I understand, that was not a thing prior to the past two years. A lot of Gil is doing it.

Obviously. Look at everyone becoming an RIA. Yeah.

Lightspeed. Lightspeed. A16Z, whoever else.

General Catalyst. GC, etc. Yeah.

Thrive, whatever else. Yep. Those businesses, those roll-ups, the big firms are going to tell you that they're all AI native and building workflows and automation to drive margin increase and efficiency.

But I wouldn't say I've heard anything to that end. Actually happening. And so they're going to need partners and firms to come in and help drive that margin efficiency.

My understanding, the playbook that's being run right now, it's like step one. People are finding, first, let me buy the thing first.

[Cyrus Shirazi] (41:03 - 41:04)
For sure.

[Desmond Fleming] (41:04 - 41:12)
And then once I have control and in the guts, I will then start to drive the margin efficiency. And I think the intent is to build the software.

[Cyrus Shirazi] (41:12 - 41:13)
Definitely. That's the intent.

[Desmond Fleming] (41:13 - 41:42)
But these guys are all investment bankers. Yeah. So what margin efficiency are they going to be able to drive?

Well, I don't know. Investment bankers can code now. For sure.

I mean, yeah. My dad was an IB and he went to college and studied CS. Yeah.

So speaking out of the two ends of my mouth. I do believe that we basically will see these firms need different sorts of partners to help them drive this margin efficiency. Makes total sense.

[Cyrus Shirazi] (41:43 - 41:43)
Yeah.

[Desmond Fleming] (41:43 - 43:18)
That's just like one random idea. A lot of people talk about how OpenAI, Anthropic, Gemini are going to kill a lot of businesses. Yeah.

They were just, I think it was might've been OpenAI, Rate Limited, Granola, I think, or released, it wasn't Rate Limited. They released their own version of the note-taking app. Really?

Which is like Granola's core. It is Granola's business right now. Yeah.

So I think that businesses encounter very unique and specific problems that their customers are facing and are able to build very unique and bespoke and custom solutions to solve those problems. Yeah. I'm somewhat skeptical of a big company like OpenAI's ability to go and launch competitors, which will have feature parity compared to all of these different tools that are some degree like wrappers.

Totally. I don't know what exactly will happen to those businesses, but my gut tells me OpenAI's probably just going to buy them. Yes.

Because they still want to own the end customer and those customers have done, whether it's through integrations or specific workflow, they've done enough to differentiate and it's not worth it for OpenAI or Anthropic or X or any of these guys in the interim to rebuild all of that. Exactly. But they still want to own that customer.

[Cyrus Shirazi] (43:18 - 43:18)
Exactly.

[Desmond Fleming] (43:18 - 43:44)
Yeah. And to train any model, you need proprietary data. Yeah.

And to solve these custom and bespoke point problems and build these point solutions, the only way that they're able to do that is because they have all this custom and proprietary data on the problem that their customers are facing. Yeah. So I think when it comes to all these massive LLM companies, we're going to see a continued trend of a fuck ton of acquisitions.

[Cyrus Shirazi] (43:45 - 43:45)
Yeah.

[Desmond Fleming] (43:45 - 45:45)
And I think that is the way that they will become the everything app. And I think that... How do you use the models today?

Do you use Chats GPT? Do you use Anthropic? Predominantly OpenAI.

Okay. We use it when it comes to transaction categorization. We use it when it comes to parsing documents to be able to extract relevant piece of information via OCR to then...

You don't use a reductor or a heron or anything in the middle? We do not. I would say that a lot of our approach thus far to date has been to solve the problem to the degree where we are in the 80-20 rule, where we'll solve the 80%, which is 20% of the work, which is less of the work, and then we'll leave the other 20%.

Because when we build these other components of the stack, that 20% and the work required to be able to successfully execute on all of that will become far easier. Right? So one problem we encounter today is when a customer uses Build.com or RAMP for invoicing, we need to go and get that invoice to then create a schedule if it's prepaid, adjust a service period if it needs to be accrued, etc. We're not natively able to get that data from RAMP because the customer put it into RAMP, then it needs to come back into our system. Yeah. But once we have everything running through our system, we are now able to solve that problem in a far easier form.

Where instead of needing to solely rely on trying to figure out some weird way to get all of this data from RAMP, we have it all natively within our operating system.

[Cyrus Shirazi] (45:45 - 45:46)
Yeah.

[Desmond Fleming] (45:46 - 46:56)
Going back to how we're going to be able to close books for businesses way fucking faster when it's all running through our system. But today you've got like the integration. We have an integration.

That pulls the information. But we don't get that data. Yeah.

Yeah. Got it. What is your advice to the version of you from two years ago?

What are the top three lessons you've learned thus far in building Haven? And obviously going from zero to five in 20 months in a really challenging space where there's a lot of incumbencies is very impressive. I would say fail more.

So put myself in positions to learn. I personally, throughout my career, I've probably learned the most from my failures. Yeah.

I've had a lot of them. Yeah. Care to share?

Yeah, for sure. When it comes to basic examples, sales. Yeah.

I've lost a ton of deals. Definitely closed a bunch. Yeah.

But I would say I've become a better salesperson not because of the deals that I've closed, but because all the deals that I've lost.

[Cyrus Shirazi] (46:56 - 46:56)
Yeah.

[Desmond Fleming] (46:56 - 47:04)
I'll give you another example, which is you got to really build for your fucking ICP.

[Cyrus Shirazi] (47:04 - 47:04)
Yeah.

[Desmond Fleming] (47:04 - 48:20)
And I made a mistake last summer and fall at Haven where I tried to go and build a sales tax product. But not all of our customers have sales tax needs. Not every business has a sales tax obligation.

Yeah. It's dependent upon the product that you sell, the markets that you're in, etc. And I lost, for a second, my eye on what the customer really, really needs and what our true atomic unit was.

I was like, let's go build an all-encompassing tax platform to help solve all of accounting, tax, etc. for businesses. But the thing that I learned was that's not going to drive any efficiency for the other parts of our platform more broadly.

Mm-hmm. And, well, yes, we're going to consolidate some of the financial operations for a business by having this kind of all-in-one when it comes to tax. It's not as hair on fire of a problem as being able to deliver financials on a daily cadence for businesses.

If you're long-term, we want to serve the 30 million SMBs in America.

[Cyrus Shirazi] (48:20 - 48:20)
Mm-hmm.

[Desmond Fleming] (48:21 - 49:16)
Every fucking SMB has cash flow issues. Yep. Period.

And the only way to help them navigate those issues is give them more real-time visibility into their financials. And the only way to actually be able to do that is to unify the stack and build something where you don't have to be reliant on all of these different third-party tools and then bring them all together into the GL at the end of the day. And so I believe in that moment, I made the mistake of listening to the customers who were asking for a solution for sales tax but not listening to what problem does our entire customer base face.

There's something readily right in front of you that you could execute against that would make more money for Haven, which is good.

[Cyrus Shirazi] (49:16 - 49:16)
Mm-hmm.

[Desmond Fleming] (49:16 - 50:11)
And then there was an abstract problem that you need to keep the main thing at all times. And there's an opportunity cost or time spent on either of those things. Totally, totally.

And so in the interim, as we go... And when you do invoicing, you have to be able to help with sales tax, right? Mm-hmm.

Because you need to, on the invoice, add sales tax as a line item so that people can charge their customers for it and then go over and make it to the relevant states that they have customers in. So going back to the build versus buy decision, in the interim, our decision will be to partner in that context. But the customer is always right.

Always listen to customer feedback. But when you get it, take a step back to think about, does this apply to every single one of my customers?

[Cyrus Shirazi] (50:12 - 50:12)
Yeah.

[Desmond Fleming] (50:12 - 50:59)
And if you're going to go horizontal, you better be building fucking products that solve problems for your entire customer base. Yeah. Not just like one subset of the customer base.

Totally. I love it. What's something that our growing community can do?

How can we help you? I don't have any asks. I mean, always looking to meet amazing entrepreneurs and help them however I can.

Fundraising, introductions to prospective early customers. If you want community and you want to meet other founders in New York, hit us up and we're happy to add you to our community. When's the next Poker Night?

I don't know. Let's get it on the book. Let's make it happen.

Yeah. I might be traveling for a little bit and my co-founders might be, but we'll make it happen. We'll make it happen.

Cyrus, thanks for coming on the pod. Thank you. Thanks for having me.

Yeah, of course.