Credit-led Growth — Fireside Chat with Manny Medina & Doug Landis

Julia Nimchinski:
Yeah, two more B2B legends, welcome to the show, Manny Medina and Doug Lendis. Doug, you need no introduction, many you as well, but briefly, Doug Lendis is the CRO and co-founder of StoryPath, and Manny Medina is the co-founder and CEO of Paid, and former CEO and founder of Outreach. What’s in your agenda Coas? And, the stage is yours. Many, Doug, what are you using?
Obviously, StoryPath, obviously paid. What else?

Doug Landis:
Oh, well, before we actually… before we get there, Manny, what’s up, man? Dude, so good to see you. We’re.

Manny Medina:
Good to see you, man, it’s been such a long time.

Doug Landis:
Totally. So, Julia, Manny and I are just gonna geek out here for a minute, and then we’re gonna get into some fun discussions. You know, it’s interesting, I don’t know if a whole lot of people know this. I’m gonna go back in time here for you. Yes.

Manny Medina:
Do it, do it.

Doug Landis:
My mom is calling me right in the middle of a webinar. I love that. Thanks. Hi, Mom, she’s listening. You know, I don’t know if there’s a whole lot of people know this, but before outreach. Your company, Group Talent, 2011.

Manny Medina:
Like… Oh my god, those were miles.

Doug Landis:
Totally! Talent marketplace for connecting employers with job seekers, or software engineers, I think that was the focus. Yeah. And you guys almost went under in 2013, I think that’s right, and then you pivoted.

Manny Medina:
That’s right.

Doug Landis:
outreach. So effing cool, and… and then you decided to do it all over again, because it actually worked?

Manny Medina:
Are we mad?

Doug Landis:
Mad? Are we totally mad? By the way, the cool thing is, like, the pivot was, you would actually build… and this kind of reminds me of, like, the Slack story. You would have actually built some technology or capabilities that everybody really loved, the ability to organize and send out automated emails, and that then became outreach. Is that right?

Manny Medina:
Yep. Yeah, yeah, yeah, so it was two pieces. It was, it was personalization before it was cool, and it was actually done by people. Like, people are actually, you know, we had a little marketplace of writers who would come and write, like, personalized subject lines and opening sentences. And you know what’s funny?
Like, our highest performers were, like, this, you know, this, this, this bargaining comedians, like, people who actually did, like, a stand-up comedy, who would, who wrote for us in the morning, you know?

Doug Landis:
By the way.

Manny Medina:
I kid you not, and they had, like, the highest performance. So it was that and follow-up. See, the automated, so the personalization and the follow-up generated 40% replies, and that was unheard of back then. So that was, that was the, you know, and we were, like, swimming in meetings, but not making any money, so that’s… that’s what we ended up doing.

Doug Landis:
Oh, that’s so great, I love it. And of course, we were duking it out, because I was on the team sales loft back then, you know, with Kyle and all, largely because Emergence Capital was an investor, and I worked at Emergence. Such a… such fun history and such fun memories. And now here we are again. You know, kind of rebuilding or building again in this new world, new age.
And you know what’s really interesting is kind of tee up this conversation, and I’m so excited, because, dude, you’re, like, you’re such a badass. By the way, most people probably don’t know this, you were, like, the third employee at AWS, is that right? Like, AWS…

Manny Medina:
Fourth or 3rd, yeah, something very, you know, it was very small. Yeah, it was, like, it was barely an idea. It was sort of one of those, like, Jeff musings. I was like, what if… what if we charge compute by the drink? Is that even possible? Like, team, go find that out. So anyway…

Doug Landis:

All of this to me says, like, you’re the right person to bet on, because you seem to have this way of thinking about or being on the precipice of kind of a new way, whether it’s a new business model or a new way of actually engaging customers are helping customers transform their own businesses, and it’s, and I think it’s kind of apropos for this conversation, because I feel like we’re kind of in this massive fundamental rewrite of enterprise software, really, if you think about it.
Yep. For a long time, we built systems that people logged into, and now we’re building systems that can actually act on our behalf. Like, they can act for us. You know, everybody’s thinking about cloud co-work, and skills, and cloud code, and it’s forcing everything to kind of change. Product design, how we think about product design, product releases, pricing, trust, governance, and value.
And how do we think about value, and how do we measure value? And I think this is a… a, a, a… it… part of the reason why you created Paid is because there’s, like, this, like, okay, we now don’t just have people that need to get paid, we have agents to get paid.
And so I’m curious, like, you made this shift, because, like, Outreach was one of the preemptive companies that was leading the charge in the world of seat-based SaaS, right?

Manny Medina:
Yep.

Doug Landis:
What did you see in the market that made you believe that the model was starting to break in the world of AI?

Manny Medina:
So there is two… two components to the Genesis of bait. Number one was this… the fact that If you think about the seed model. When the moment you write a contract to sell a seat, you are giving your customer innovation up to the point at which they bought the seat. And you don’t give him anything else. You see what I mean? So, and you do it professionally.

Doug Landis:
Don’t they get, don’t they get the additional innovation that you’re releasing in the product if you’re pushing additional releases?

Manny Medina:
Most of the times, they don’t. So, like, remember in our world, you used to lock it so that can just sell add-ons, or you can sell the bigger seats.

Doug Landis:
So true. Yeah, that’s true.

Manny Medina:
So, like, so, you know, you’re locking the seat, you give the, you know, your customer the product as of this moment, and then you don’t get any other innovation, and then you’re, like, struggling to get back into procurement, to get… sell an add-on, and, like, I’m like, the software is moving faster than ever. You see what I mean? And, like, you are.

Doug Landis:
Yeah, totally.

Manny Medina:
landing the seat, you’re, like, you know, you’re robbing your own future, and you’re, like, robbing your customer from the benefit of your own innovation. So, like, that… you know, I’m gonna curse a little bit. That shit just drove me crazy, because it was so… so stupid.

Doug Landis:
Okay, but I’m gonna challenge you on that, because… and this is where it’s gonna get really fun, and we’re gonna kind of go off script.

Manny Medina:
Yeah, that’s.

Doug Landis:
Because we’re… we are seat-based, and we sell licenses, straight up, and we’re pushing new releases every week. Of our product. And huge releases and updates, like, I would argue, every month. And the product just changed and grown dramatically. Our cost basis, though, hasn’t really changed.
Like, we’re so efficient because we are using swarms of agents to help us build, to help us operate our company, because everything that we do and use is AI-first and AI-native, just like I’m sure I know you are as well.
And so, the reality is our customers are actually getting this benefit, so maybe the downside is, meaning we’re not locked, meaning we’re just… we’re releasing it and just letting it go in the wild. I think maybe the downside is more for us, because we’re actually not capturing the financial value of what we’re releasing on an ongoing basis. So maybe that’s kind of the shift.

Manny Medina:
That’s kind of it.

Doug Landis:
towards the… yeah, the shift is more towards.

Manny Medina:
It’s like you’re.

Doug Landis:
Customers are actually getting more value than we’re actually getting paid for today.

Manny Medina:
I mean, that’s exactly it. It’s like, you’re not getting paid for the innovation that you drive. And, like, let’s assume for a minute that you’re hyper-efficient. You’re still consuming a lot of tokens. So, like, right now, in debate, at this moment. You know, a medium to high-performing developer consumes as much in tokens as they consume in their base salary.
So now an individual is costing $400,000. It’s crazy, right? Now you say, oh, but I have to hire fewer of them. Well, you do until you don’t. you know, your ambitions only gets larger, your customer’s appetite for consumption only gets larger, so you’re, like, this only trends in one direction. So your R&D is out of control, and you have no way to monetize at the same time that you’re shipping.
So that’s problem number one, put that aside. The second problem, and this is actually, like, the story, like, the genesis of paid. was… I was having a conversation with Alan Diginson, who is the CEO of DocuSign. And my team is swarming the accounts because we’re doing an expansion, we’re upselling, we’re renewing, doing all the things.
And Alan straight up asked me, like, Manny, how many fewer people do I hire next year to maintain my rate of growth, given that you’re gonna bring all these agents in? And I’m like, I’m staring at a contraction event. You see what I mean?

Doug Landis:
Scarying.

Manny Medina:
Had a flat renewal.

Doug Landis:
You see what I mean? He’s about to lose revenue, he had to He’s like, I don’t want to hire more people. Yeah.

Manny Medina:
Like, the words that are coming out of my mouth are precluding me for expanding this account. And, like, there’s no way to fucking monetize it. I’m like, this is crazy talk. So I went back to my IT team and asked them, like, what do we need to do? To switch it to… to detach the agents from the seat? And, like, the answer is nothing.
You can create a fake SKU and sort of, like, contort it into some kind of pretzel to, like, make sure that, you know, there is a zero, you know, dollar SKU and whatever, but, like, this is stupid. Like, you know, Zora and NetSuite and Salesforce Building were keeping me away from growth. And the model was just broken. So that’s… that’s when I was like, this is… this is nonsense.

Doug Landis:
That’s so interesting. I mean, okay, this… the scary part is this conversation is gonna maybe make us rethink our pricing model, but I’m totally open to it, I’m here for it. Because as I think about this, I’m like, yeah, we’re actually increasing capacity for enterprise sellers, right?
So, like, as a sales team, you don’t necessarily need to go hire that replacement if you just lost somebody, or if they left, or if you had to let them go. Maybe you can do more with less. So let’s think about this. So, as we shift from these This, like, people, you know, something that people use to actually… something that acts on our behalf.
as businesses, how do we have to… what changes need to happen first? Is it… you need to rethink the product, and the product architecture? Is it the actual business model? Is it the operating model? I mean, pricing and packaging is always one of those things that feels like it’s something you figure out after the fact.
You’ve got, like, the value that you can deliver, and are you actually solving a real problem, and… you know, does the product actually work, and how do we think about how we’re… our delivery mechanism, etc? But, like. What’s the order of operations in terms of change, and what have you noticed?

Manny Medina:
Yeah, so, you know, we… because we now do this every day, all day, you know, it sort of like follows a similar recipe, which is… so, first of all, as a software business, you have to work back from what the customer is getting. Always. I mean, the customer bought you for a reason, right?
Like, it’s a job to be done, there is an improvement somewhere, there is a savings somewhere else, and that has to be tightly quantified, you see what I mean? And then tied directly to the agent activity that is driving that result.
So, that’s actually the first step and the hardest step, is that if the software can be tied to the result that you’re driving, or the task, or the action, or the outcome, or whatever that is, then you over… you… you sort of, like, cross… You know, the hardest and the first step. Right.
That’s the… that was the problem with seats, is that seats, sort of like was a… was a dumpster truck of features, right? All in the seat, of which you use probably 20 You see what I mean? And, like, you could not assign a seat to an actual result.
So now, you actually have the luxury of saying, okay, so this is what the agent is doing, because it’s working almost autonomously or semi-autonomously, and this is what you’re getting for it. Now, there will be problems with attribution, et cetera, et cetera, et cetera, but at least you can associate a task That has almost always a human equivalent value for that task. Right?
So if you’re reviewing a document, you’re writing a story, you’re doing some research, you’re generating a meeting, all those things would have consumed human time. So at the very least, you have a legal-to-legal pairing there. And then you have time savings, and then you have cost savings, and then you have cost… You know, then your revenue expansions, things of that nature.
So, like, that’s the first step. It’s like, map what your agent is doing to how the customer perceives value. And this is the thing that is super interesting. It’s not all customers would perceive value the same way.

Doug Landis:
Right. So that’s where you… Yeah, you just… you literally took the words out of my mouth. I’m like, yeah, but you’re assuming that everybody defines value the same way. You’re also… there are also some assumptions baked in there that the task, if you will, is really clear. And also, underneath that is, it’s not just one agent that’s usually assigned to that task.
Like, in our case, we have a swarm of agents that are doing a boatload of tasks across the board for, you know, an enterprise seller, and it’s like, well, how do you assign value to that one task when there’s multiple agents that are actually helping to complete that? So there’s a whole bunch of… you just unlock…

Manny Medina:
There’s a whole bunch of them.

Doug Landis:
I just got…

Manny Medina:
100%, 100%. And so, like, this is the beauty of where we are right now, is that because each customer is consuming your agents in different ways for different reasons, you want to have a framework by which you show to every customer how they’re getting value, and how they’re getting value over time, too.
And this is the funny thing, is that we all think that value needs to be captured in the form of dollars, but they don’t have to be captured in the form of value. It could be captured in the form of time savings. Yeah, acceleration of roadmap, they could be capturing, like, risk avoidance.
Like, we deployed a customer where it said, like, for every 20 exceptions that we find, we prevent, you know, 5 regression risks. Like, you know what I mean? Like, everybody has a heuristic in their head that you can just apply, and just show that your agent is driving that result, or preventing that failure.
And as long as you do that from the customer back to you, I think you’re going to be in good shape.

Doug Landis:
So this is largely what you’re talking about, is this notion of outcome-based pricing. Right? Results-based pricing, however, whatever you want to call it. Where does that… where does that work really, really well, and where does it actually start to get much more messy? Like, you know what I mean?
Like, well, let’s take existing SaaS companies that are, like, bolting on AI capabilities, who already have, you know, they already have their SaaS seat model, and now they’re trying to throw on some, like, consumption pricing on top of that, which really gets crazy, or they’re throwing some outcome pricing on top of that, that gets messy.
Like, where does it make most sense, and where is it maybe, like, stick with seat-based pricing?

Manny Medina:
Yeah, no, that’s a great question. So, there’s a… there’s a matrix, a 2×2, that wasn’t invented by me, it was invented by Maravan Romanujan, and he has this thing called Attribution Autonomy Matrix. So, if you think you have four quadrants, you have high attribution, high autonomy.
you have, you know, low attribution, high autonomy, low autonomy, low attribution, high attribution, and then low attribution, low autonomy. So on the high attribution, high autonomy, where the agent is performing the work, and you can assign over 50% of the value to that work done by the agent itself, then outcomes or outputs are your best bet.

Doug Landis:
Hmm, okay.

Manny Medina:
And as long as you are defining the outcome or the output in a way that the customer agrees that that’s value to them. then that is your best bet, and it’s the best bet for two reasons. One is that it’s really hard to rip you out. This is why companies like Sierra are doing so well, even though the pricing is obscure. These are companies like, you know, Intercom are doing so well.
you know, even with the price that means transpiring, it’s like, the outcome of a customer resolution is pretty defined. So that one is clear.

Doug Landis:
Pretty black.

Manny Medina:
Where it’s pretty black and white, but there is ones that are less clear that are still have… that are still up there. So, for instance, If you’re avoiding a chargeback. That’s all the agent.
If you are, you know, capturing… if you’re able to recoup more cash on a, you know, on a person that wasn’t paying you, and the agent did all the follow-ups and figure out where they were and all that, a collections agent, that’s also very attributable. Many times, when you’re resolving a ticket, you get rerouted to a sales call.
to upsell you something after you, you know, you got your ticket resolved. When the agent does that, it’s also autonomy and attribution high. You see what I mean? So, like, the interesting thing is that we are getting more and more, like, as the agent gets more autonomous and more intelligent, you know, there’s sort of, like, the shifting towards that direction. You see what I mean? So…

Doug Landis:
Yep.

Manny Medina:
as the agent becomes more forward and predictive around what it is that you need, then you’re gonna see more moving in that direction. Let’s talk about the other thing, which is… oh, go ahead, go ahead.

Doug Landis:
Well, I was gonna say… no, finish what you’re gonna say, because I want to bring in, like, things like, how do you price for better judgment? Like, as human beings, like, of all the agent is giving me more insight and more context and, you know, and the ability to actually make a better decision.
I can, you know, I can choose where to prioritize my time with a customer or, you know, what bugs to go fix first, right? So, to me, that’s kind of better judgment, better execution. How do you incorporate that into value and both pricing? How does that fit into the matrix?

Manny Medina:
Yeah, yeah, no, I think that’s a great question. I don’t know that you can, and that’s where… Yeah, no, no, like, I’m just being honest, like, I think that.

Doug Landis:
Yeah, yeah.

Manny Medina:
You know, you’re…

Doug Landis:
I wanted a simple answer!

Manny Medina:
No, there’s no… yeah, so it’s actually… it’s funny, so let me share a story with you. So, you know, we as humans, are used to paying for effort, so there’s something called the effort fallacy.

Doug Landis:
Yes, yes.

Manny Medina:
If somebody shows up at your house, and it’s like, let’s say you come back from a party, it’s midnight, and you’re… somehow your key doesn’t work, and you need a locksmith to come in and let you in. Locksmith comes in, 15 minutes you’re in, and it boosts you a $500 bill in front of you. You’re gonna be like, what the hell?
Like, it’s just 15 minutes of your time, and $500 for 15 minutes is… Whereas, if a locksmith comes into your house at midnight, and goes back and forth to the shop to find new tools, and it takes in 3 hours to do the work, and puts a $500 bill in front of you, you’re willing to pay that bill, even though it’s a worse experience for you.

Doug Landis:
Yep.

Manny Medina:
So we’re putting agents and traumatizing them, and actually doing the same thing, when in reality, you don’t want to spend any time with the agent. You see what I mean? You want to spend more time where you’re supposed to be spending time, not sort of, you know. Going around the app and, like, clicking stuff.

Doug Landis:
So… Yeah.

Manny Medina:
you know, unfortunately, or fortunately, paying for effort works right now. So if you see the agent consuming a lot of tokens, if you see the agent doing really complex stuff, if you see the agent, like, really spending time redoing your calendar so that you can focus on the right things, you’re happily willing to pay for that. That’s just the way the market clears. You see what I mean?
So you can go ahead and charge for it, and eventually we’ll get to a point in which you have… you know, at this point, like, I run… I have, like, 3 AI assistants. I just run them because one… all of them are customers, and all of them do something different for me.
you know, and I… some of them are good at some things, and some of them are good at some others, and I, you know, right now, nobody’s perfect at all of them, so…

Doug Landis:
Right.

Manny Medina:
Yeah, that’s kind of where it is, but, like, but actually, to answer the question before around seed pricing, I think there are some areas in which you’re still… the seed pricing is so sticky that it’s going to be so hard to move. Like, when you have a system of record. They haven’t charged by seat forever. Yeah. And I don’t know that that’s gonna change too much, to be frank.
I think you still need the concept of a seat for a while, and, you know, you’re not gonna fight gravity on that one. So that’s sort of like… The map, if you would.

Doug Landis:
Yeah, I was on a call with a bunch of operators, like, 250 of them, and I asked, I’m like, what percentage of your whole go-to-market tech stack is still SaaS? And the minimum response was 85%. So, if you think about it, so we’re talking about 15% of their tech stack right now, currently.
Now, that’s gonna change over time, is, you know, AI, some sort of AI-native tooling, so it’s like, alright, so guess what? So, largely, they’re still paying seats for a lot of their technology that they’re currently using. It’s the new stuff that’s coming out. And as we think about the new stuff, and I know we’re almost at time here, which is a bit of a bummer, but that’s all good.
And I do have some customer calls that have to come up… that are coming up, because, you know, we’re running startups. That’s right. I think one of the important things that we have to consider, at least in this world, and I’m curious your thoughts as you think about pricing and value is, like, how do we think about governance and accountability for these agents?
How do we think about, what does good governance actually look like in the world of these pure agent models that we’re actually paying for, their value, their workflow, their execution, their, you know, their outcomes. How do we actually think about, like, managing them or holding them accountable the same way that we hold people accountable?

Manny Medina:
Yeah, those are great questions. I don’t think that there’s an easy answer, but I’ll give you a handful of data points. So one is that right now, we’re still in the world where people are hiring more than one agent to do the same work, and it’s kind of the same thing that you hire reps.
Yeah, so, like, you know, any one shop that you see out there has more than one AI SDR agent, or more than one, you know, inbound routing agent, or more than one support agent. So this… people are actively trying a bunch of stuff. And then you can compare it against each other, right, to see which one has higher efficiency. So that’s one way to do it.
The other… the other sort of vector to it is that there was, you know, as usual, you know, you can make fun of it, as usual, Europe is regulating the hell out of this. So, like, there is an EOAI Act that is coming down the pike. That is kind of like GDPR, you know, 2 or 3.0. Then it has a lot of things around, like, you know. The model that you’re using, certification, model drift.
And that’s gonna put a best practices standard that people are gonna snap to, and…

Doug Landis:
Hmm.

Manny Medina:
You know, if nothing else. that’s gonna be a good way to do it. I think right now it’s still too wild west-y. Like, even when I tried to look for.

Doug Landis:
Coleus.

Manny Medina:
of value. I have to mix it up, you know what I mean? And I have to, like, look at research papers and try to figure out where it’s at, like, otherwise, you know what I mean? Where there’s nothing else.

Doug Landis:
We’re still, like, finger in the air.

Julia Nimchinski:
Yeah, I have to do it.

Doug Landis:
Man, this is… This is so not… this is so not enough time. This is so not enough time.

Manny Medina:
We’re just getting warmed up.

Doug Landis:
Haha!

Manny Medina:
But we have…

Doug Landis:
I’m just getting worse.

Manny Medina:
Good to see you, man. How you doing?

Doug Landis:
What’s up, Jay?

Julia Nimchinski:
Yes.

Doug Landis:
You’re welcome. Thank you. Manny, we’ll… Manny, we’ll pick this up on our own… in our own conversation. Let’s take it to LinkedIn.

Manny Medina:
Oh, 100%, 100%. Alright, brother.

Doug Landis:
Good seeing you. Bye, Julia.

Manny Medina:
I do it.