Mike Duboe:

Pranav, great to have you here. Thank you for joining us. We have been working with each other for some time and you've always been one of my favorite people to riff with on the topic of measurement, but a lot more than that. So maybe for folks that don't know you, Pranav has had a long-time career as a marketer, leading originally as a YC founder, but then leading growth at Dropbox, Magento, Bill.com, Pilot, and more. And now you're running Paramark, which is a Greylock portfolio company. But I think you are also — and I'm biased on this — maybe the smartest person out there on all things measurement and B2B marketing. And so I wanted to get into some of that with you today, and more. So welcome. Thanks for being here, Pranav.

Pranav Piyush:

Thanks for having me. I don't know if I will live up to the smartest man alive, maybe my co-founder is, but yes, I do know a few things about measurement and I'm super excited to chat about the whole topic. It's such an interesting time for the category, so let's get into it.

Mike Duboe:

Yeah, it's funny, Rishabh and I, we host this — it's been a while — but we used to host this dinner series at our office here, and one of the things we would joke about is keep a timer for how long it takes for the conversation to devolve into measurement. It was like CMO series dinners and eventually things kind of go to that. So I guess we're going to do it right at the start here and maybe ask why did you decide to dedicate your career or this chapter of your career to this problem?

Pranav Piyush:

Very good question. So I was a VP of marketing right before I start Paramark. And I had seen all types of conversations about measurement myself and was a little bit disillusioned. I actually didn't want to work on this problem. So the initial idea was something else altogether, but I went and talked to a bunch of CMOs and VPs of marketing in my network. And, like you said, after 10 conversations I'm like, Ah, man, everyone's still struggling with the same thing, which is how do I have a reasonable conversation with my CFO and with my CEO about what's working and what's not working? Where do I put my next dollar? That was really the spark for us of like, All right, it's not just me who's struggling with this, everybody else is struggling with this. Let's go build a solution that actually does it right.

We looked at the landscape, it wasn't interesting. This is back in 2022 and a lot has changed in the ecosystem, but that was the reason. And one thing that I tell all prospects and customers — and we'll start with the zinger — creative is Batman, measurement is Robin. The point here is measurement is not the end-all be-all of marketing. In fact, it is a support function. And the reason I bring that up is anybody who's saying that measurement is going to single-handedly change the course of your business is just lying to you. You still have to focus most of your effort on the creative. That should be the core of your business as a marketer, and then we're there to support the measurement stuff that should be table stakes for any business that's growing and scaling.

Mike Duboe:

Yeah. Well, maybe let's rewind and talk about what broken measurement led to this proliferation of new companies and then people caring about doing things the way you do, maybe to set some foundation here. There was a point in time where it became consensus that hey, direct click-based attribution no longer works and the way to do things is probabilistic methods. So incrementality testing or media mix modeling, and different companies emerged, took a strong point of view on one or the other. Paramark is ... Well, why don't you describe to people your stance on this and maybe a little bit of the history that led to your growth and your focus on this?

Pranav Piyush:

Oh my god, this is a very interesting discussion item because there are some things that have changed and there are some things that have stayed the same. And I'll talk about the things that have changed. The big tailwind for the entire category was really all the privacy changes that Apple made, number one on the list. And so I was in the middle of it, this is 2020, 2021, iOS 14.5, everyone's heard the story. But when Apple made that decision to limit the amount of tracking that's possible on the platform, it just made everybody wake up and kind of recognize that they're going to have to change things and how they measure things because Apple, iOS, the whole ecosystem, is just such a big part of the modern ... both consumer and business ecosystems. So that was the first big tailwind. The second one, surprisingly to me, was the whole ZIRP era and the whole conversation from the CMO, CFO, CEO going like, Hey, we can't keep spending money the way we were doing in the late 2010s or even the early 2010s.

Something's got to change. And so I think the focus on the financial responsibility for the CMO just became much, much higher. The stakes were much higher. The third thing that I talk about a lot, actually, is the fact that Google and Meta got saturated. I don't think many people talk about this. If you think of 2000s and 2010s, you could just show up on Meta and Google, and even if you didn't have great stuff going on yourself, because those two platforms were growing at such an aggressive rate, you're just going with them just because you're there, because the whole town square is there. What happened in the late 2010s or now the early 2020s, well, they reached a certain amount of maturity and saturation that just being there is not enough. So now you need to find arbitrage in different ways.

That's the third piece of the puzzle. And suddenly you have the Reddits, the Pinterests, the Snaps, the TikToks, and there's a lot more diversity that you have to try and figure out: How do you break through? So the combination of all of those three things, I think, just reignited the conversation. Now, I did say that some things didn't change. The fact that event A happened after event B is treated somehow as causal, that has never been true. It will never be true, but people just ignored it and just assumed that event A caused event B. And Mike, you said something about it being consensus now that that is not the right way of measuring things. Oh my God, that is not true. It is not consensus still. The number of people — I'm sure Rishabh has a point of view on this — that still talk about first touch, last touch, multitouch surprises the heck out of me.

Rishabh Jain:

Yeah, I think my point of view, which I am not shy about, is that MTA is used as a safety blanket to defend the decisions that you're making as someone who is operating the machine. Whereas other forms of measurement are the forms of measurement if you actually have business outcome responsibility. I think it's a very aggressive point of view, but I still maintain that, roughly.

Pranav Piyush:

There is another interesting point here — and I use this to kind of break through the conversation sometimes — humans are audiovisual entities. We need to be able to hear things, see things, touch things. That's how we make sense of the world around us. I give this example: The sun rises in the east and sets in the west, right? Well, not really. The sun isn't rising and setting, we're revolving. And so if you really understand what's happening in the world, you have a very different mental model, but we make it easier for ourselves by using that language because that's how we just communicate. And I think a lot of what's happening at MTA is very similar. We try to simplify the world to a set of observations that we can easily communicate about and the reality might be quite different from it.

Rishabh Jain:

I guess I'm wondering, on that topic, MTA does have this wonderful property that it's simple. The reason I think it gives a lot of comfort to people is it simplifies a lot of complexity. Now, whether it's correct or not is a totally unrelated question, but because it's simplifying, that helps. Whereas incrementality is actually not simple. It's a much more difficult concept or traditionally a much more academic concept. How are people moving it from something that's difficult to understand in a more academic concept to something that people can embrace in practice? And what do you think are some of the things that have happened over the last two years that have enabled people to actually be able to execute on this?

Pranav Piyush:

Yeah, it's a great question. I think you're spot-on on this observation that MTA has this property of being very simple, whereas incrementality, MMM, [is] much more probabilistic and therefore more academic to really grok. And I think that's the challenge for this entire space, for Paramark, for others who are operating in this space is how do you make incrementality in MMM as simple and easy to understand and grok as MTA claims to be? And I don't think anybody can argue that that has happened. It hasn't happened, right? Let's be very honest, because if it had happened, you would've had a billion-dollar, $10 billion, $20 billion company in this space. We're all working towards it. What you have to do is balance what is correct with what is simple and actionable, and that requires a new type of product design. It requires a new type of UX design that allows you to not hide away the complexity but still tell you the truth.

It also requires the user on the other side of the screen to be willing to handle the truth, right? Can you handle the truth? Can you handle the fact that the $300,000 a month that you're spending on Pmax may not be incremental, or may be incremental? I don't know. Right? So it's a big shift. The thing that has been very nice, actually, and I give credit to Meta and Google here, is they have done a fantastic job increasing education and awareness. Facebook's open source, Google's open source, LinkedIn has just joined the fray with their conversion lift, right? So credit where credit is due, they have lots of challenges with how they do measurement, but a few things, a few areas where they've pushed the ball forward and leveled up the conversation. And then obviously all the vendors in the space — Paramark, I'm going to name names here, the Hauses, the Recasts, the Measureds, the LiftLabs. I have love for all of them because they are increasing the level of conversation in the ecosystem.

Mike Duboe:

Pranav, one of the things you referenced earlier was the role of CFOs in these decisions, and I think you've taken the tack to sell to both CFOs and CMOs, and that dynamic can be tricky to play into. If you think about some of the companies that you work with where there's a very healthy CFO-CMO dynamic, how would you describe that and how should forward-thinking [CFOs] be thinking about marketing and setting up their CMOs for success as well?

Pranav Piyush:

Yeah, we made a very strategic decision in the very beginning to build everything at Paramark to support both the CMO function and the CFO function. What I tell CMOs is you don't want to play a defense, you want to play offense. What that means is you don't want to show up at the quarterly business review or your board meeting trying to justify your existence or trying to prove your impact. What you want to show up with is a series of bets that you're going on offense and finding net new growth. And when you have that conversation, you'll suddenly find the CFO's skepticism kind of melt down a little bit and the guard comes down and you're having a real conversation about what are growth levers. It is not easy because most CMOs and marketing leaders, they end up in the corner a little bit.

They have to be defensive because of just the cultural dynamic that's been set up in most organizations. But if you can leverage your measurement partner and tooling to effect that cultural change within the C-suite, then I think you go on offense. Now, one example that I give is I tell every prospect and our sales team does, Hey, can we bring in your CFO or your strat fund team into the sales process now? The reason we want to front load that conversation is if they don't buy into the methodology and the process now, six months later when you show up with results, you're going to have a very hard time convincing them. So you need to pre-sell this, and that's why this is not just the CMO sort of tool. This is a full business tool that you're buying. There are some customers who will tell me, No, no, no, don't worry about it, Pranav, the CFO gets it. And I'm like, Nope, nope, they don't. If you've not had a direct conversation about this, it's not going to go well.

Mike Duboe:

Yep. Let's stay on this topic. I want to understand, let's talk about the topic of planning and forecasting and how really forward-thinking functions do this because you take an extreme, you set a budget at the start of the year, marketers will spend to that, and the results you get are what you get. You maybe take another, I dunno if it's an extreme, but kind of another end of the spectrum, which is, Hey, hit a ROAS target, spend as much as you want up to that and max out growth. I think it's rare to get large-scale companies that operate that way for various reasons, but I'll ask the question again, maybe with the context of planning and forecasting: What does "great" look like here?

Pranav Piyush:

Yeah, it's a great question. Now, I think there is no one answer on what "great" looks like, because ... Let's be very crisp about this. When you are working at a company that's doing a few hundred million dollars in revenue, whether it's D2C, whether it's SaaS, whether it's mobile, whatever it is. Few hundred million dollars, you need to have a certain amount of predictability in your financial planning process. That's just a given. You're pre-IPO, you're going to go IPO at some point, or you're raising your next round. And what ends up happening is you have this financial plan that has a certain amount of parameters baked in. You're aiming for 60% growth at this EBITDA margin or this sort of gross margin, and you have to be able to show your progression towards that next phase of growth or that next big milestone.

So you got to know the constraints. The thing that great companies do within that environment is share those constraints with the marketing team. Everybody is on the same page about your top-line growth and your bottom-line growth targets, and there's extreme clarity about what the assumptions are. Now, then what "great" looks like is you have a top-down plan. Absolutely. You must have a top-down plan that says, Hey, for FY 26, we're going to spend this much money, and guess what? Our best case, worst case, and base case of marketing's contribution to that top-line growth number looks like X, Y, and Z. This is a joint planning exercise between your marketing, marketing analytics team, and your finance team. Great companies have that conversation early and often and they do it in a recurring fashion. You're doing this on an every quarterly basis.

Then you have a list of hypotheses, a list of bets. People will use different languages for this, but you list out here are the five goals that we are going to go to try and go from worst case to best case. And that's where experimentation comes in. And there's recognition that this top-down plan is a top-down plan, but the actual execution is going to be very different. It's going to be much more iterative, it's going to be agile — just like a product team, just like an engineering team, just like a sales team. You're applying that same mentality to your marketing organization.

Rishabh Jain:

How do you realistically execute with that rate of learning cadence or iteration based on what you're actually seeing in your marketing measurement? And how do you think about what the organizational setup needs to look like in order for someone to ... So you're going to a CMO, you're saying, Hey, here's how you should be thinking about it. The CFO loves this because to them it sounds like stage gating, basically, they basically hear stage gates. And so now they're asking you, Hey Pranav, how do we actually organizationally set up to execute in this way because we have not previously operated in this way? What's the first thing you're telling them to do to actually be able to embrace this way of operating?

Pranav Piyush:

Yeah, the first thing that I talk about is this is a top-down decision. You're not going to have a director or even a VP-level person be able to pull this off as a change management exercise. It has to come at the C-suite, right? So the CMO, the CFO, the CEO have to be bought in into this way of working. That's step number one. If you don't have that, everything else is meaningless. The second part is setting up your own cadence. And so this is where we're getting into the tactics. What this means is you always have a roadmap of testing and you have a monthly — if you can do even more frequent than that, fantastic — but you have a monthly readout on what's working, what's not working. This is based on your model outputs and your ongoing observability, and you have a very agile roadmap. With our best customers, we're talking about doing 20 to 30 tests in the second half of this year.

So you can imagine you're running two, three, four tests at the same time across the country or across the globe, and they're saying, Hey, can we do 30? Can we do 40? Can we do 50? And so that level of intensity, these are our most demanding customers. I love them, but I'm also like, Oh my God, can you handle this amount of change? Let's talk about the tactics here. We're talking about, okay, launching new channels, we're talking about launching new — completely new — creative and positioning. We're talking about leveraging their organic social and their paid social. We're talking about new spend levels, and everybody from the VP to the director, to the analytics team, to the finance team have a clear-eyed view that to get to that velocity of testing requires everybody to be on the same page. So a weekly meeting where we go down and look at the roadmap and we're saying, this is what's going to come up next. Oh, hey, we just had a big strategy meeting and we need to change things. It's very agile, it is very, very agile, and that's hard to pull off if you don't have that culture set up in the organization. If you are still coming from the world where I set the budget at the beginning of the year and I just do a bunch of media buys and I just wait and see what happens, oof. You're going to die. That organization's not going to survive.

Rishabh Jain:

Yeah, maybe I use this as an opportunity to pivot onto what can AI enable here? I hear this speed of testing, I hear all these things, and it's hard not to feel like, Gosh, AI must be highly enabling to create this change that you are seeking to push people towards. So maybe I just start with the highest-level question of, when you think about the way AI can impact how you can run the company and measurement practices, where's the first place where you feel like the biggest impact is to be had? Is it in the speed? Is it somewhere else? How do you think about how this technology shift enables this paradigm shift?

Pranav Piyush:

Yeah, it's a good idea. It's a good question to push on. I think of it in two buckets. There's the creative bucket and then there's the measurement bucket, going back to the point that we started talking about. I think AI does a lot for the creative, and I don't think that's something that anybody would argue against. There's a ton of stuff that we can do with gen AI on the creative side. I'll give you my own personal examples. Every time I'm coming up with an idea that I think is kind of nice and worth putting some money behind, we test it with AI to say, come up with five new revs, come up with ... there's a whole framework. I did a webinar on this just two days ago so it's fresh in my mind. Lots of opportunity there to cut the time — the zero-to-one time — for coming up with new concepts, on riffing on new things, whether it's copy, whether it's video and everything in between.

On the measurement side, the models have to be quite different. So we talk about AI at a very high level. The models and the AI that you need is a lot more mathematical in nature. The machine learning, the statistical models, the linear regression models, that's where the innovation has been for the last 15 years, and that's enabled a lot of growth in terms of what's possible. If you had to do this 15 years ago in terms of the rapid iteration of testing plans and measurement, you were hiring PhDs and building out in-house data science teams. None of that is necessary anymore. I would argue you still should hire data scientists for certain things, but not your tactical measurement stuff. The thing that we are investing a lot of time and energy in on the measurement side is twofold. One is how do you run more models in parallel to give you an even better understanding of the opportunity space?

And what that means is, instead of following a sequential — like, I do a test design and it takes me this amount of time to do one test design — can we spin up 10 possible test designs at the same time? And what ends up happening is our customers are looking for that agility because their business plans are changing all the time. And so you might have a discussion on Friday about doing a TikTok Pulse Up and by Monday you're like, Hey, actually we also want to launch a Pmax Pulse Up and by Friday you're like, Oh, actually we also want to launch a CTV. So how do you leverage AI or more agentic stuff to make that going from zero to one on that "I have an idea" to launching a test much faster. There's automation, there's running machine learning, there's make it one click publish to the platform, all of those things become a lot easier with AI. There's a second piece on the measurement side, which is all about interpretability — if that's a word — of results. Can I just talk to my platform? I open up my mobile phone and I just say, Hey, what does this test tell me about should I spend more money or less money? Dumb it down to a chat experience. I think that's going to come very, very soon for everybody in the ecosystem as well. So lots of applicability, it's just very different from the gen AI creative experimentation stuff.

Mike Duboe:

Staying on this topic of AI, one thing I've observed is that the composition of marketing teams is changing quite a bit, and in some cases you don't have marketers running marketing. And certainly, teams are getting smaller, but I think that the DNA that is running these marketing functions is changing quite a bit. What are you observing here? How does this impact Paramark, and how does this impact how you think about building your marketing function?

Pranav Piyush:

Very good question. I'm thinking about our customer base and what are we observing in our customer base. There's a couple of things that are interesting. So one thing that I'm noticing is the folks who are adopting incrementality and these advanced measurement capabilities are, by virtue of doing that, a lot more technical, statistical, data-informed by default. Even if they don't understand some of the data science behind it, they are a lot more numbers-driven, and I think the next generation is going to be even more of that. So I am seeing a generational shift in just the marketing persona. I'm not just talking about 25-year-olds, I'm also talking about the 4-year-olds and the 50-year-olds who are ... Those are the ones who are adopting this tech. The second piece that I'm noticing is there's a big alpha in creative and the spice and the human element. The way I think about our marketing team, you either hire for that taste and that creative element, or you hire for that extremely tech-forward, data-informed marketer.

Everything else in the middle, you are going to have a tough road ahead. What I find is some teams are very technically savvy and they outsource the creative piece because that's just how their teams are built. And for some it's the exact opposite where they have great creative taste in-house and they outsource the technical pieces. You can imagine there being a very nice hybrid format there. I do think it's going to be fewer team members overall. I'm of two minds [about] what's going to happen on the agency side. I actually think there's going to be a ton of AI-native creative agencies, like a mini explosion, Cambrian explosion, of those rather than big holdcos. That's just my take because you have to be agile to live up with all the changes that are happening in the ecosystem. So smaller, leaner teams, more supporting teams that are outside, and then measurement is just table stakes. And I think you're going to use, obviously you're going to use Paramark if you're a modern scaled marketer. (laugh)

Mike Duboe:

One reflection I'm having as you're talking through this is, hey, one could actually draw the conclusion that being a quantitative or heavily quant-focused marketing team is synonymous with performance marketing and you're going to move to only do things that are measurable and that's going to come at the expense of brand. Now I think part of using these methodologies, you actually, instead of some kind of simplistic attribution model is you are actually more likely to realize the value of doing brand efforts and more creative marketing. So I believe that brand can be measurable. It's funny, we had Tim Doyle from Eucalyptus, mutual friend, here on the pod recently, and he's probably the best marketer I could think of who deeply understands both sides of that. But for those who would draw the conclusion that hey, being overly focused on measurement just leads you to doing more performance marketing and maybe dismissing some of the other stuff, how do you react to that?

Pranav Piyush:

I think it's a false perception in the industry. Our earliest customer, the first big test that they did with us was TV, and it's for a mobile app and it had 50% to 60% incrementality. Now, it was a little bit more expensive than their other channels, but this idea that you can't be a tech-forward company and use channels like TV to find growth, it is just misplaced. You're just not thinking big enough. What it tells me is that the last 15 years of, honestly, this is where I think Google and Meta have done a little bit of disservice, is they've habituated everyone to think about what's going to generate a click, buy now, try free, 50% off, and guess what? All of those things existed before the digital channels. Coupon codes came from newspapers. There's nothing new about it. This idea that brand and performance are somehow different things is just not true. Good marketing is going to create positive mental associations, whether it shows up on a TV or a YouTube stream or a TikTok short. So I say it to everybody, really think about your priors and your assumptions about what is good marketing and think about your own personal buying experiences. And if you do that, you will realize that what people call brand absolutely works. We have a big labeling issue of we've taken a few things and put them in the brand bucket, and that's just complete nonsense.

Rishabh Jain:

First of all, I totally agree with you. So I want to start with that before I get to the question, so that way the context is clear. I just struggle to believe that if you said this to a CFO that they would say, Sign me up. And maybe this picks up on a thread that you were starting to embark on earlier with this idea of like, Hey, can you help me understand? And so you're chatting with the operating system to help you understand what's going on from a measurement perspective, but how do you think about what needs to be true to get someone to actually agree with this framing? So the good and bad of what Meta did is that they unlocked greater marketing dollars. They did that by giving more security through last-click attribution. Last-click attribution's greatest benefit, which is also its greatest flaw, is false precision. There's this false precision around like, Hey, here's the thing that led to the thing, but that false precision, it turns out, is great for CFOs because they can build a model in their head. All financial models are also based on false precision. So just walk me through how do you get someone there and do you think that the complexity of the measurement can be sufficiently abstracted with AI and interaction through chat, or what's your vision of how we get there? What are people doing in order to actually feel this picture?

Pranav Piyush:

Yeah, it's great. So let's talk about that exact example. That customer example that I gave you where they launched a TV campaign and they did it through an experiment to be able to see the impact on conversions. If we had just thrown money at it without running an experiment, I think that would be a colossal waste of time and energy because you'd be looking at the results and going like, Well, was that brand or was that just seasonality or was that the offer that we also did at the same time? No, no, no, no. Let's talk about it in a true experiment-oriented way and have the CFO, the CMO, the media buyer, all be on the same page about what is the hypothesis? What are we testing? What is the next action? If we get this result, we're going to do X. If we get this other result, we're going to do Y. So the framing for the CFO always has to be tying back to experimentation. This is not just a magical answer box that will give you answers based on anything. You're running causal experiments. That's the framing for the CFO. The other framing that's very important is, Hey, 80% of your experiments are going to fail.

Let that sink in, right? 80% of your experiments are going to fail. This is not Pranav saying this, this is not Paramark saying this, go talk to the people who run experiments at Microsoft, at Booking, at Uber, at Airbnb, any of these large companies. They will tell you that the meta analysis across all of these experiments conclude the fact that four out of five tests are not going to get you the growth that you think they will. So setting that expectation with the CFO, that you need to set aside a budget for testing, not everything's going to work on the first try, you need to iterate. And guess what? All of these things are true for your product roadmap, for your engineering roadmap. We've had these discussions forever, it's not a foreign concept. You just need to ground them in thinking about marketing in the same way, which is it is a series of educated guesses and bets. It's a series of experiments, and you rinse and repeat. If you can get that culture change at the CFO level, you have somebody who's going to support you forever.

Rishabh Jain:

I want to take this opportunity ... Because I know Mike and I have heard a little bit about your vision about if you do this super well on the measurement side, where does it take us next. When someone executes this well, what happens next? And what should this eventually lead to in terms of how an organization operates?

Pranav Piyush:

Yeah, it's a great question and obviously we think a lot about this. From the very beginning, our thesis was quite simple, that if you fix the dysfunction and the distrust that exists around measurement, if you can get across that hurdle, you open up the opportunity to help with planning and then eventually with execution. Planning is a very natural extension of measurement, and that's already happening with our customer base. Everyone's asking us for like, Hey, can you help us do FY 26 planning? That's the hot topic right now and will continue to be till the end of the year. But execution is the next piece, which I think becomes a lot easier. I'll talk about why this is. Again, the vision ... And if you take a giant step back and I think about how marketing teams operated five years ago, you've got some stuff in Notion, you've got a bunch of stuff in Asana, you've got a bunch of stuff in Google Slides. You ask the CMO at the quarterly business review, Hey team, what's our roadmap for execution for the next quarter? You get a little bit of a blank stare, you get a little bit of anxiety, and then they go ask their team to prepare a slide deck, and then none of it is tied to measurement. And you get this really messy story of what the heck are we doing as a business, as a marketing team, to grow this, to grow revenue? You've lived this, Rishabh. Mike, you've lived this. I've lived this. And so, our thesis...

Rishabh Jain:

I'm just imagining this really funny meeting. And so I can't help but laugh because I can imagine exactly how this goes. But yeah, please continue.

Pranav Piyush:

Five years later, I think what's going to happen is there's going to be a platform — I think that's what we're building — where you show up on a Monday morning and you log into your marketing operating system and everything from your list of ideas to your experiments tied back to measurement is all in one place, and the speed of going from an idea to a fully executed campaign goes down to minutes and hours. That efficiency allows you to increase your volume of testing and therefore your volume of learning. And if you adopt that mindset and that operating system, you're just a lot more efficient in figuring out what's working, what's not working. Now, I will also add, all of this only works if you have something worth selling. You've got to have a great product, you've got to have great positioning, you got to have great pricing. This is the communications and the go-to-market part of it. There's a whole bunch of product pricing, promotion, placement that we haven't even talked about, and that's a whole bunch of stuff that marketers should be a part of that conversation, but very often that doesn't end up happening, which is a shame.

Mike Duboe:

Yeah, I mean I think on this topic, there's a lot of ... There are many companies right now pitching autonomous marketing teams. I think very few of them are rooted in a long history of data on what actually works for that specific company and team. And owning that brain of efficacy should power a much more tailored and effective both experiment roadmap as well as eventually some semi-autonomous marketing function too. So I think understanding what works is actually probably the hardest part of that all, and that's kind of the dataset you're sitting on. I also want to address Rishabh's question on, Hey, would a CFO really go for this? In my personal life, take a brand search study uncovering, very quickly, dollars wasted, which lift tests could uncover a lot better than last-click models. As we all know, brand search gets an unfair amount of credit in those. Typically there's a study like that that can be done pretty quickly in a company where you more than recoup the cost of a SaaS investment. And actually that's kind of eye-opening to, Oh man, this stuff that my attribution models are telling me actually works is not at all incremental. Even this notion of incrementality has only been in the zeitgeist recently. So I don't know. In my life, CFOs go for it, but again, limited dataset.

Pranav Piyush:

No, you're 100% right. I think the objection that we hear a lot is people are set in their ways. And so it's a lot about the change management of even if they agree with that concept that, Mike, you put out, that you can run a test and you can see the impact, but there is just inertia. Humans are not rational beings. (laugh) It's like, Oh, am I going to have to go change everything in my financial model as a result of this and the amount of change that's going to cause me? So there's a class of finance people who could be in that bucket, but I completely agree with you that if you talk about it from an experimentation and lift studies perspective, the ROI is just so simple and easy to communicate to the finance audience.

Mike Duboe:

Yeah, I mean there's the famous quote — I forget if it's Ogilvy, it's embarrassing — but 50% of my marketing budget is wasted, I just don't know which half. If you could make that 40%, that's worth its weight. And I think that's part of what Paramark is doing. So anyway, jumping back into the topics, Paramark got started with more of a focus on B2B teams, which is interesting. A lot of other companies in the space focused on D2C eCommerce, and that is where a lot of the performance-advertising dollars are going. You focused on some very large and notable household names in B2B to start and I think resonated with that type of marketer. Now you serve both, but talk to us about why that starting point and maybe as you think about some of the best teams on both sides that you work with, what is most similar and then what's most different between those?

Pranav Piyush:

Yeah, it's a good question. I mean, listen, part of it was just my background and my network, because I worked at PayPal, Dropbox, Adobe, Bill.com, and that came with a bootstrap network that I could tap into. So there was a little bit of that. There was also a actual hypothesis that we had that B2B is much harder to solve for than B2C. Longer sales cycles, more complicated datasets, more complicated set of channels, smaller sample sizes. All of these reasons made it harder for this stuff to work in B2B. For a long time we actually got that as a pushback of like, Does this work for B2B? And then we showed them that yes, it does work for B2B and got over that hump. So if we can solve it for B2B, we can easily solve it for B2C because the dataset, the shape of the data, the marketing channels, I think it's actually much easier to solve on the B2C side.

Now, it gets a little bit more complicated when you get to omnichannel B2C, so you're talking about Amazon and Shopify and retail and third-party retail, and suddenly you have a more complicated dataset to work with. So that starts to look much more like B2B, actually, where you've got sales and marketing and outbound and inbound and all that fun stuff. So there's lots of parallels. The one thing that is quite different, and this is not actually a property of B2C or B2B, it's more about the consideration cycle. You could have a $50,000 consumer product — cars — you could have a $50,000 ACV B2B SaaS product, right? And so the consideration cycle, the size of the purchase changes how you think about measurement. What is a leading indicator, because you don't have six months to wait for that consideration cycle to be complete before you call the test. So as your AOV goes up, you have to be agile and nimble to think about how are you going to measure the success in a short enough time period that gives you the ability to learn and iterate. And that applies on both B2C and B2B.

Mike Duboe:

Going back to some of those early B2B customers, why did they ... The business has grown a lot over the last year, but if you go early on, my observation is some very large notable CMOs entrusted you as a pre-seed startup to handle these pretty big consequential decisions. Why was that the case and maybe what lessons could you generalize around that for other founders listening?

Pranav Piyush:

Yeah, it's a good question, and funnily enough, I think the story has stayed pretty consistent from that early customer to even now. We were literally talking to a CMO yesterday who was talking about why they are even talking to us given their spend and scale. It boils down to three things. One is, can you make this actionable and simple for a marketer? There's a lot of data scientists that are a lot of platforms that sometimes can over index on the data science and not make it actionable enough for a CMO for them to trust and believe. So can you talk CMO and can you talk data science at the same time? That's a very hard skill, in my opinion. If you're building a company that can do both, you'll have a lot of success. So that was our secret sauce in the early days, and it continues to be that's how we hire people that are working with customers.

The second piece is, I think the founder market fit piece is very real. I can talk about my real stories of, Hey, I tried this thing when I was a VP of marketing at Invoice2go, I tried this thing when I was running growth at Dropbox Business, I tried this thing when I was doing Magento or Pilot, and that becomes a very easy way to have a conversation with a prospect. Even when we hire salespeople now, we're looking for people who have real experience in that domain, and that's how you build these relationships. If you're going to sell a $100,000, $200,000, $300,000 product, people want to hear that you have walked in their shoes to some degree. And I think that helps a lot. So talking the right language, the founder-market fit. The third is you're going to do things that are not scalable in the early days.

Classic, do things that don't scale. And that means that I'm here not to sell you a tool, I'm here to solve your problem. Very often in the marketing measurement space — and I would imagine this is true for other business as well, and Rishabh, I'm curious to hear your take on this — if you can't actually solve the problem, then you're not going to have a sticky business no matter how good your tool is. This goes back to the point of ... The conversation I was having yesterday, the CMO said, I just don't have anybody in my team who can leverage an incrementality solution if I don't have the extra help from the measurement partner to drive the week-to-week execution. I just don't have that skillset on my team. So understanding that and building the platform and the solution and the service to take it all the way to the end state of you launched the test, you achieved the outcome, you got the next action, let's rinse and repeat. People were looking for that, and that's hard.

Rishabh Jain:

Yeah. My quick view on this is — and we should definitely talk more about it another time — is I'm starting to experience across all different types of software including us that yes, people not only want the tool, but they want someone to service it, whether you as the company decide to service it or you have a tight partnership with someone else who can service it. People want to buy outcomes, and that's becoming more and more common. Whether AI is accelerating that, is causal to that, it's hard, who knows? And it's not particularly relevant, but it is definitely becoming the case that people are expecting outcomes more than they are expecting tools at this point. I wanted to ask, though, you were saying that people trust you because of your background and that's how you're hiring salespeople and you were thinking about the scaling. Yeah, my last question here is, as you think about the stage of the journey you're at —feeding on this idea of how should other founders be thinking about this — how do you think about which elements of that go-to-market can be scaled? Because founder-led brands are — whether it's a B2B brand or B2C brand — are becoming more and more common. I'm wondering how you think about what your role becomes and what roles then you need to hire around you to support this founder-led brand that you're building at Paramark, and how should other founders be thinking about it as they go through their early- and mid-stage journeys?

Pranav Piyush:

Yeah, the first thing I would say is, I think this is one way of building your company. I don't think this is the only way to build your company. A founder-led brand is one template, but I think there are many other successful templates. I know many companies that have successful businesses where the founder is never on LinkedIn, never on Twitter. So please do the thing that is going to be useful for your business. This is what worked for us and we're going to leverage it. I do think that what's happening now is you need employee advocacy. And what that means is ... We haven't done this successfully yet. Everyone's heads-down busy working, so they don't have time to show up on LinkedIn and do the advocacy on our behalf. But I do think that there's going to be companies who get their ... We call them growth advisers.

These are post-onboarding team members who are talking about ways in which they're helping their customers. Hey, this is how we solve this really complicated gnarly measurement problem. Hey, this is how we solve this really complicated scenario-planning problem. And you talk about that openly. You are authentic, you are direct, you're transparent, and that becomes your marketing. And Sam was talking about it in the last podcast, how you can take that and do the white listing and the boosting and all of that fun stuff. You can do that with your own employee content, and I think that's going to be the real unlock for us. The one thing that we look for when we're hiring customer-facing team members, whether it's sales or growth advisory, we're looking for people who are intellectually very curious about this space, and they have the intellectual firepower. They may not yet have tons of marketing background, but they can pick it up because they understand statistics, they understand machine learning, they understand data science, or they have tons of marketing experience and they have a really strong head on their shoulders, and they can pick up the data science within the first couple of weeks of onboarding with us.

That's the skillset that we look for. And then let them loose, let them have fun and go help customers, and that should just work.

Mike Duboe:

Yeah. I want to close with a related question, and maybe to put a finer point on what Rishabh asked. If you run your own incrementality test on your set of tactics, marketing Paramark, founder-led marketing or otherwise, what tactic do you double down on and what do you start to remove from the mix?

Pranav Piyush:

Great question. We've been very lean in how we've operated, but if I were to do that, paid social, so organic has worked really well for us. We're going to double down and do a lot more on paid social. Second is podcasts. We've had our own podcast, I've done a ton of podcast appearances. I think it is absolutely going to crush for us. We haven't done a bunch of sponsorships, but we've done a few. We've had a mixed bag with events. Events are good, but it's a lot of work and you got to be prepared for that. So as a lean seed-stage company, I don't know. It could work, but it's hard in my opinion. Just the execution is hard. Yeah, that's the only one that I'm less excited about in the short term. Yeah.

Mike Duboe:

Great. Pranav, it's been a great chat as always. Thanks for jumping on with us and yeah, we'll chat soon.

Pranav Piyush:

Thanks for having me. This was great.

‍