SFG 37: Andy Byrne on how AI is changing revenue ops

‍

In this episode of the Startup Field Guide podcast, Sandhya Hegde chats with Andy Byrne, co-founder and CEO of Clari. Clari was started in 2012 and last valued at $2.6B. Its revenue platform has been helping enterprise sales teams create visibility and predictability within their organization for over a decade. It helps over 1,500 companies including Cisco and Databricks improve their entire revenue process.

‍
Be sure to check out more Startup Field Guide Podcast episodes on Spotify, Apple, and Youtube. Hosted by Unusual Ventures General Partner Sandhya Hegde (former EVP at Amplitude), the SFG podcast uncovers how the top unicorn founders of today really found product-market fit.

‍

If you are interested in learning more about some of the themes and ideas in this episode, please check out the Unusual Ventures Field Guides on coming up with a founding insight and working with design partners.

‍

‍

Episode transcript

‍

Sandhya Hegde

Welcome to the Startup Field Guide, where we learn from successful founders of unicorn companies, how their startups truly found product market fit. I'm your host Sandhya Hegde and joining us today is the CEO and co-founder of Clari, Andy Byrn.

Clari was started in 2012, last valued at 2. 6 billion dollars. And it's a tool I have used myself. Every time someone said, okay, are we going to make our quarter? Clari was the go-to-tool we would open up at Amplitude. So really excited to dive into the story. Clari is a RevenueOps platform. It's been helping enterprise sales teams create visibility and predictability within their orgs for over a decade now, and actually works with over 1500 companies, including companies like Cisco and Databricks to improve their entire revenue process. So going back to 2012, Andy, could you share what was the origin story?

Andy Byrne

Origin story is one word. Pain. I've been in and around large sales teams, building them, scaling them. And I just had so much pain trying to run what today people call the revenue cadence, every meeting of every day of every week of every month of the quarter. And it was just from rep to exec. It was so hard. And so we worked with Sequoia Capital, who's our largest shareholder. Went to both Jim Goetz and Doug Leone over there. We started to brainstorm around this idea that this is just not right. We were in a CRM, and the CRM wouldn't do what we wanted it to do, so we'd go into Excel.

We'd have Excel hell from rep to exec, and then they would then turn on BI. So the industry calls us the three-headed hydra, the monster. And we thought there's something's wrong here. And we started to analyze the CRM database. And we realized why there was this export, this jump. It's because there were some technical issues there that it wasn't built to actually run the revenue process.

And it wasn't built from the ground up to actually do artificial intelligence and predict will you meet, beat or miss, which is arguably the most important question in business.

Sandhya Hegde

And when you first started Clari what was it like? What was the reaction you got from your other folks in the ecosystem. Because I've experienced that often when you take a process that's typically run on run in Excel and in worksheets and try to kind of start a company around it, you will often get a lot of skepticism from folks who are like Excel works and is this a feature or a platform?

What was the early reaction if you go back to 2012?

Andy Byrne

What I'd say is it's funny. The people always think, hey, it's the most romantic, stage of the company. You don't have a number to hit. You're just you're creating something from nothing. I would say that it was the most stressful time because there's so much uncertainty and you're not sure and you're getting a lot of rejection.What I'd say is that the big moments where we felt like, wow, these are big leaps is, we started analyzing all the human behavior that was out there. We'd look into a CRM, like a Salesforce, and we'd watch everything that everybody's doing in that CRM, and we'd historically snapshot everything.

And then we started saying to the customers, we're going to go into your email and we're going to look at all the emails and we're going to go into your calendar. And we're going to use team linking for exchange in Gmail and Office 365. You know what we're going to do? We're going to monitor everything everybody does.

And a lot of folks are saying, you're going to do what? That's crazy. And what we realized is that used in the right way, all that data and being able to do machine learning on that data and predict what's going to happen. It was like a game changer that people actually got over this reluctance of this is like big brother.

I'm not going to deploy this because what are my salespeople going to think? And what are the leaders going to think? So the bet was that data is so important, so much more important than CRM data, that people would put it to good use. And we started to see people say, Oh my gosh, I've seen things that I've never seen before. I'm eliminating blind spots. I'm a better manager. I'm a better rep. I'm a better exec. So that was a moment where a lot of other folks they were thinking about maybe investing in the company said, No one's ever gonna buy that. We had a lot of that early days and that was stressful and uncertain, but what did you do?How do you get over that? You just keep working closely with the customers. We had a handful of them. We had HP, Okta, Nutanix. And they said, keep building, keep going. And they were our design partners early on. And that's where we started to get some momentum and some wins. Helping a rep close their deals faster, helping a manager drive more revenue, and helping an exec boost the predictability and accuracy of forecasts is a big deal. And the company took off.

Sandhya Hegde

I have so many follow up questions for you. So if you go back to your work with those design partners, do you remember any surprises? Because I feel like the original vision for Clari feels so consistent with what you are doing today. Like there definitely wasn't a big pivot, but I'm curious like were there any surprises for you as a founder in terms of what you expected the product would look like exactly versus, where it ended up after feedback from Okta, etc?

Andy Byrne

Yeah, for sure. PRD was what we call the golden triangle, rep, manager, exec. We didn't know what the hell we're going to call the product. We just said, let's go use a new type of database that's built from the ground up to do machine learning, predict and answer the most important question on where we have risk and pipeline and deals and reps and segments and geos. And we called it a predictive sales management solution. It wasn't around revenue. The big aha, the big surprise for us was that we started to see organic license growth outside of sales teams. We started seeing marketing and pre-sales started coming into our platform. The post sales motion to run upsell, cross sell, and whitespace analysis and SaaS parlance, that's the net dollar retention motion, right?

We started to see all of these people that are revenue critical employees that were not in sales coming into the platform and we didn't change the platform at all. And we didn't actually even think about those use cases, so I remember coming back to our small office in Mountain View, California and saying to the engineers, Oh my gosh, so much bigger than sales. It's around revenue. And at the same time, we were seeing the rise of the CRO, Chief Revenue Officer. That title did not exist 10 years ago. We started to see proliferation of APIs and data. And the rise of artificial intelligence. So we had a handful of things. You got to have a little bit of luck.

And I think that was some luck for us where we did not realize we're building a revenue platform. And that was so much bigger, and that was our second liftoff. So first liftoff was solve for the Golden Triangle. Those people in the sales hierarchy, right? And got to about 150, 200 customers, and Bain Capital did our B.

And then we did follow on with Madrona and Sapphire. And when we realized that it's about building a revenue platform. And running that entire 13 week cadence and getting all revenue critical employees. We had another liftoff again, that had follow on rounds from Silver Lake and Blackstone.

Then in late 2021, luck makes you look smart. We just got lucky. What happened is Clari was starting to show up in their board meetings, And the partners at those firms were seeing Clari was actually being used by their highest performing companies. And so we actually engaged with those partners and raised some follow on rounds and we've continued to continue to scale.

Sandhya Hegde

I want to go back to the golden triangle.It's actually typically harder to build products that have multiple stakeholders and users from day one, right? You don't have the luxury of saying, I'll just like, make it magical for this one and so you have to think about multiple stakeholders. So puts more pressure on your minimum viable product. The bar is higher. I'm curious how you approach that. What was your intuition? What were some of the friction points and having it be really intuitive and work out of the box for I would assume the RevOps persona as well over time, you're just adding more, but could you share more about that?

Andy Byrne

The one thing I'll say is that there are so many times when we'd come back from a customer, we'd go, what are we building? Is this right? There's so many anxiety moments, right? Or imposter syndrome moments as an entrepreneur where you're just second guessing and you go why is it that they're not absolutely falling over themselves and you have to go deeper on the why, as we all know that you have to eliminate the happy ears when they're excited.

You have to actually become a skeptic and say why are you excited about this thing I want to build for you? And go deeper, deeper, deeper and do that level of really rich diligence. What we always knew as we looked at this three-headed Hydra that was being used is we started to study basically the army of people from rep to first line, to second line, to third line, and we started to interview all these people.

It was pretty straightforward and we went, Oh, my gosh it's so much more than a rep productivity problem, right? You got issues everywhere. And so we methodically said, let's just simplify. Let's start with the rep and we did. That's the first thing we had a mobile app for the rep that helped them close their deals faster.

And of course you can't build a company on that, right? How much are they going to pay for that? 10 bucks a month. And so we always knew because we saw this super interesting intersection between the rep and the manager, and it wasn't necessarily about building for. the people It was about building for that intersection between those people where we felt like we had aha moments going, Oh my gosh, look at that collaboration. And then above the rep, we saw there's the execs and there's collab that happens with between the frontline manager and second line on all the way up. And we realized all these people, what are they doing? They're collaborating around revenue and the industry calls it revenue collaboration.

So that was how we saw, we pieced together all of these different moments and we literally just interview, tell me about what you do in those meetings, tell me about the tools that you use and what sucks about it. And there was so much pain. We just felt like this emotional fiduciary responsibility, like an obligation to go fix this.

Sandhya Hegde

Makes sense. I'm curious how you thought about go to market strategy, right? Obviously you need a top down motion if you're going to access confidential data from a company's CRM, you're not going to do PLG if you are working with the HPs of the world. And I would assume the bigger the company and the bigger the team, the more the value prop of Clari-like alliance, right?

But you also have the opportunity to do something bottom up that is just focused on kind of rep productivity. How did you think about go to market motion strategy and what were the things that worked for you in the early days to build awareness, to accelerate sales cycles, et cetera?

Andy Byrne

There's a lot of questions in there, so I'm going to just start with the first question on how did we think about go to market? I don't think we were very elegant around go to market. If you look at when we started the company, it was 2012, 2013. Really didn't get going in earnest until late 2013. Really didn't start actually having a real go to market until 2014. PLG wasn't a big deal in 2014. It wasn't really the rage. And we always knew that we would need to build a platform to have companies like we have today. Johnson Johnson. ADP Perkin Elmer, Sodexo, Cisco, and others, right?

We knew that we were going to need an enterprise motion. My DNA is enterprise. And I'm older generation, so I'm not a PLG person. So the answer is that we just knew that we wanted to be doing the seven and eight figure deals that we do today. And so we just were very focused and disciplined on having an enterprise motion.

Sandhya Hegde 

Got it. Makes sense. And you have already talked about being the center for revenue. I'm curious, how do you think about the other opportunities that Clari has, especially since you're already in email, you're already in the manager, rep coaching workflow. How did you think about what are the other adjacent opportunities and how do you in general decide what Clari should pursue and what Clari should not pursue given it's like endless opportunity.

Andy Byrne

 There's definitely endless opportunities. I don't know if there's a straightforward answer here, but we looked at the world as, hey, there's all this structured and unstructured data. And where can we analyze the structured and unstructured data to help people do the best revenue collaboration and how can they govern the revenue process? We call it revenue governance, collab, collab and gov. And we started to see opportunities around other data that's out there, namely conversational intelligence data, which we acquired a company called Wingman that we fell in love with. We're using an alternative CI solution that we started and started evaluating this really great product from this small company in Bangalore, India.

It was doing incredible things and we fell in love with it. We acquired the company and we started to use it. We are not just doing it for better coaching, but think about taking that unstructured data and using it to inform your forecasts and to use it to inform every one of those meetings that we're going into where if you're the rep and I'm going to go inspect your pipeline and we're not going to do coaching, but we're going to understand how are you going to hit your number.

That's additional data. So there's moves that we're making around what we call data, structured and unstructured. The second way we think about it is workflows, right? So we think about our orientation, our origin, we always thought about, hey, how do we run the revenue process starting from the golden triangle, right?

From rep to exec and running that internal workflow cadence, one on ones, pipeline, inspection meetings, slip deal reviews, forecast calls. So we are thinking about that and what's the database that we need to build to run those. On the external workflows,we started to realize, Hey, there's opportunities for us to extend our value footprint.

That's where we acquired Groove which is a company that does sales engagement. And, we were internal users of Groove, we fell in love with the product. So it's a similar story where it was pretty straightforward for us to extend our value footprint. So we're always looking at hey, is there an ability for us to accelerate value delivery for our customers in two dimensions, data and workflows, and now a third dimension around around artificial intelligence.

Sandhya Hegde

Makes sense. And, bringing us to AI, I'm curious obviously you were already an ML first company, right? Like you were in the prediction business from day one. What was the kind of moment in time when you looked at the, the birth of LLMs and you know, saw the opportunity or, and how did you think about what is the right way for Clari to adopt some of the more novel developments in AI over the past two years?

Andy Byrne

I'll just wanna make one comment on the predictive artificial intelligence, if I may, is my learning on this to share with the community is it just is fascinating to watch people go from skeptical to curious, to reliant. And people using Clari all the way up to the boardroom, right?

People using it to say, okay what call are we going to make to the street for these publicly traded companies? Or what is Clari calling? In the early days the customers saying, oh, that's so cute. Oh, that's really cool. And so it's been fascinating over a decade of run-ups to see how hundreds of thousands of people have changed the way they think and the reliance like the way we look at our phone. Nobody ever questions. Are we going to be there at the time that, Google Maps or Apple Maps tells us we're going to get there. So we're just relying on it. What I'd say in terms of the generative AI work is when did we start to realize that it was very powerful? It was about two and a half, three years ago when we started getting access to these LLMs.

And it was just through Sequoia and our other connections through our board, we got early access to these things and there just started these experiments that were popping up all over the company and they were feature-flagging and turning this on and saying, Andy, check this out. And, for us to be able to go into a deal. And we call it Rev. ai, have that go through using large language models and tell me, Andy, of these 150 emails that have been going across here's how that deal is going. And here's what we think should happen next by this person. Here's the action items. I was floored. And so we started officially turning on a lot of distributed work around LLMs. And along with our predictive and our NLP work, we're now doing this generative work, and we go back to looking at that 13 week cadence of every meeting of every day of every week of every month, and we're expecting, okay, where is the moment where we can drop in generative AI? And make a massive difference for the people that are collaborating. How can we use generative AI for the executive who's trying to govern and make sure people are doing the right things, the right accountability, all that stuff, and the way we think about it is, how do we shrink the time to revenue, the time it takes somebody to get results from their revenue process? And generative AI is making some profound moves across all of those, we call them revenue moments.

Sandhya Hegde

Maybe a bigger question. I'm sure you're thinking about what would sales even look like in 10 years, right? It is a very human relationship and language-oriented profession. Even if your company is following a product led growth model, all of the expansion is extremely sales-driven and relationships matter, understanding matters. And then at the same time, we are seeing,a combination of workflow automation, smaller parts being accelerated as well as I would say a lot of early experiments and full autonomy, right? Can we fully automate outreach and meeting booking for cold outreach? We have seen interesting things around personalized demo generation.

How do you think about what a mature sales motion and a sales org would look like in 10 years. What would be your kind of wild guess?

Andy Byrne

Yeah I would say that if you look at the world through the lens of create, convert, and close. Just to try to simplify what's going to happen and how we create pipeline, how we convert it and how we close it. I do think that if you go back and you talk to what did the customers think and how are they thinking about deploying it? There's a lot of after the fact what happened moments, and I will, just tell you that those numbers of okay WTF  tell me why that happened and the back and forth going into the data and looking at, okay, wait, are conversion rates right? And there's a lot of post mortem in the revenue process.

And what I would say will happen is that a lot of that's going to go away.

We're going to be catching that so much earlier. And there's not going to be a, oh, what do we think is going to happen? AndI think what happens is in the create, if I just take each one of these in turn as we think about top of funnel creation  I think you're going to find that there'll be higher quality pipeline that is created. There will be a really deep and rich understanding of how multi-product actually works in a company. I talk to companies all the time that are multi-product, huge multi-billion dollar companies. You know what they say to me? We suck at multi-product. We're horrible at it. And they don't even know how to look at the data and segment the data and run A/B tests from the website and what I call throughline all the way down to enabling the rep to then run a multiproduct motion.

I think that gets so much better and we get insights on create, on multiproduct. What I'd say on convert is I think you're going to see the rise of the super human frontline manager. I think a lot of conversion issues are not around rep productivity. It's around the manager's inability to understand what the hell is happening across pipe. And if you're a manager, you have 10 reps. They have 10 deals. On Sunday night, you're trying to figure out where are you going to spend your time next week, right? And what's going to happen is this generative AI, Rev dot AI is going to come in and say, okay, here's where you have risk. Here's where you have upside. Go. And imagine doing that across thousands of frontline managers at Cisco. The level of productivity on how we convert, level conversion will be fast. And then on close, I think that there's some really interesting paradigm shifts that we'll see around buyer and seller collaboration, that we've not seen before that will allow people to really understand where they are in the buying journey, or where they are in the selling journey, and that you will see the ability for generative AI to go into both structured and unstructured data.

I'll give you an example. You're seeing a lot of data. We pull data from Snowflake, PostgreS, BigQuery now, right? We're analyzing all of this consumption level data. And then you also want to see the relationship work that's being done. That is data feed 1 is structured. Data feed 2 is unstructured, right? How do we then get the machine to go in, look at the history of deals that were very similar to this deal that's being executed, and then discern is this tracking in a good or a bad direction, right? That's really hard to do. Now I think as I think about create, convert, close, What is going to be the most important thing around the future of how we run revenue is what we call a revenue database. Customers have a CRM database, they have a marketing database, they have an ERP database. And now, for the first time, they have a revenue database, which is probably pretty important. It's the most important business process that they're running in the company. And it's that, and it's really the data that sits. The structured and unstructured data that's sitting there and being historically snapshotting. And using that data to run those workflows that'll drive better create, better convert, better close.

Sandhya Hegde

 Makes sense. And, you brought up revenue database. How do you think the CRM providers need to like really meet this new moment where, you know, underlying architecture of where CRM providers started is based on pretty much what everyone feels like is an old world now that doesn't exist anymore, but obviously they're like some of the most sticky technology ever invented.

What do you think if you were running a CRM company, I won't name which one, how would you approach this moment?

Andy Byrne

I  think they're approaching in the right way, frankly. I think that they see that there is a  revenue database peacefully coexisting with a CRM and an ERP and a marketing database makes a lot of sense. Who doesn't want a revenue database, which we think is going to be the most important database inside the enterprise.It's not replacing the CRM. The CRM is still critically important. If anything, it's actually making the CRM investment better.

Do you think there will be an opportunity for new types of CRMs to be born that are maybe more specialized for like different industries, at this time? I do. I do think that we'll see the birth of new types of databases that are built from the ground up with generative AI in mind and workflows in mind. I just think about the speed at which we see people being able to create content. We were talking in the beginning of this call before we jumped on here about art, right? And how fast you can create art. I do think you're going to see the proliferation of new types of databases and workflows that will disrupt a lot of existing incumbents for sure.

Sandhya Hegde

I've been investing in go to market tech for a while now and often every time there's a new startup pitch where they say, here's a particular kind of problem in a workflow and we want to build a point solution for it and spread from there, it often seems like the sales tech landscape has gotten really crowded with those, right?

Some of the, even like series B, C stage startups I talk to will have 50, 60 different go to market tools that they have strung together and can justify the need for . They all do some specific thing they really want. What has been your perspective on seeing that evolution? Because I think it's like very much happened in the last 10 years. What's your perspective on that? And, where do you think it's headed in terms of either there being a lot of consolidation or a complete disruption given what you can do with unstructured data today?

Andy Byrne

Yeah, I think so a couple of data points for the audience. One,CNBC dropped their recent jobs report earlier this year, fastest growing job in the United States is revenue operations. Blew us away, frankly, as  starting and creating the category. There's 500, 000 jobs on LinkedIn. So that's crazy. If I look across all of the venture data, over 300 new companies have been built over the last year that are focused around RevTech, right? That's one company a day. That's crazy. A lot of these SI communities, if you look at, okay, so the likes of Deloitte, Accenture, PwC, Sloan, they're building revenue transformation practices, right to help them come in and so what's driving all of this? I do think that if I was to simplify what's happened and the tipping point that's occurred is people realize revenue is not just an outcome. It is a business process. It happens to be the most important business process in the company. Hard stop. Like no one can debate that all the way up to the board. It freaking matters. If you look at the value proposition of a company that's trying to come into RevTech, if I look at the security landscape versus the RevTech landscape, security landscape provides technology that protects your downside, right? To make sure there's no issue in terms of no breach that would tank your stock. In the case of RevTech, not only does it protect your downside so you can answer the most important question in business, will you meet, beat, or miss, but it actually gives you upside too and has the ability for you to, if I think about the big force that everybody is thinking about, all the venture folks are talking about revenue leak, right?

And that's the revenue that we're leaking. Whether it's, leads that get created and nobody follows up on them. It's targets that you've given to reps and they never actually penetrated those targets. These are all examples of revenue leak and it's a big deal. We've got one data point for the audience is BCG had released a report last year that said that the data suggests that up to 2 trillion worth of revenue is being lost every year across all B2B companies.

We did our own analysis at Clari Labs. Up to 15 percent of your revenue is lost due to revenue leak. So imagine all these companies that can add another 15 percent of growth. That's crazy. So I think that's what's driving all of this new venture money, all of these new startups. And yeah, when it's a hot market, it's a crowded market for sure.

Sandhya Hegde

And, at the same time, especially given the market correction and lots of layoffs and go to market teams, it's also actually probably right now a very tough time for early stage sales tech teams and teams that are struggling with a lot of partial churn from their customers downsizing.

What would be your advice to a new founder just getting started in the go to market tech environment? What would be your advice in terms of like how to think about your idea is it is it a strong enough idea to build a company around?

Andy Byrne

You have to play in two different worlds. They have these polarities, right? Pole one is, you gotta think long-term and believe in your vision and you gotta go for it. There's so many times when I'd articulate the vision and people would say, that's so far out. Of course it's far out, like I'm 10 years in and I feel like our best work is in front of us. You have to just have that long term vision, focus on that, and then whatever is in your control. A lot of the macro stuff is not in your control, right? And we're gonna see the craziness of inflation hitting 9%, now it's down to 3%.

We'll probably see a pause of interest rates and, we'll see maybe some better moments in time for tech. But, you can't be distracted by any of that. As an entrepreneur, you just got to focus on, go grab a handful of your early customers that are believers, that want you to build this.

And then you have to do super creative. things to make sure that you can start deploying this and start learning and creating those learning loops and then growing your footprint and trying to get a commercial mindset and then trying to land a handful of your first customers,  no one should be scared about macro.

Just focus on what you can control. And the entrepreneur, if they're really entrepreneurial. They've got so much goddamn grit and they won't let macro affect them. They'll keep going.