What is really valuable now is like, oh my gosh, that person gets my business. If we can't figure out how to make agents and humans work together, the whole idea of AI fails.
What do you think is the future of the SI ecosystem?
Customers want to buy outcomes, not hours.
Welcome to the last mile, our conversations with enterprise AI leaders to demystify everything from FD as a function, the future of AI native services to how people are actually deploying AI inside large enterprises. I'm Ivory, founder of Contour, and I'm here with Nick Mehta, the former CEO of Gainsight as well as Snicker Doodle.
Snicker Doodle has busted in here. He's my dog. He's going to share the wisdom with us. It's awesome to be here.
Yeah. Thanks so much for being here. I'm excited for this conversation.
So we last saw each other at a philosophy event actually you hosted and we were talking about AI and consciousness. AI is loneliness. I know you also started the New York Foundation lately. Tell me what does a typical Tuesday look like for you now?
Oh my gosh. It's crazy just like it is, but crazy like we're so lucky to be able to work on such amazing things, you know? So I feel anytime I'm like really busy, like today I was telling you I had like 15 meetings today, and it's just so energizing meeting entrepreneurs, seeing what everyone's working on. Also just seeing people talk about the implications of all this stuff that's happening. None of us know. Anyone that says on X what they know that's going to happen more than a month out is totally making it up. But what I also love is the generosity of spirit of people trying to help each other, which I think has always been true, but I'm seeing it more and more particularly on the job front.
This New York Foundation that Clara Shai started, an incredible entrepreneur, and then she brought me on as co-founder, is basically all about helping companies figure out how they can help people find that right on-ramp to get that right next job to get into the world given all the craziness with AI. Careers are so in flux. And then for job seekers, people coming out of college, Gen Z folks, anyone, how do we help them navigate their career, find their next role? And so building tools, resources, content, particularly initially focused on young people because right now it's pretty tough. There are a lot of people looking for jobs. It's very hard. And so you can look up New York Foundation to learn more, but it's really exciting to see the work that's being done already.
Yeah, I love the work that you guys are doing at Newark Foundation. Maybe let's go back to also the early days of Gainsight as well. So over 13 years you built Gainsight into a over billion dollar company. You guys owned over 70% of the customer success software market. Now in the early days I remember there were only a thousand CSMs maybe less.
What was it like back then jumping on into Gainsight as CEO and what gave you conviction that the market was going to be huge and not really a niche, which it might have looked like at the time?
When we started the company, there was this benefit of naivety that we did not look on LinkedIn to see how many CSMs were there in the world. Because the basic idea, as some people know, is Gainsight initially, we do a lot more now, but initially is building software to help CSMs do a better job retaining and expanding their customers.
We should have done a market sizing. We should have looked on LinkedIn and said how many people are in CSM. And later on I did, I looked and I was like oh my god it's only like a thousand. So if your software costs, I don't know, back then you'd sell per seat, right, a few thousand per seat, our TAM was like two million dollars. And so what we realized is actually for us to build a software company around customer success, we had to really focus on evangelizing the concept of customer success.
What that looked like is two core things. Helping companies say why do they need to invest in customer success as a concept, right? So helping CEOs and boards. Now it seems obvious, nowadays every company's like oh god, you got to retain your customers. Back then it wasn't so obvious. Many people lived in the world pre-SaaS where once you got a customer they would just stay with you for life whether they liked it or not, they couldn't switch. And so for a lot of these people it was brand new.
We spent a lot of time educating the market and then educating them how to do it. And that's everything from like building a post-sales organization, which obviously you're thinking a lot about now, and building a customer success team and how to compensate them. And we're just really lucky to be part of that story of that career going from something like a thousand people in 2013 to like today several hundred thousand people. So it's just kind of a job of a lifetime to be able to be on that journey with them.
Yeah, that sounds amazing. And how did you guys initially define CSM operationally?
Well, so the company that really gets credit for the term CSM is Salesforce. So credit where credit is due. Salesforce kind of created everything in SaaS right before SaaS kind of really was a thing. And so the story goes, we wrote many books on customer success, but one of them tells the story of 2004. Allegedly there was an offsite at Salesforce and they had sold a lot of software and they started seeing churn. And Mark Benioff, as the story goes, said we need a role to focus on really making sure customers are adopting the software, getting value, staying and ideally growing. And that was a CSM. He came up with the term CSM.
But in the early days there just weren't that many SaaS companies. So if you go back to that time frame, which I know is like a thousand years ago, it's like literally pre-ChatGPT if you can believe there was a world such. And you had Salesforce, you had Workday, you had a handful of other companies, Box, Marketo, Okta, maybe a few dozen materially sized companies. So the term was not widespread.
What you found is, and it's very tied to what you guys are doing now, people were doing things like the role of customer success manager but they were called different things: account manager, service account manager, support escalation engineer, all these different terms. And we'll talk later about FDEs, but there wasn't a common vernacular that everyone could kind of rally around.
One of the things I think we tried to do was say look this is a legitimate thing. We ran a big conference on customer success, still do, wrote five books on it, all just to try to make it like there's a standard here. Because I think a lot of times what happens is there's a new idea and everyone kind of independently is doing it, and this is what's happening with FDE. And then eventually somebody's like hey you guys are all doing the same thing, you should also call it the same thing, call it FDE, call it CSM. And then we were proud to play a small role in making that happen.
That's awesome. And also in the early days, were there any fun stories maybe that people don't know?
Oh so many, I mean so many stories. One story that I've told a couple times but not super bright is we just didn't realize this was going to be a big category. So we were like okay, seems like there's people doing this job, they have different titles. So we had this really crappy office in Mountain View above a bar in downtown Bay Area for folks that are listening from abroad.
Our office was like this really, I don't know what your office is like, but it was like the carpet was peeling off the floor. There was like one bathroom for everyone. It's just not a good office. No way in shape or form should anyone work in that office. But it was our first place and we were like hey we should throw an event to get CS people together, or customer-facing people together. And we thought let's throw this event.
We basically went to the local grocery store, Safeway since you're here, and just bought the cheapest wine and cheese we could and put them out on our conference table, which was a ping pong table. And just sent out an Evite and said hey if you're working in customer-related roles in SaaS you should come to this event. And we thought like 20 people would show up because Mountain View is not San Francisco, it's not as exciting. And this office was nothing to write home about. And the wine and cheese was not very good. So no reason should you come to this event unless you wanted to meet other people like you.
Literally it was like I don't know 75 people and went late into the night. And it was so cool because it wasn't about us. It was like people wanted to be together. And so basically every single thing we did subsequent to that was just people wanting to find other people like them.
And again, I think this is happening now in the FDE world where it's like gosh I'm reinventing the wheel myself. Are there other people like me? And part of it you want to learn from them. Part of it is you still don't want to feel as alone. Like you're like okay I'm not the only one doing this. I would always ask, we've done hundreds of events at Gainsight, and I'd ask people at the end like what did you take away from the event? And the biggest thing they would say is there's other people like me. I'm not alone. And I'm not dumb. Like I'm on the right track because they thought like they're just off on their own. Turns out everyone's doing that too.
It showed up in all kinds of crazy ways. Like we made music videos with our customers. Customers made music videos about customer success on their own. We had like a sing-along at one of our events, which is weird, right? How do you get a bunch of business people to sing along? It's because there was a community. And so that's the thing that's really powerful and I think an opportunity here is people want to connect with people like them.
Yeah. We take a lot of inspiration from how you guys built Gainsight in terms of the events and community focus. We actually just hosted our first event, 100 plus FDEs came. I think it's really magical what happens when you put a lot of like-minded people in the same room. I'm curious, that first event that you guys threw, what was it like? What happened when you put so many CSMs into the same room?
Yeah. I mean it was just basically like we didn't do anything, right? We didn't provide any kind of content or venue or anything and they just kept talking and talking and talking. And that same thing happened. The conference that we run every year is called Pulse. So kind of in the little CS world, it's pretty well known. And so we do that in the US and we do it in Europe.
The first Pulse was in San Francisco, May of 2013. And it was the same kind of thing where we had a ballroom that had like 100 capacity and like 300 people showed up. And I think somebody set off the fire alarm, which probably was one of our competitors, I'm assuming, trying to crash our conference. That's pretty logical. And literally we all had to clear out. But it was like one of those times it was so memorable, people didn't mind that the fire alarm went off. People didn't mind that the coffee ran out.
Lots of times when people go to a conference they complain about all the stuff, right, I understand. But the people didn't complain that much because they were like oh my god this is the best thing, I get to be with people like me. The vibe was that. Later in the later years we actually, one of our Pulse conferences, we made the whole theme Welcome Home because it's like a place where there's other people like you. And so that's what was really, it was truly a delight to do those events because it felt like you saw the human side of these business people.
Yeah for sure. And did you find that any of the CSMs also had differing experiences of how CS was run at different companies?
They did, but it's not like they thought they were radically different. But probably similar to your experience when you like break the atomic pieces out, it was like there was a few archetypes. Some of them were more technical, some were more salesy, some were more project-managey. Some of them might have had a commission type plan, some of them might have had a more traditional base type plan. Things like that were a little different. But there's a lot more common than different.
And then they could share that commonality and then over time you could kind of create different sub-rules. Probably it'll happen in FDE. Right now FDE is sort of like everything, but obviously some FDEs are more like consultant types versus engineers and so on. And the same thing started happening where there was a customer success engineering role, which is actually a lot like FDE, and then there was a CSM and customer success strategist. So there were some differences but actually I'd say a lot more was in common than different.
Gotcha. Yeah, that makes a ton of sense. Also you've said before that you guys were sort of forced to create a category, called customer success, because there were no customers.
Yeah, we had to find someone to sell to.
And you've also said oftentimes that you don't necessarily want to create a category if you don't have to.
Totally. It's a lot of work.
Curious what made you feel like you had to create a category here and also what it takes to do category creation.
In this case, it was funny because the idea, there were three of us who had the idea for Gainsight. Two guys were working on their own, I always had a similar idea, and then we got kind of introduced through an investor and that's how they all came together, Battery Ventures. And I'd run a SaaS company before Gainsight and they'd done a SaaS company. So basically we just felt this pain. We're like oh we should go solve this pain.
A lot of startups are just like somebody feels a pain in their last company and they want to go solve that in the next company. Very common. And so we hadn't thought through how big the market was and we should have thought all those things. We didn't at all. But we were just like we want to go do this. And then we by necessity were like we got to go create a category, I guess.
And it's like everything, it was much harder and took longer than we thought. Like it wasn't one of those overnight success things, but it was incredibly rewarding at the end. And so when I talk to new people about it, because I do get asked a lot, I think one thing is just defining what that means. It's like anything where that's too broad of a term, kind of like FDE. I think there's, okay, there's a new job and a new business process and there's new software, which is basically what Gainsight was, that's like the highest complexity level. I highly would not recommend it, but it was fun.
And then there's like oh no the job is the same but there's actually just a new type of software people could be using, right? That's a different kind of category. You know where you're like okay cybersecurity has been around for a while but like for example agentic security is a new thing. So that's like an example. You might be the same buyer but a new type of technology. FDE is much more like Gainsight where it's actually kind of a new role. So it's a little bit different, it's a little harder I think in that way because you're trying to evangelize a new role and a new software stack and things like that.
So one thing is defining like what do you really mean by category creation? Sometimes honestly it's not creation. You're just like I'm just the better version of that previous thing. Sometimes it's actually category aggregation. Like you look at what Databricks has done which is obviously like all-time great business success story, and what they've done now is aggregate so many categories in what they do. Same with Rippling, same with Ramp, like all these great companies where they're like you know what I'm going to take ten huge markets. Same thing with CrowdStrike, like pretty much almost every great company is like okay I'm aggregating categories, I'm not creating all net new categories, but each of them to the buyer thinks of as individual and I'm going to add them up together.
I think one thing you probably see from friends of yours who are entrepreneurs, before you would wait to go that category aggregation, multi-product, for years, and now so many people do it like in year one or two. That's a whole other topic.
Yeah, makes sense. And whenever you're trying to do something as difficult as category creation, I'm sure people were bundling you into other categories. And I think one particular example is for instance the opposite of maybe Gainsight, like Snowflake, where they were saying customer success is maybe not a function or everyone's involved in customer success. Frank Slootman, Chris Lynch, they had this saying. Do you think that's also a way to drive customer outcomes or do you think there's one right way to do it?
I'll give you context because Frank Slootman has a great book basically on running a company. It's a very good book by the way. Frank Slootman's legendary. I should say if Frank's watching this you're awesome. But obviously for people that don't know, he ran a company called Data Domain, sold it to EMC, ran ServiceNow, he ran Snowflake, and now Snowflake, truly the Michael Jordan of CEOs or whatever.
He writes this book and it's a very good book, Amping Up I think it's called. And I'm reading this book and I'm like oh that's awesome, high performance bonuses, and the chapter says something like don't do customer success. I'm like what? Why does this billionaire have to pick a fight with me? That's not fair. And so I think I met him once at an event and I have super respect for him. But obviously in that particular case I beg to differ. At some point maybe I challenged him to a cage death match. He never heard it. I just said in my conference I was like if you're watching Frank.
But let me make the case of both sides because I actually think it's worth intellectually honestly analyzing both. So one argument is you don't need a role called customer success management because everyone is in the job and it's everyone's job. And by the way, I think there's a lot of truth to that. And in some cases your company is set up such that everyone else is going to fill in those gaps. I can just tell you there's a lot of companies where they're not run that way and therefore there's a huge gap. The customer is sold to and then there's somebody to pick up the call if there's a support issue, but nobody's proactively working with them.
One challenge with customer success, because I'm always a believer you have to reinvent, is sometimes the CSM could only advise the client and they couldn't actually go help them configure the software. What's really interesting is FDE actually, I think, is potentially the next evolution because it's like I'm not just going to tell you what I recommend you do. I'm actually going to help you get that done. I'm going to help you configure it, change your prompts, whatever your harness, etc. I'm going to make this actually work in your environment.
So I actually think there's an opportunity to reinvent customer success as a kind of mega concept. You'll still have some of the roles, but I think we always said at Gainsight like there's customer success as the idea. It's like who doesn't want your customers to be successful so they stay and spend more money with you. Everyone wants that. Obviously you're a bad CEO if you don't want that. And then some companies you need a role to quarterback them. And to Frank's case, maybe some companies you don't.
Now funnily enough after Frank left ServiceNow they built a huge CSM team. So it's also probably a preference thing. Some CEOs want it, some don't. They run it differently. But my own view is that no matter what you call the role, you need to marshal everyone in the company around customers adopting your technology, getting value so they stay with you and grow. Nobody would debate that I don't think.
Yeah, makes sense. And you mentioned configuring software, right? Another profession you can call it professional services that you guys had a pretty big department there. I actually talked to Josh and Pete, the team there. Is there any parallels between professional services and FDE?
It's kind of funny because you find professional services, FDE, and customer success, and it's like you could kind of almost orient any of those three and add the other two in. They're all, there's a little bit of too much overlap if you have like a big engineering team and a big PS team and a big CSM team and a big sales team, right?
But I do think that professional services teams are incredible and they've been around for a long time. Sometimes they work for the software vendor and they're in-house, like for example Gainsight still has a great professional services team. And sometimes they work for a third-party consultant like Accenture or system integrator, etc. And they're implementing software on behalf of the client but they don't work for the vendor. So Accenture implements a lot of Salesforce for example. So they're two different models for professional services.
All those professional services kind of follow the same rough concept of you've got people that are managing the client engagement and project, sometimes they call them engagement managers, project managers. And then you have some of the people that are doing the actual technical work: technical account managers, technical architects, solution architects, there's a million titles for them.
I think the thing that I love about the FDE world, and maybe let me say three specific things. Number one is that in these new business models that are token-based, it's a more strategic function. So companies are investing more. They're just hiring. They're willing to pay more. FDEs make a lot more money. And PS traditionally it was like okay we want to minimize the cost there. And so by paying more, you're getting just amazing people working for these companies.
FDEs are not just about generating services revenue. They're going to drive consumption to your product, which is going to drive token spend and revenue growth. Arguably as important as a salesperson, I think you'd probably agree.
And then number two difference is that we kind of broke this old paradigm that you can't modify any code. Obviously FDE, who's the kind of godfather of all FDEs, can they modify code? And now each company's a little different. Some of them let them modify core code. Some you're just modifying some harness stuff or some prompts or whatever. But they're technical. They're not just configuring a UI. They're actually pretty technical in some of these companies.
And I think number three, which I love, is they're also bringing a lot of prescriptive guidance. The combo of deeply understanding the client's needs, the best ones I've seen, probably you're seeing too, is they really get to know that client super well. And they come in with a perspective on how this thing should be done in Decagon or whatever product they're implementing.
Those three differences, now why is that all possible? Well, companies are just realizing it's more strategic, so they're investing more money and more time. I think professional services teams would have loved all that, but they were often treated as more of a cost center and just how do we minimize that? And what I think is cool now is people realizing, look, software is kind of becoming a little bit more of a commodity, more of a cost center. And the value is actually in how you get it deployed, and sometimes we call that FDE.
Yeah, it's interesting that professional services world you were thinking of it as a cost center but FDEs, a lot of teams don't even think about how much they're spending. They're trying to hire as quickly as possible. And you mentioned that there are five landmines actually, you called it, that when you grew your professional services team to 100 plus people that you were running into. What were those five landmines and do you think we'll see these in the FDE world as well?
Yeah totally. Some of these I might not remember all five but this is a post I made from a month ago. There's some things that are going to be challenging. So I'll just rattle some examples. When you have a people-dependent thing unlike software there's some brittleness to it. So for example you are growing your company in sales and you have all these new clients coming on board and you always either have too many people in FDE or professional services relative to the new sales or too few.
If you have too many, you're burning more money. And if you're too few, your clients are waiting to get started. Either both of them suck. I actually heard about some company where every other quarter they stop selling so they can have capacity in their FDE team or whatever to deploy. And so that's not a great situation, but when you have people it's harder because you have to hire, you have to ramp.
Second thing that I think is quite challenging is you fundamentally have variance. At the simplistic level, put aside the non-determinism of LLMs, traditional software is predetermined, so therefore it works the same when you use it or I use it, and you put in one query into a database and I put in the same query you get the same results. And it's not true with humans.
So one thing that happens in professional services and FDE is it's very people-dependent. In fact, I think Jason Lemkin talked about this, who's a well-known SaaS investor for the one person that doesn't know that. He had this thing on one podcast, on the Twenty VC podcast, where he said they had an FDE that built some process for his company, SaaStr, and then they were on vacation and nobody can do anything because that FDE is not there. All the knowledge isn't there and it might be in that person's head.
This very people-dependent nature, it also means as a client you might have a great project if you have the great FDE, you might have an okay project if you get the okay FDE. And same thing was true in professional services. So variance is an issue. I think you run into like human issues like burnout. I mean gosh these FDEs are like well compensated but they're also working really hard and it's a tough job because sometimes if you're working with big customers you're going on site.
And if you fly every single week on site and you're trying to have a personal life or a family it's tough. And you're like as a CEO founder...
You're like, I want to get that client super successful, but I also don't want to burn my team out. We implemented a client early on at Gainsight, a big well-known software company, I won't say who. And they were so hard to deal with that literally my professional service person who loved Gainsight, she quit our company because she's like, 'It's destroying my life.' It was horrible, you know, and I could blame the client, but it's my responsibility too to shield her from that, right? So I still think that was something. There's a lot of stuff on the human side that's quite challenging, of just like how do you maintain consistency? Then you have this issue of you don't want to end up with one-off stuff that years later is hard to maintain. So you know you can imagine a world where somebody comes in and sets all this awesome stuff up and then like two years later it's like, okay, well who set that up and how do we maintain that? And now good news is LLMs are pretty good, it's better now because LLMs could interpret what it was, but there's stuff like that that we're going to deal with there. So I think there's just a lot of stuff that broadly is about how people are more complicated than software, and people are powerful and they're great, but they're not the same as software. They're not as fungible, you know, and so I think we just have to not kid ourselves that FD is going to be super easy, you know.
Yeah, makes sense. And the burnout problem especially seems to be happening a lot. Like FDs are the easiest people in the company. Oh my gosh. And I feel like that comes back to a lot of resource planning challenges, right? Like I know professional services teams, they have big resource planning software that they use to make sure that doesn't burn people out. I mean I'm sure it still burns some people out since it's just such a difficult problem. Do you see that kind of translating over as well to the world?
A hundred percent. Yeah, the resource planning is like a huge part of it. Obviously there's the resource planning itself, which I think traditionally was approached in very much like traditional project management, right, like assigning people, and obviously with agents and stuff you can do much better. But then there's the uncertainty of the demand, which is actually unfortunately the hardest problem in resource planning. So like if you knew exactly what your demand was going to be, you could be like, 'Okay, this person needs to be in Kansas next week. This person needs to be in Florida.' And you can do it all right, right? You could be like, 'This person's worked four weeks in a row. They need to take a vacation.' The problem is the demand is uncertain. It's spiky. You know, think about like software which is often closed at the end of the quarter and all of a sudden the next day everyone wants to go live, and then for like the month after that nobody has anything to do. And you know this is the stuff that I think is still going to be a challenge. I think AI can help a lot, but frankly like one of the things that we should all be prepared for is some of those challenges still might remain.
Yeah, for sure. And there's a lot of talk about software being a commodity, of course. You've said before that maybe the differentiator is actually experience, and maybe that's partially the post-sales work, part of it is services. Yeah, would you be able to elaborate on what you meant by that?
Well, I think some of this is process-limited. I feel like all of us on X every day we're like, okay, taste is the new moat, distribution is the new moat. We're just grasping for straws. We're like, is there a moat anywhere? I hope so. So I take a little caveat around that. But I do think that one thing I've thought about recently a lot is how for the customer, let's say not a tech company, like a regular real company that built something or sells something or services something, it's a real company, you know, not in Silicon Valley, for them it's so overwhelming, the number of things, the choices, the uncertainty. They read the headlines, they're not on X 24/7 like some of us might be, right? They're not obsessively looking at every new model, and they're like, 'I don't know what I'm supposed to do.' They see, okay, Anthropic has this stuff. I'm reading about these things. There's this company that does something. They all sound the same. And so I think there's some aspect of experience that's truly about getting their trust and then keeping it, right? Or earning it and keeping it. And so, how do you be that person that's like, 'Look, I'm going to help you figure it out.' And this I think is where the attractiveness of FD is, where you're like, 'Oh, I don't know anything about what to do, but that person seems really smart and they seem to care about my problems. They're going to come learn about me and they're going to go do something for me.' And you're probably seeing that all the time, even like VC firms we know and stuff, like we're like, 'Bring in an AI expert and just go figure it out for us,' right? And so I think experience or perhaps trust could be a differentiator as we go forward because I think people are really overwhelmed.
