Robert Morier: Welcome to the Dakota Live podcast. I'm your host, Robert Morier. The goal of this podcast is to help you better know the people behind investment decisions. We introduce you to chief investment officers, manager research professionals, and other industry leaders to help you sell in between the lines and better understand the investment sales ecosystem. If you're not familiar with Dakota and our Dakota Live content, please visit our website @dakota.com. Before we get started, I need to read a brief disclosure. This content is provided for informational purposes and should not be relied upon as recommendations or advice about investing in securities. All investments involve risk and may lose money. Dakota does not guarantee the accuracy of any of the information provided by the speaker who is not affiliated with Dakota. Not a solicitation, testimonial, or endorsement by Dakota or its affiliates. Nothing herein is intended to indicate approval, support, or recommendation of the investment advisor or its supervised persons by Dakota. Today's episode is brought to you by Dakota Marketplace. Are you tired of constantly jumping between multiple databases and channels to find the right investment opportunities? Introducing Dakota Marketplace, the comprehensive institutional and intermediary database built by fundraisers for fundraisers. With Dakota Marketplace, you'll have access to all channels and asset classes in one place, saving you time and streamlining your fundraising process. Say goodbye to the frustration of searching through multiple databases and say hello to a seamless and efficient fundraising experience. Sign up now and see the difference Dakota Marketplace can make for you. Visit dakotamarketplace.com today. Our guest today is Melody Koh. Melody is a partner at NextView Ventures, a seed-stage venture capital firm focused on what the firm calls the everyday economy, companies that meaningfully shape how people live, work, shop, move, and interact on a daily basis. Melody brings one of the more uniquely multidimensional backgrounds in venture capital today, spanning investment banking, venture investing, entrepreneurship, hypergrowth operations, and product leadership. She began her career in technology and media investment banking at Evercore before moving into strategic venture investing at Time Warner. After earning her MBA with distinction from Harvard Business School, she founded a venture-backed startup in the personalized wine subscription space, but before transitioning into operating roles in technology. Melody later joined Fab.com as a product manager before becoming one of the earliest product leaders at Blue Apron shortly after its Series A financing. At the time, Blue Apron was still an early-stage startup with a small team and an ambitious vision. Over the next several years, Melody helped scale the company through roughly 25x growth while building and leading teams across product management, product design, analytics, data science, all as head of product. In 2017, she joined NextView Ventures, becoming a partner in 2019. Today, Melody focuses on seed stage investing while also leading the firm's AI and data initiatives. Her work increasingly centers around how artificial intelligence is reshaping workflows, company building, product development, and the future of work itself. Public commentary from Melody and NextView reflects a strong focus on AI-native businesses, workflow automation, operational defensibility, and the growing importance of proprietary data distribution and customer integration as foundational AI capabilities become more commoditized. Her investing style appears deeply informed by her operating background. She is known for working closely with founders at the earliest stages of company formation, particularly around product development, scaling challenges, organizational design, and go-to-market execution. Melody earned her BS in Commerce with Distinction from the University of Virginia McIntyre School of Commerce, and she is here with us today on the Dakota Live podcast. Melody, thank you so much for being here and congratulations on all your success.
Melody Koh: Oh, thank you, and great to be here. Thanks for having me on.
Robert Morier: Yeah, it's a pleasure. You and I met about 4 years ago. We were judges on a pitch contest for, I think, Harvard undergrads.
Melody Koh: I think it was Harvard. I forgot it was MBA or undergrad, yeah.
Robert Morier: Yeah, I can't remember either. How many pitch contests have you done since then? I was curious about what your track record is like for judging those.
Melody Koh: I actually don't know. I try to excuse myself from those because those are very time-consuming. But we, uh, me and my partner Rob especially try to divvy up the work, uh, with, uh, Harvard community, you know, it's more of a community service and giving back. So, so both of us do a combination of, you know, venture office hours, judging and things like that.
Robert Morier: Yeah. If, if you had to put yourself back in that seat and you are now looking at yourself presenting, uh, what have you learned over the years that now in your opinion makes for a good fast presentation, something that really has some efficiencies?
Melody Koh: Very clear, strong statement about what a problem is, why is what we're solving, like, why is this problem important? Why should we care? Why are you the people who will have the best shot at solving it? And what the solution is. And I think a lot of times founders over, you know, I have this very prolonged description about the market and da da da da da, and then like on slide 5, you're still like, wait, what? What are we doing here? What are we solving for? Why is this important? Why should I care? So I think those are probably important pointers and just really quickly going through like, why is this a big problem and why do we care and why does it matter?
Robert Morier: I think about when you were at business school, you founded your own startup in that personalized wine subscription space of all things. Tell me Were you having a glass of wine and you're like, you know what, we need to create a subscription service for this?
Melody Koh: I mean, it's not exactly as that, but like somewhat close. I was reflecting on a personal problem that I had about going to a wine shop without making educated decision. And I felt like more people like me, a lot of people like me have the same problem. So, oh, why don't we try to understand people's taste preferences? And be able to send them wine that matched their taste. Despite having been a venture investor, a bit of a junior one before business school, I think the number one mistake there was, the market selection was horrible.
Robert Morier: So you knew there, there, there needed to be a better solution. You created one. What did, what did you learn through that process, particularly around execution and what founders actually need from investors?
Melody Koh: My, co-founder quit on me while we were in accelerator program together. So the number one lesson there was pick the right team. And, you know, now fast forward as an investor, I think there are actually research stats on this. The number one thing that kills a startup is founder dispute, departure, split up of sorts. So team risk and founder team risk is actually one of the biggest risks of early-stage startup building. Even if that didn't happen, I think the company still would have not worked, and that was because the market selection was horrible. I wasn't enough of a student of the market, I would say, in terms of understanding the operational complexity and the regulatory complexity. It is okay to have complexity. The problem just has to be worth solving like, I, I kind of think about a reflection post hoc that it is okay to go through a lot of effort to solve the problem, and that translates into your margin and OpEx and all the other stuff that you have to do, CapEx. But if the TAM is big enough, that is okay. And unfortunately, in this particular case, the problem was very, very complicated, but the prize was not big enough for a lot of other structural reasons. So I think that was, that was a mistake.
Robert Morier: You know, we always talk about people and the importance of people and the importance of teams. We're jumping ahead a little bit in terms of your current role, but how do you mitigate that risk? So how do you mitigate the risk of, of two co-founders or a team being able to scale at the rate that the investor is hoping for while not imploding?
Melody Koh: Your previous question actually was about what investors can be helpful with. And I think a lot of it is about team and coaching. Because, you know, the reason I brought up my, my specific experience with my own company back in the day, we really try to— when we meet a team, we're trying to understand, for lack of a better phrase, like, who is the alpha? Who is like, what is the dynamic of this team? Is it like equal economics and governors and voices? Is there a clear alpha CEO and then everybody else follows the leader, which is, by the way, nothing wrong with that. It's just a different dynamic style. But we, we want to make sure that we understand the dynamics, the skill sets, the personalities, and based on our experience, we can see that, okay, this will work, or there might be challenges very near term. And then, you know, other things we dig into are Do they have a shared history? Do they actually know each other really well? It is not required to have a shared history, but reflecting back to my own experience, I met my co-founder, you know, first week of business school. We're classmates, we're sectionmates. So we sat in the same classroom for the entire first year of business school and we had no shared history. I didn't actually know this person and we just teamed up because, you know, she seemed smart and And she was intellectually interested in it. And I think people don't realize unless they've started a company before, startups are hard. The brainstorming parts are fun, but execution and the grind is actually not that fun. And there's gonna be lots, lots of ups and downs. So I think having, uh, as an investor, having an understanding of the team's composition and why are they teaming up, like what are the underlying potential risks in terms of them understanding each other or not understanding each other? So that's kind of the, the pre-investing diligence that we really try to get comfortable, obviously aided by first-party interactions as well as references on and off list. Uh, post-investing, I think there's a lot about really being emotionally, like emotionally present as an investor. I, I feel like a lot of times our role is actually a coach or a therapist and not like an officially trained one, but because early stage is so dramatic in terms of its ups and downs. You know, I think half of our job in terms of support is actually being a steady emotional presence. And then the other half is more the tactical stuff, like, oh, here are the reps that we have about business and team building and stuff like that. I think on the post-investment side, like helping the team, you also mentioned scaling themselves, right? That's actually one of the biggest challenges. Because the jobs are very different. Uh, being a, you know, pre-product market fit founder trying to find product market fit to scaling 10x to scaling 100x to being a pre-IPO stage CEO to a post-IPO stage CEO, that's like 5 different jobs. And I think very few people can actually grow like that. And I think that our job to the best we can, ultimately is on the founders, but to the best we can, to provide resources and peers that are 1 or 2 steps ahead of them for them to really understand, oh, what good looks like and build a peer network, right? CEOs or founders, because I think unfortunately the only ones that who understands the founder journey are other founders. It is a very, very weird job because it's really not a job. So, so those are the types of things that we try to do pre and post investments.
Robert Morier: What kind of coach are you? Are you the kind of coach that holds your founders accountable? Are you the kind of coach that doesn't say much? You kind of let them find their own way? We hear a lot about coaching when we particularly focus on VC and earlier stage. But, you know, coaching means a lot of things to a lot of different people. And people, people receive coaching in a lot of different ways. Some people are very coachable. Some people are not. Some people need it softly. Some people need, you know, maybe a tougher hand. How do you, how do you consider yourself as a coach?
Melody Koh: My job is not to tell the founders answers because I don't know the answers. The best method is for founders to land on the answers themselves. To the best extent I can, I try to ask questions to get them to figure out what the answer is. Because when you're the one figuring out an answer, you're like, oh, I like that. That's my answer. As opposed to my investor told me so and so, so I must do it. Nobody likes to follow what other people say, I think. So I think that the better techniques is to ask questions and point them in some direction. I would say the other thing is I have, when I was an exec at Blue Apron, I actually received training, coaching, as a, you know, Blue Apron paid for coaching because I was a rising leader in the company. And it was very helpful for me. And one of the most helpful things that coach did was reframing a situation with a different metaphor to get me out of my kind of mental blocker. And I think that's another really useful technique to leverage when having conversations with founders. And because ultimately the founders know their business the best, the founders know the dynamics and the situations on the ground the most, like we can never replace that level of understanding. So I think the best we can do is to really bring in a different perspective that's not depriving them of their opportunity to think for themselves.
Robert Morier: While you're doing that, you are hoping that these companies are going to introduce more process in order to scale, but you also don't want that to curb or kill innovation. So, so how do you think about that balance as both an investor and then evaluating the leaders within these companies?
Melody Koh: I think process is a means to an end, meaning, right, what is a company? A company is kind of an organism, right? The, the, the inputs are Capital, which buys you people, and people then mix with how they work, which is called culture, and then the process and outputs, product, business model, etc. Process's job is to make sure your people can produce the stuff that you want. I think you want to really keep that lens as opposed to, oh, you know, I've heard companies of this stage or this amount of people do this, so thus we must. But I think it's like A lot of times you have to think about first principle. Why, you know, oh, people do performance review. Why they do performance review? Why they do it once a year? Why they do this every 6 months? What happens if we don't do performance review? Do we have other things that we do that achieve the same objective as performance review? I'm just picking performance review as an example. And the other thought I have is, which I think not a lot of people do, is every single process you introduce You're focusing on like, oh, what am I going to gain from the process? Oh, more orderly workers or visibility for me or whatever. But then I think I would think about what you're giving up. What is the cost of the process? Because everything has, you know, it's a double-edged sword. Oh, that means 30 minutes of people logging XYZ or— and then you have to think about is the cost is worth what you're gaining. And because if you don't do that, then you're very likely going to introduce a bunch of processes, and that compounds in terms of cost. And then you don't— it kind of runs away from you.
Robert Morier: Yeah, that makes sense. Is that skill, you know, understanding what you're giving up in order to grow and introduce processes, do you think that's one of the attributes that separates companies that scale successfully from those that tend to break under growth pressure? Is that recognition of, of the, of the opportunity cost of things?
Melody Koh: Um, I would say it's the people that you bring on. Meaning, you as a CEO— CEO is not going to be the almighty person who knows everything. In the very early stage, as if you are the CEO, you might task yourself with that job, like, oh, I must know engineering, product, operation, marketing, sales, blah, blah, blah. As you scale, your number one job becomes, let me hire really smart people who I think know how to do those things. So that I think that the judgment is actually an extension of who you hire, who you entrust, and what they tell you, what they filter to you, and what they're proposing. So, so I actually think, you know, The question on scaling, I actually think the number one thing is talent. Like, what is the talent that is inside the organism that is producing? And because the talent that you bring in directly impacts the culture that they can shape, or the existing culture can accommodate or not accommodate these people. And, and I think there's a lot of different philosophies around talent. I think with AI right now, you know, I think that the, there's a, there's a talk of the revival of the T-shape people and traditional execs, meaning traditional large company execs who are, oh, I, I know how to hire a big team and run XYZ process and blah, blah, blah. And then they haven't like been in the ground executing for 20 years. I don't actually think those people are gonna cut it. I think the people who are gonna cut it are both can operate, can operate at the strategic level. So like the T at the top, As well, they can just, they can jump in and be IC at any minute because with AI now you can do that very efficiently and you're actually gonna be managing a hybrid of agents and, and human beings as opposed to just human beings. And then funny enough, speaking of process with agents, you built agent scaffolding as opposed to human processes. So it's like the, the management technique's gonna shape and change, but we're deviating from your original question. But like, I, I, that's kind of, kind of how I think about scaling and people and talent and all that stuff.
Robert Morier: No, no, it's important. And you've written about this. You've written about scaffolding agents and how— so I'm curious, and we can get back, we will get back to where NextView Ventures is today and what you're focused on in what you call the everyday economy. But how would you explain the difference between a scaffolded agent and a standalone LLM like ChatGPT in this context of having to understand the root of this tea? The strategy is there, but now, you know, you've got to, you know, got to go up and down the pole, I guess, if, with that, for lack of a better metaphor.
Melody Koh: I like, I like using employees and management as a, I think it's a very good analogy when it comes to thinking about working with AI and agents. So ChatGPT, everybody has used ChatGPT. The ChatGPT experience, every time you have a new chat, I would analogize that as you are talking to a very smart 26-year-old who graduated from Harvard and maybe had a PhD of some sort. And, but every time it's like this fresh new person showed up at the door. He's like, never got trained by your, what your product is. And then you were trying to be like, in 30 minutes, really top to bottom, explain what you do as a job and function and company and strategy and priorities. And obviously that's like very taxing, and the very smart 26-year-old Harvard grad, uh, graduate might not actually catch all that. And then you have to do it again next time you, you spin a new chat session. A scaffolding agent I would— you would think it as the same raw talent but has been working at your company for 2 years and was in every single meeting, was taking notes. You've had weekly one-on-one with this person for 2 years. Every week you review what worked, what mistakes you made, and then you two discuss, okay, if you made that mistake, how would you redesign how you work so that next time you do the same work, this mistake hopefully doesn't happen. And then that process has been going on for 2 years, and that's a much more pleasant person to work with in the context of your company, uh, as opposed to, you know, the fresh person that shows up at your door every time you start a chat session. That's kind of how I would describe the, the visual differences when you work in these 2 different types of settings.
Narrator: Dakota Marketplace provides the industry with the highest quality and most up-to-date GP, LP, private company, public company data, as well as performance benchmarking and a number of other different datasets that the investment industry is fueled by. We're excited to announce that all of that data is available now via our customizable and bespoke APIs. You can create an API that will plug into any CRM or use some of our pre-built APIs with Salesforce, Snowflake, and a number of other CRMs that you can find on our website. To learn more about Dakota's API offerings and create a custom package that best suits your business, go to dakota.com.
Robert Morier: No, I think that's important, particularly as we think about what NextView Ventures is doing. So when you think about the importance of that scaffolded agent, uh, the ability for senior management, you know, to, to be able to go up and down that tee. How does that translate to what you're doing today at NextView in the context of these everyday economy opportunities that you are identifying?
Melody Koh: The articulation of everyday economy really is about mass market end user problems that are worth solving with technology. And the reason this framing hasn't changed that much since I joined the firm is because technology, again, is just means to an end. The, the key is we want to solve problems with technology leverage because it has a lot of implication about scalability, margin, and all that, interpretability. Now it just so happened that the technology now is AI, and most of the companies we look at are leveraging AI, solving every— everybody's problems. You know, in, in B2B, there are a lot of— obviously the labs are doing a lot of things, the foundational labs, Anthropic and OpenAI. But you, you think about what I just described as like the scaffolding versus the non-scaffolded version of it. The labs are what, what they're gonna do that, you know, for example, uh, Anthropic launched Agents for Finance. Now they have agents that are coached and scaffolded to be able to do DCF, discounted cash flow models, comps, pitch decks, and, and things like that, which we all, a lot of people listening probably recognize like, oh yeah, that's what I do. Part of it. But if you think about like, what is a specialized version of finance and what is the type of knowledge that's required to do the specialized version of finance? Well, pick accounting, you know, closing books, uh, closing books for generic random business X or physical business Y. The reason closing books in accounting hasn't been 100% automated is because there's so many exceptions. Oh, this expense. Is because of this context. So thus this should be categorized as that, or this should be categorized as this. All that stuff needs company-level data, industry-level data to properly train and scaffold so that you can imagine a startup or an independent company that is not one of the labs, if they're willing to spend the time and energy and hours to talk to 100 of these end customers to then get the data to design an agent that actually understand how to categorize expenses for mom-and-pop coffee shops. I'm just making up an example really well because they know exactly which of these filters for the coffee machines and beans there should be categorized. And so that 95% of that can be automated as opposed to 60%. Then that is a differentiated thing that actually creates value for those business owners that they will gladly pay for. That the Anthropic version of Agent for Finance does not do. But you can use the Anthropic version to be as a building block because you're like, okay, you know how to navigate Excel to start. That's hopefully like a good example for like a scaffolding in a business context, right? Like in a consumer context, very similar, right? We have a company that is, uh, building a chief of staff for parents and You can say that, oh well, if you are technical enough and enterprising enough and you can set up an Claude in a Mac Mini and give it permission to go into your Gmail inbox and your Google Calendar and train it to read all your kids' school emails, to parse out all the right things to pay attention to, like tomorrow is a rescheduling of the crazy hair day, and Friday you have to pick up 30 minutes earlier and then deliver it elegantly in text to you and your husband. Yeah, other people can do that. Why does the company exist? And the answer is most people don't do that because it's too much work. And a generic, if you ever, if you're listening to this like, oh, that's a good idea, and you try to do that by a one-off ChatGPT session, I guarantee you will fail. Because there's a lot of many other things that you need to do to again scaffold the agent with giving it a lot more context about you and your families and where they go to school and, you know, who drives kids on what day in order to deliver that experience in a complete package. And that's what the company can do, hopefully, if they execute well. So that's like a consumer example. Then, you know, briefly touching on like, how do, how do we, how do we invest? We are also a business. Venture capital is a business and it's knowledge work. And guess what? AI is here to take knowledge work. So a year ago we decided to be more aggressive about redesigning ourselves. And by that, I mean, I, I think I'll use hedge funds as a good analogy because I think this podcast obviously has a lot of audience across the the asset spectrum. 20 Years ago, there is no quant hedge fund. Every investor thinks themselves as a special snowflake in a, in a, you know, oh, I'm, you know, this is art and pick stocks. And I mean, I'm oversimplifying, but I'm sure that's, that's not exactly what hedge fund managers say, but I think that's like the, the, the, the narrative. And then now everybody has a quant strategy. But that doesn't mean they all do the same thing. That, that also doesn't mean that there's no human investors. The difference is now that the thing that can be delegated to machine will let the machines do it, and the human then move up the judgment ladder to do things that are hopefully non-consensus. And all the hedge funds competing are gonna be competing on the new non-consensus frontier. And, but the, the good news is that we're, You know, if the thing the machine can do more efficiently, we should let the machines do it. And I think venture is at a very, very beginning of this. I think most of my peers will say that, especially for early stage, most of my peers will say, oh well, but we are different. We don't have numbers. We invest in companies that don't have financials. So of course it's art and I'm an artist. That is lazy thinking because If you really think carefully of how you actually make decisions, of the 10 things that you mentally run through, there might be 5 that you can decompose and say, okay, I can let the machines do that if I set very clear, like, what is the success criteria of those? And then there might be remaining 5 that you still have to think about because models are not, for example, not as good at seeing the forest. They're very good at seeing trees. At understanding facts, but models have their own advantages in terms of processing large amount of data and they're reasonably good, no pun intended, reasonably good at reasoning today. I think two things that we believe is one, models are getting smarter. They're already pretty smart, and I think most humans are bad at understanding the exponential trajectory in terms of how technology improves. And then two, to my earlier point, we believe that this is What we do is a decomposable set of steps, and then we are better off spending our human energy on things that are truly very difficult for the models today. So that's kind of a little bit of backdrop of how we're thinking about redesigning, like if we're, as if we're starting a firm from scratch with nothing. Like we're very not precious about what stays and what goes, and that's kind of the, the, the motion that's Uh, you know, still early days, but that's, that's kind of the active motion that's happening inside the firm.
Robert Morier: So when you think about that company, the Chief of Staff for Parents, which, uh, I'm thinking out loud, I, I hope it can help pick up the picture frame that my dog ate last night at 2:30 in the morning. Uh, but as you think about a company like that, what are the attributes? And this is obviously consumer, so one leg of, of your stool, but when you think about a company like that, How, how is a company like that sourced? Um, and then what does the underwriting process look like as it relates to, you know, ensuring that that business is going to be, is going to remain durable in today's environment?
Melody Koh: One of the founders was introduced to us. Uh, so it was a classic venture network-driven, uh, sourcing. You know, I don't, I don't think in 10 years venture stays exactly that. Like, I think that would be one of the main motions. But if you think about the classic assumption of that sourcing motion is that there's natural limitation in terms of humans' networks. And, um, you, you, you don't wanna scale your team linear. I mean, a lot of firms actually do. That's why they have giant AUMs. You, you historically, the way to scale that is by hire more human beings to cover more network nodes. I would just say that I think that's not exactly what I think in 10 years that's not what we'll be doing for, you know, as majority of our motion. But that, that company today, that, uh, was an introduction by a co-investor. Product is very easy to build, relatively speaking. I'm not trivializing their product, but like, after I talk through it, you're like, okay, I know exactly how to build that. So building that is not a problem, and that is not a secret, you know, secret sauce of any, by any means. The, I think the two things that are enduring One is customer obsession. There is a difference of the how the product express themselves depending on how well you know your customers. Like, pick the parenting thing. You're a parent, I'm a parent. Like, we have all of our quirks and qualms about blah blah blah blah blah, like, you know, how we manage our households and children and this and that, and If you have two teams, one spoke with 50 parents or 50 households, one spoke with 1,000, the one that spoke with 1,000 probably will build a better product if they understand how to translate what they heard into the expression. And really, that doesn't mean they have to build everything that they heard from 1,000 households, but that just means that they have the ability to translate that and prioritize that in a way that The product is just like 10 to 20% better, more delightful, more aha, faster to value. Like those are all those little things actually matter. Because if you think about that particular type of product, once you onboard it, if it's pretty good, you're not going to churn. Like if it's delivering value every day, you're like, yeah, I'm going to— I don't want to go move to something else because the setup costs high. I feel like reauthorize my Gmail account and this and that. So, so I think customer product obsession is a category. And the other thing is you just have to run faster. Like assume every market is dynamic and, and, and competitive market is efficient, especially with AI right now. There are 10 teams going after any plausible sounding idea at any given point in time because everybody's getting the same information and inspiration about, oh, Oh, now the model can do this. Oh, now the agent stack can do that. Oh, here's like a problem on the ground that hasn't been picked up by the labs. Like you just have to assume at any given point in time, 10 to 20 teams going after the exact same problem and they're all good. They're all quality teams. I think the, there's no true defensibility other than the classic, you know, motions like network effects, but those are hard to come by, right? So you just have to be able to run faster, but what can you do to run faster. You, your team just has to be better. And especially I think the nuance right now in 2026 is that traditional definition of fancy teams, I actually don't think apply 100%, meaning, oh, you worked at Meta and then Google and then blah, blah, blah, blah, blah. So you're like highly credentialed. That doesn't mean you know how to leverage AI as well as someone sitting next to you who didn't get those credentialed experience. And the difference is that if you don't, if you two are given the same Codex or Claude subscription, but that other person knows how to set up Codex, so they run 10 agent sessions concurrently and doing work at night and is 10x the output as you, and that times 10 people on each company's teams, you're gonna, then is very clear who's gonna be ahead and who's gonna be behind. So I actually think that that's That is something that every operator and founder are, are, are honestly like, everybody is trying to figure out what are the types of people they hire? How do we know who is the right team? And, but I, I do think coming back to the original question, it is about velocity, but velocity has a lot to do with what is the team composition and how the internal culture is shaped, especially during this very dynamic time when I think we're still trying to figure out what is the new right way to work, because I think the past 10 years of right way to work is all getting shattered with the tools that we have right now.
Robert Morier: Let's say, um, you know, you do start to grow and you're scaling and you're thinking about these attributes that you believe are gonna define the next 5 to 10 years. How do you think about that in the context of your own business? So you, you may not be hiring today, but If the resume and the traditional pedigree is, uh, open to question, how would you then approach hiring a candidate in today's economy?
Melody Koh: Venture is a weird business. I, I almost don't wanna use us as an example, so I would use like the non-investor team of venture, right? Because that maps more classically to regular real companies, like, because venture is that your company by any means. It's, you know, we're, we're investment firm. So let's say non-investment team. I think the number one prerequisite is that you have, you have to know how to use AI. I mean, that's a generic sounding thing, but to be a little more specific, everyone has like, let's call everyone has a functional label. Like, oh, you're a marketer or you're a product person, or I don't know, your background is in ops. And if we're hiring for a functional role and we're comparing two candidates, If one candidate knows how to supercharge her expertise, let's pick finance, in finance, and that person can 5x her output, but she knows how to review the work so that it's not slop. She knows how to set it up so that she can do 5 things at the same time that are lower level. You know, let's say we're hiring a mid-level person, but we don't want to hire a junior person because Asians can be junior people, right? Like, why are we hiring? You know, so that's it. By the way, there's a whole separate problem in society in terms of the junior gap in terms of training, you know, if someone knows how to do that, that's 5x the person's output versus the other person who doesn't know how to do that as well. Of course, I'm gonna pick the 5x output person. And then you should take a step back. How do you identify this kind of people? A lot of people don't spend time or, you know, go deep with like the, the using, using AI, so to speak, or like I'm just gonna call it more bluntly, using AI agents, because everybody use ChatGPT, not many people use agents, is because I think they put a mental barrier in front of them, say, oh, I'm not, I'm not technical, I'm not an engineer, or I'm that, that's not me, like, I'm not a that type of person. And I think that's the biggest challenge for people to get it. It's actually not a real barrier. It's just a mental barrier. And the reason I say it's not a real barrier is because the programming language is English. So you don't actually need to know programming. You just need to like know what you want, clearly articulate requirements, be able to make trade-offs. When you don't understand, have the model explain itself. Okay, so this goes to the second attribute, which is you kind of have to be a student. You, you, I think you want, you want someone who likes being a student, who likes learning, because I'm learning every day. Like, I, I, you know, especially when it comes to using AI, like, things are moving so fast. New things are launching every day, capabilities, models, Harness, all that stuff. And, and I'm pushing the agents to do things that I wasn't pushing them to do 3 months ago. But now I'm like, okay, well, we did this, now we try that. Like, let me outsource this to you to see how you do that. And so you're— I think you have to— you cannot be doing this just because like, oh, my boss asked me to use AI, so thus I grudgingly do it. You— I think, yeah, I could— the personality type has to be a little bit like a learner who is excited about learning And then I say the last thing is I, I think you want a little bit more of the system-level thinker because end of the day, especially if you don't have management experience, when we talked about managing agents are kind of like managing people, you kind of have to think about you're orchestrating a system so that your system can be productive. And I think Honestly, I think that's a hard thing to interview for, but I think it's just through conversations to really understand how people think. It's not any of the hard skills, it's really about personality attributes because I think given the opportunity, you know, that, that type of person can show up and quickly tinker and pick up things and, and it's not for the sake of tinkering, it's really for, so you can be a better version of you at work. And hopefully you'll be happier too, because you don't have to do all the grunt work that you don't like to do. And that's kind of the point with using AI..
Narrator: For years, legacy data providers have made private fund performance benchmarking complex and expensive. That's why we launched Dakota Performance and Benchmarks, the first-ever benchmarking platform built by people who are using the data themselves every single day. We've made our benchmarking affordable, customizable, and very, very easy to use. You can log in to Dakota Marketplace today to start creating your own benchmarks and viewing our created benchmarks, or you can learn more and book a demo at our website at dakota.com.
Robert Morier: When you think about, uh, your, your LP partners, so the allocators that listen into the show, um, another way to ask the question that I asked you earlier about what you look for in, in a hire is what are the LP partners that you work with looking for in you? So what do you, um, present as, as your edge, as your competitive edge relative to other providers who are doing the same thing, or, or at least operating in, in generally a similar space. So what, uh, why does NextView get hired?
Melody Koh: You're right. Like there are probably like 1,000 C funds, if not more. Um, I think that landscape's changed dramatically since my 3 partners started firm 15 years ago. I would say the short version of that is I think we bring— so earlier I said there are some components of our job that can probably be automated. Or delegated to AI, but there's a lot that still needs to be top and bottom, um, human in the loop. And to author the scaffolding, you kind of need to know how to direct them all. Like, that's the key. That's why, you know, there are so many generic, oh, here's like an AI product for VC thing. We don't buy any of those because I'm like, well, I don't want your version. I, I wanna like, open the box and author our version because we have a lens in terms of what we're looking for, how we evaluate. And back to the employee analogy, if some really smart person shows up, we're going to try to train the person to be a principal that is a NextView principal as opposed to the generic other firm principal. What we do have is 15 years of on-the-ground institutional investing experience in the seed asset class since the beginning of C-stage investing. And that is hard-earned lessons and insights and judgments that we can then use to do two things. One, we can use to author for the stuff that we can actually increasingly, you know, innovate around in our own process. We can build a version that we believe is more robust because that combined— like, if you compare us to a manager that just started, I think I, I, you know, obviously he's talking our own books, but I believe we're gonna have more advantage because we've seen market cycles and invest in like 250-odd companies at pre-C and C stage. So that is the layer of judgment that we layer on top of the machine, right? And then two, you know, there are still some aspects of our job that requires, you know, the, the, the, the, see, you know, that there, there is some part of that requires obviously not only human interactions, but also ability to build trust and credibility and to guide founders. Like some of those very difficult conversations post-investments, right? Like those are still gonna benefit from the compounding experience that we have. So I actually think that, you know, my partner Rob has a good analogy. He's like, NextView is like a classic car. On the surface, but with like an EV powertrain underneath.
Robert Morier: That's helpful. I appreciate it. You mentioned this, the 15 years. How, how has seed investing changed in your opinion since you entered, uh, venture capital?
Melody Koh: I was a junior investor in 2009, 2011, um, not at NextView. This is pre-business school. I was at a corporate venture team, uh, at Time Warner. Back then, and NextView started 2010. First fund was 2011. So like I was, so it was roughly about the same time, you know, when NextView first started, there were probably like 10 or not even seed funds. Most of them are West Coast and most of them are graduated from super angels back in the day. Um, so NextView actually was one of the first institutional seed funds. With more of an East Coast base. So very cottage industry, very small. Seed was a million dollar round. Series A was 3. I remember Blue Apron Series A in 2013, 2014 was $3 million or something like that. $5 Million, I think. Who does $5 million Series A these days? Um, fast forward now, we probably have 1,000 seed funds. I think seed is definitely a mature class, and I honestly think that, you know, my partner Rob has authored a couple of blog posts about one. One was the scary one, which is the seed crisis, the crisis in seed. And then the second one is a more pleasant one because you offer some thoughts on solutions. Um, but there, there are structural dynamics that change, like exit markets, public markets. You know, if you're a sub-$5 billion market cap company, you don't get any coverage. You cannot survive as a publicly traded company. Like, so there are other dynamics around exits that have really put pressure on seed managers because seed managers used to be able to say like, okay, if we have a couple wins in the, you know, $500 million to like $5 billion range, it's 3 to 5x the fund. But If the exit environment is more challenged in that range, then you have to compete with the mega funds and say, okay, we have to get into the $10 billion and up outcomes. And that, I think that's, it's not, it's not that it's not doable, it's just a different strategy. And I think a lot of the C managers, including our peers and ourselves, are really rethinking like how to execute against the new reality. Notwithstanding, you know, at the entry there's, there's the competitive pressure as well. And I think that's why, you know, we're trying to be pretty aggressive in like rethinking how we do our jobs and how, how can we redesign, you know, our, our day-to-day motion because it's, it's, we don't believe that, oh, if we just continue to do what we started doing, you know, I joined a firm in late 2017, early 2018. So, uh, even since then, the market has changed quite a bit. Both in terms of valuation, round size, competitive dynamics. But I think we are all aligned on we cannot just be, you know, we should not rest on our laurels in terms of, oh, because this worked and early funds have strong returns because this is how we do things back in the 2010s, that's how we should continue to do things. I think that's something that's pretty clear. And I think the direction that we're going after is a little bit what I alluded to earlier.
Robert Morier: You mentioned round sizes. How has it changed in terms of fundraising objectives? So when you think about going out to market and you're raising for a fund, how has that process changed? Because there's no shortage of commentary on how difficult it is and has been for essentially everybody. It's—.
Melody Koh: It's—.
Robert Morier: I don't think— I don't think anybody's been able to, to dodge it. So when you think about it from your perspective, a large portion of this audience are also fundraisers. So there are people out there, you know, trying to convey a competitive edge or a reason for why they should be at the table. And when you think about in the context of the evolution of seed investing, how has that conversation evolved as it relates to fundraising and what the expectations should be of the LP who are investing in a strategy like yours?
Melody Koh: Yeah, I mean, I don't know if I have any like secret sauce or magic bullets, right? Like, what, what do LPs want? LPs want to make money. So you got to— I think you got to articulate a strategy that demonstrates the math and the assumptions on how you're going to, how are you going to make the returns. I would say that, you know, as it relates to Venture very specifically, right, we're benchmark— like, because of the DPI and liquidity timeframe, you know, at all said and done, we're benchmarked at a pretty aggressive return threshold in order to compete against other asset classes, right? You know, if we're holding periods no longer 10 and closer to 15, given how long some of these companies take to exit and show up back in, in distribution. And, you know, I, I think different LPs have obviously first, first of all, right? Like different LPs based on like who they are and what the institutions that they, they, they represent, they have different ways of thinking about, is it a, you know, a pie chart with allocation by asset classes? Is it, you know, generalist model, best ideas compete against each other, everything comped at an IRR basis, like, so everybody has like a different, and then, you know, your different operating cash needs for the organization that you work for. So I think everybody has a little bit different way of thinking about what is their return threshold and what do they care about. Uh, especially as it relates to like cash planning and cycle. But all that being said, early stage venture, like if we ought to exist, we have to deliver parabolic upside. I think it's as simple as that. And because otherwise we don't deserve the illiquidity because like illiquidity is, is a little too long. Now that being said, I think the other thing I think is especially for venture managers What you'll want to demonstrate is how do you engineer liquidity? I think we have a slide in our fundraising deck that's like, here is a wall of liquidity. And it's not just, we just sit around twisting our, you know, fingers and wait for a natural exit event. We had decent amount of secondaries that we've engineered, and especially because we're early stage investors that are not multi-stage funds, it is less controversial for us to exit. And so I think actually on a go-forward basis, like a lot of venture managers probably should have someone who is always thinking about liquidity. I mean, in, in private markets, that's pretty hard to execute in a perfect way, but having that framing and demonstrating how you've done it as a manager to try to be a little more active than passive. Engineering liquidities for your LPs. I think that's pretty important. And maybe like, that's like one tactical thing I would share that I think has worked because the market has changed so much. You also cannot just rely on like, you know, we're a firm with fortunately good track records of our earlier funds, but you can't just rely on that and say, oh, because of that, so thus you know, what we do on a go-forward basis is also going to work because the market dynamics on the ground is a little different. So I think it is very important to articulate what is the new kind of market situation in your asset class and what is your go-forward motion that is going to work, and like, why is that going to work, and what are the key things that you believe that will play out over the next couple years. I think that with AI right now, it is very, very tricky because honestly, I think like nobody knows, nobody knows a lot of questions. Nobody knows how large the lab's gonna get to, what are they gonna eat, what, where's the value gonna accrue to infra layer or the app layer. B2B, consumer, like all that. But I think, you know, if we think about like what we do, there's probably like 100 research that's shown that the most important thing about a company's outcome at the concept or seed stage is about the team. So I think, I think that's, that's kind of how it is like a constant because team is the core I probably talked about people and team like a lot in this, in this, in this conversation already, right? Like that's the constant that navigates the market situation over the next 10 years from the point of your investment. So like, that's the only thing that you can trust. And I think, think about us as managers, that's what we're trying to articulate in terms of how we're going to get to the best teams and how we're going to evaluate the best teams.
Robert Morier: No, I think that's wonderful. That's a full circle moment for a podcast episode, starting with the team and ending with the team. But I don't want to end just, just yet because we get a chance to talk about your market views and some of those opportunities, uh, that you actually write about quite a bit. Could you, can you tell our audience a little bit about Ground Truth? So where was Ground Truth born from and what do you, you know, what's the goal of sharing, you know, your views on what's happening in the markets? And then we'll ask you a couple questions around it.
Melody Koh: Yeah, as an investor, historically I, I wrote, but I didn't, I was not one of those prolific writers. Like my partner Rob is way more prolific writers. Writing comes very easily to him. And because I, I never really felt like I have a lot to say. Well, that sounds, that sounds— because I was like, well, you know, there's so many other investors are so smart. Like, I don't know that I don't have anything specific to say. And then because of the new scope I took on, I start working much more closely with AI every day. And, and I start having these moments where I'm like, oh, that's very interesting. Oh, that's very interesting. And then I realized, like, not a lot of people around me actually know what I'm talking about or know what I'm thinking. And then I was like, oh, I should just start. I should just start writing. Initially, I actually didn't have a very specific objective. I was like, oh, I'm just going to muse in public. And then I realized that this is— the seat I'm in is actually reasonably unique perspective because I think there are actually very few investors that have the, the privilege to spend so much time in AI. Um, because I have a hybrid role, like I don't do 100% investing anymore, right? Like I'm like building the machine that invests supposedly, but I've had like so many years of institutional investing experience. So I kind of have an investor lens and I have the builder lens current, and I also had a previous builder lens that is not building with AI. And building Insight like a real company as opposed to like a venture firm. So I have thoughts I want to share. And then so that I start writing. What I would say is that once I start writing, I found my conversation with founder to be a lot easier because a lot of times before this I'll have thoughts, but I'll have to regurgitate my entire thesis in a very short conversation. And now I'm just like, well, I wrote about that. Here, I'll send it to you. You can read it and then we will get into the pitch to talk about your thoughts on how you're attacking the market, given that you've read my thoughts. So I don't have to spend 10 minutes introducing what I think about what you're doing. Um, so, so that has been like a nice side benefit.
Robert Morier: What do you do with it? And you've talked about this a bit in your writing. What do you do with that company that's still doing nothing with AI? You know, particularly for very judgment-heavy work. Where do you think the hesitation is? Is it technical limitations? Is it organizational inertia? When you are faced with a company who you've sent one of your pieces to, and then you realize they're really doing nothing with it, what's the approach from your seat?
Melody Koh: Well, if it's a company that might be investment target, I would not invest. So, you know, I think But to, to, you know, what we talked about so much is how do you, how do you stay ahead in a very fierce competitive field as a startup? I mean, the prerequisite, I mean, I would say that even for companies that are building AI companies with AI products, they might actually not be as good at leveraging AI to run their company. Those are two different things. So, so when I'm diligencing or thinking about investing, I kind of want to get a feel for how strong they're going to be in terms of building the execution flywheel for their own organization, because that is the velocity that's going to come out from that company. But to answer your question on like, what's the hesitation in general, right? Like, I think it's a combination of was a couple things. One is, if you have not used the higher— like, I have the blog post called, uh, The Four Relationships with the AI, and it talks about like the four levels. If you're only being on level one, meaning single ChatGPT sessions, you cannot open your mind to what's possible because you have not seen what's possible. And if you've been to level three or level four, that's all of a sudden every day you're seeing opportunities on the floor. I'm like, oh yeah, that can be done, or that can be Got like, I think it's a little bit of the, the awareness limitation if you're not seeing the higher level of what's possible with the frontier right now. And I would say bluntly, like the frontier right now, if you not use a, unfortunately if you not use a coding agent like Claude Code or Codex or Cowork, that is not the frontier, right? Because agents is where the frontier is. With the, with the agent harness and the labs have delivered, and unfortunately, or fortunately, I don't know which way, coding agent is the most advanced form of agents because that's the first mature market. But don't get intimidated by the name a coding agent because you can use it to do a lot of different things. And that's what Anthropic's trying to do with Cowork, for example. And I'm sure OpenAI is gonna come up with something soon. But I think like the awareness limitation is one. And then obviously all the other stuff that you said, inertia, You know, pick venture capital firm as an example, classic investment firm of any sort, right? A lot of your audience manager side, if everybody's an investment GP, our incentive is not to build systems for the rest of the company or the rest of the firm. The economic incentive does not align with doing that kind of work because historically investment firms are built with individual manager and GPs delivering a sleeve of track record doing whatever they can. So at best, I think you see individuals pushing their individual innovation limits and try to get ahead of other individuals. But I think it's very hard and very rare for a firm-wide version of that.
Robert Morier: That's helpful. So with all of that in, in context, what are the areas of technology or consumer behavior that you think are going to remain the most underappreciated in today's market? What do you think is being overlooked that people should be spending more time with?
Melody Koh: Consumer has not been a very hot sector for venture for past, I don't know, 6, 7, 8, 6, 7 years. And I, you know, I think that Consumer, I think there is an interesting potential opportunity for consumer because all the focus and energy has been B2B enterprise AI adoption because that's where the nearest, the clearest near-term dollars are at. Oh, make worker more efficient, replace the worker budget, right? Charge a third of the worker salary. But I think that this generation of AI has a lot of attributes when it comes to new form factors, new interaction modalities, especially with agents. I think that nobody has really cracked the, like, what is the, what is the Claude or personal pet agent? For individuals, right? I'm not saying exactly that's like the form that would exist, but I think that's like one example to illustrate. Um, and, and I think there's a lot of like surface area for exploration and tinkering. It, it, it— I almost think that the early Web 2.0 version of tinkering hopefully will come soon, and that there's just all sorts of weird stuff happening back then, like Twitter, Tumblr, like people were just tinkering and building weird, seemingly weird form factors. I think the key, you know, now that being said, like one of the key challenges, consumer, is the distribution has been locked down for a while. The growth, the, the, the early growth of virality engines in the 2010s that are close to free no longer exist. But I think there are gonna be ways to circumvent that. Hopefully. So, so I think that's probably an area that from like where we sit in, in terms of where we are in a, in a asset class spectrum, that's an area. If you like pick, you know, venture investor XYZ, a lot of them used to focus on consumer 5 years ago, but they all haven't done anything consumer for the last 5 years. So I think that's probably one area that is under probably underappreciated compared to market consensus.
Robert Morier: That's helpful. Thank you, Melody. Uh, before we close, I do have one more question. You've had some unexpected appearances. I saw on Cash Cab and Top Chef. These are, these are the hard-hitting questions that we wanted to ask you in the beginning of the conversation, but we saved to the end. Um, how did that happen? How did those happen?
Melody Koh: Cash Cab was actually pretty random and fun. I mean, I don't even know people know Cash Cab, right? The show doesn't exist anymore. I think I think.
Robert Morier: Uh, well, yeah, yellow taxis don't barely exist.
Melody Koh: So, but, um, I was, I was a junior VC in New York City in 2010s, and back then there were probably like 30 junior VCs in town total. Um, they're probably like 800 now. But if it was, we one, uh, someone from, from that crew will organize like a monthly dinner. So we're at this barbecue place in Flatiron, and then 4 of us were walking out, we're trying to share a cab. Going like Hell's Kitchen and up to like Upper West. And then we hailed a cab and we walked in. This is me plus 3 other junior VCs. You know, a guy from Bessemer and then the guy at Franco R.E. And then another guy at like IA Ventures. And then, and then me. And then we walked in and the thing just like, welcome to Cash Cab. We're like, what? This is, this is real. This really happens. And then they're like, oh, well, okay, so this is This is Cash Cab. If you guys want to do this, where are you guys going? Let's make sure we have 20 minutes because they need enough time for trivia questions. And if you want, please go walk out the cab, and we have to properly film you walking in, but you have to act as surprised when the lights go off again because we have a car following the cab to film the walking into the cab. So actually, it does— it did randomly happen. Like, it was not staged.. And then we had to sign NDAs and we're like, VCs don't sign NDAs, but we had to sign NDAs because, you know, they're trying to— it's just for the show. And then we actually, 4 of us won, I don't know, like $300 each after, at the end of it, pre-tax. Yeah. Uh, so that was, that was pretty good. Um, Top Chef was slightly more boring. I was at Blue Apron and we, uh, were partners with, uh, Top Chef. Uh, the company was part of Top Chef, so each season we'll get a winner. And we'll make a recipe and put it in a box. And my, my buddy was the head of marketing at the time and she's like, oh, they're filming the Restaurant War episode in Denver. We can send 2 people. Do you want to go with me? I was like, I was an obvious— I was a cooking show fan. I was like, yeah, of course I want to go. So 2 of us went. And for those who don't know or have not watched Top Chef, there's always like an episode called the Restaurant War. They basically whittle down the contestants to about, I would say, like 8 people. Per team and then divvy them up. They have to open a restaurant in 24 hours, including front of the house, back of house, design, menu execution. And we're the patrons, the fake— well, fake patrons who got allocated to Restaurant A or Restaurant B to eat, but we actually can eat. And then the cameras will be going on around in the fake restaurant and they'll capture people's face. I literally was on TV for half a second. And me and my friend were practicing what we're going to say that sounds really smart as a commentary about the food. And I remember I said one sentence, it was like half a second of a frame, and that was gone. So, so that was a, that was a Top Chef experience.
Robert Morier: Well, you were well prepared.
Melody Koh: We were like, oh, the camera's going to come to us. We have to say something. We have to be prepared something really smart and, and see if we make the cut on TV. And we made the cut on TV for like a second each.
Robert Morier: Yeah. Thank you for spending this hour with us on the Dakota Live podcast. It was wonderful to hear more about you, more about NextView Ventures and what you're thinking, particularly in terms of the work that you're doing within artificial intelligence. It was highly informative for myself, I'm sure for our studio audience as well, our students from Drexel University that were here visiting as well, and one Rutgers student. Thank you, thank you for spending your time.
Melody Koh: Yeah, thank you for having me.
Robert Morier: If you want to learn more about Melody and NextView Ventures, please visit their website at nextview.vc. I also highly encourage you to look at Melody's Substack, which is The Ground Truth, that we talked about in this conversation. Highly informative, wonderful insights that I regularly follow. So I hope you do as well. Thank you, Melody, for being here. And to our audience, thank you for investing your time with Dakota.