Feb 14, 2024 · 1h 11m · news
Martin Gontovnikas (Gonto): The Biggest Mistakes Startups Make When Scaling into Enterprises | E1115 · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of 20VC, growth marketing expert Martin Gontovnikas joins Harry Stebbings to outline the mechanics of product-led growth, detailing how to execute high-impact big swings, design high-retention onboarding experiences, and successfully transition from a bottoms-up motion to enterprise sales.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 17.2% of the talking time here. How this is scored →
Harry as informed peer 3.7 Guest teaching 6.1 Guest disagreement 1.8 Harry pushing back 2.2
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 17:20 Refusing pre-PMF growth hires
Gonto forcefully rejects Harry's suggestion that startups might need growth hires pre-PMF to gather data, asserting he would never hire for growth before product-market fit is established.
Hardest push from Harry ▶ 8:12 Challenging A/B testing limits with media thumbnail dataHarry directly pushes back against Gonto's assertion that data doesn't tell you what to do by citing his media company's A/B tests on thumbnails yielding clear click-through winners.
Biggest teaching moment ▶ 1:03:40 Revealing email holdout test showing 20% higher retention without emailsGonto shows how relying on high email open/click rates was misleading, revealing that a holdout group receiving zero onboarding emails actually had 20% higher user retention.
Harry holds his own ▶ 44:15 Calling out false enterprise expansion timelinesHarry demonstrates deep VC knowledge by exposing portfolio founders who justify high enterprise CS costs with low SMB revenues under the promise of future PLG expansion.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Gonto's Unconventional Entry Into Growth Marketing | 2 | 4 | 1 | 1 | Harry asks a standard introductory question about Gonto's entry into growth marketing. Gonto shares his journey from developer advocate at Auth0 to leading self-service growth, explaining how psychology and big risk-taking drive exponential results. | |
| Lesson 2: Taking Big Bets for Exponential Growth | 3 | 5 | 1 | 1 | Harry asks insightful questions on willingness to take risks and managing failure. Gonto reframes bravery as taking action despite fear and outlines budgeting rules for small vs big experiment risks. | |
| Qualitative Post-Mortems and Searchable Databases | 3 | 6 | 2 | 2 | Harry prompts Gonto on how teams practically learn from failure. Gonto emphasizes qualitative feedback over quantitative metrics and defines growth as applying the scientific method to KPIs through creative intuition. | |
| The Role and Limits of A/B Testing in B2B | 5 | 7 | 3 | 4 | Harry challenges Gonto's stance that data doesn't dictate actions by citing his media company's thumbnail A/B testing wins. Gonto counters by distinguishing B2B from B2C data dynamics and explaining how local maxima miss exponential swings. | |
| Time-Bounding Big Bets: Auth0 Outbound Case Study | 3 | 6 | 1 | 2 | Harry asks about time-bounding big bets. Gonto explains how qualitative user interviews revealed why Auth0 outbound failed, leading to a scrapers-based initiative tracking strategy. | |
| Amazon's Approach to Innovation and Big Bets | 3 | 5 | 1 | 2 | Harry inquires if large or public companies get too big to take big bets. Gonto uses Amazon's Kindle Fire failure leading to AWS as proof that isolated teams must take big swings. | |
| No Growth Hires Before Product-Market Fit | 4 | 7 | 4 | 3 | Harry pushes on the chicken-and-egg problem of needing growth hires before PMF to collect data. Gonto forcefully rejects hiring growth pre-PMF, asserting founders must close design partners first before opening PLG floodgates. | |
| Guided Product Tours vs. Autonomous Onboarding | 4 | 6 | 2 | 2 | Harry asks detailed product questions about product tours vs empty dashboards and handling horizontal software like Notion. Gonto details Auth0 A/B testing on developer onboarding preferences and template-driven user profiling. | |
| The Art of Bottoms-Up and Proactive Customer Support | 3 | 6 | 1 | 1 | Harry brings up blending bottoms-up and top-down PLG motions. Gonto explains detecting user friction patterns in dashboards and Vercel's strategy of hiring Apple Genius Bar grads as Product Advocates. | |
| Clustering Usage to Drive Freemium Upgrades | 3 | 7 | 1 | 1 | Harry asks how much value to offer in freemium tiers. Gonto details using usage clustering algorithms to design pricing tiers and shares Auth0's counterintuitive finding on multi-factor auth anti-retention timing. | |
| Portable User Loyalty and Core Value Retention | 3 | 6 | 2 | 2 | Harry asks about key retention metrics and hard paywalls in consumer apps. Gonto rejects hard paywalls, emphasizing core value retention through personal developer projects and quarterly over-quota billing. | |
| Blending Bottoms-Up and Top-Down Enterprise Sales | 3 | 6 | 1 | 1 | Harry asks where startups fail in top-down enterprise execution. Gonto explains combining motions by running targeted ads for engineering managers using availability bias when ground-floor developers adopt the product. | |
| Crafting Specific and Value-Aligned Internal Pitches | 4 | 5 | 2 | 2 | Harry shares a founder story and asks for granular advice on internal pitch messages. Gonto emphasizes extreme specificity over marketing buzzwords like 'seamless' and tailors collateral to enterprise analysts vs startup needs. | |
| Demos, Internal Advocacy, and Whiteboarding | 3 | 6 | 2 | 2 | Harry asks about effective product demos and vertical playbooks. Gonto recommends arming internal champions to demo products themselves and deploying enterprise architects on sales calls for whiteboarding sessions. | |
| When to Scale Into the Enterprise Market | 5 | 6 | 3 | 4 | Harry challenges the necessity of enterprise scaling by citing HubSpot's $28B SMB model. Gonto counters with unit economics and revenue concentration data from Twilio, Atlassian, and Auth0. | |
| Active Account Expansion vs. the Self-Selling Lie | 6 | 7 | 3 | 5 | Harry calls out portfolio founders claiming eventual PLG account expansion despite high enterprise costs. Gonto agrees that relying on organic expansion is a lie, illustrating his point with New Relic's feature bloat downfall. | |
| Understanding the Psychology of Enterprise Buyers | 4 | 7 | 2 | 3 | Harry asks what enterprise buyers prioritize. Gonto dissects buyer psychology around fear of getting fired vs desire for promotion, introducing the concept of a Chief Technology Marketing Officer (CTMO). | |
| Aligning Product and Marketing Teams in Practice | 3 | 6 | 1 | 2 | Harry asks how to practically eliminate silos between product and marketing teams. Gonto shares Auth0's onboarding conflict over sales contact prompts and how role-based segmentation resolved it. | |
| Avoiding Expectations Mismatches and Misalignment | 3 | 6 | 1 | 1 | Harry asks where product-marketing alignment fails and what profile to hire for initial growth roles. Gonto highlights marketing overpromises and defines 'hunger for glory' as the key driver in scrappy first hires. | |
| The Reality of Growth Frameworks vs. Scrappy Execution | 3 | 8 | 3 | 2 | Harry asks about case study testing in interview processes. Gonto debunked popular Reforge growth frameworks as post-hoc narratives created by winners, advocating abstract interview exercises assessing problem-solving process. | |
| Avoiding Clone Hiring and Embracing Cognitive Diversity | 5 | 6 | 2 | 2 | Harry candidly shares his own hiring weakness regarding logo bias. Gonto reveals his mistake of clone hiring, his rule of hiring someone he personally dislikes for cognitive diversity, and Auth0's $300k wasted ad campaign on unactivated signups. | |
| The Danger of Gamified Email Metrics | 4 | 7 | 1 | 2 | Harry agrees on the dangers of gamifying metrics. Gonto presents an experiment where a 10% holdout group receiving zero onboarding emails retained 20% better, prompting them to kill automated email sequences. | |
| Custom Engineering in a Commoditized AI World | 5 | 6 | 3 | 4 | Harry challenges Gonto's assertion on AI outbound complexity by pointing out commoditized off-the-shelf tools like HeyGen and Synthesia. Gonto counters that off-the-shelf automation rapidly decays in effectiveness without custom engineering. | |
| Growth Timings and Founders Over-Relying on 'Magic Pills' | 4 | 6 | 1 | 2 | Harry notes that founders falsely view growth as a magic pill. Gonto discusses timing mistakes and shares Clerk's successful YouTube growth tactic of embedding organic product usage rather than dedicated sponsored spots. |
Statements from this episode (56)
Gonto: Startups pivoting to enterprise fail if they abandon core users
“If you forget about who your users were, you eventually go out of business.”
Gonto: Psychology is crucial when marketing to software developers
“Even when you sell to developers, psychology is key to marketing because as Daniel Kahneman says, all decisions are emotional.”
Auth0: Story-focused marketing emails converted better than data-heavy emails
“In a lot of cases in the beginning, we were using emails with a lot of data and we actually moved them to have almost no data and just focus on making sure that the story was coherent and those converted better.”
Gonto: Employees can run small, measurable experiments without approval
“One thing, for example, that I'm doing now is if you have an experiment and you have an idea, you want to try it out. It's well-framed. You've done the research, you understand, you have a gut feeling, but you have a way to measure it. As long as it's less tha…”
Gonto: Fire employees who fail without learning, but protect those who iterate
“When somebody fails at something. You do not give them shits. It's more about how you learned anything. If they haven't learned anything, you can fire them. But if they learned something, that means that at least they can iterate.”
Gonto: Qualitative data matters more than quantitative data in growth
“I'm a big believer that qualitative information is a lot more important than quantitative information in growth.”
Gontovnikas built a custom GPT to query his past failure learnings
“Lately, for example, one thing I've done is all of the learnings, I have them in my Google sheets with What did we learn? Some tags and stuff like that. And I created like a GPT from it.”
Gonto: Creative growth leaders will always beat purely data-driven marketers
“Data does not tell you what to do. Like that's why growth people who are creative, who thinks about unique ways to drive exponential growth are always going to be win and are always going to be much better than anybody else that is out there.”
Gonto: B2B startups must test big swings, not small variations, for fast results
“For B to B, if you want to get results, you need to try something that is completely different because then you will get results a lot faster than if you do it in any other way.”
Gonto: Target six small, two medium, and one big bet per quarter
“So what I think about that is about planning on, okay, in the next quarter I do six small bets because they take me very little time, two medium bets, and one big bets. I do the small bets so that those give me an incremental increase that allows me to continu…”
Gonto: Internal tech stack initiatives are stronger outbound triggers than industry verticals
“We realized that the initiative mattered more than the industries. So for example, if a company's upgrading the technology stack, moving from JavaScript to Java, then at that point, it actually made sense for them to think about the authentication provider.”
Gonto: Amazon would never have launched AWS without trying the Fire phone
“And if they weren't willing to try Kindle phone, they would never be willing to try AWS.”
Gonto: Interspersing Growth Teams Across Organization Never Succeeds at Scale
“Ideally you instill it around the org. I have never seen a place where that is a success yet.”
Gonto: Track growth team output with sales-style pipeline leaderboards
“One thing, for example, that we've done To make growth people be competitive is build dashboards similar to what you do with account executives, where instead of having a quota, you actually have growth person and how much pipeline they created. And you get th…”
Gonto: Startups should never hire growth roles before product-market fit
“I would never hire somebody for growth before product market fit. I think that before product market fit, it's the job of the founder to get to see if there's product market fit. If the founder can't do what needs to be done to find product market fits, I thin…”
Gonto: PLG startups should launch closed with six to eight design partners
“I think that a product growth company should always start with design partners. You should always start closed with six to eight customers that you can find through your network, your business network. You get to better understand, get closer to PMF, at least …”
Gonto: Blank dashboards without onboarding destroy PLG first impressions
“For example, you don't do any onboarding and you just drop them in a blank, like blank dashboard with no data, even though it should have data. That is always like a very bad impression. You can't get out of that.”
Gonto: Pushing advanced features too early in PLG causes user churn
“Another one is, for example, you get them to implement features That you want them to use because they are the ones that are going to pay the most, but you get them to implement them too early before they re-understand the product. And you have to slowly get t…”
Gonto: Junior developers prefer step-by-step tours; senior developers find them condescending
“More junior developers wanted a point and click, like tell me here, so I can just get done with it. Somebody that was more senior and had more experience felt that that was condescending. So they wanted somebody to explain what are the big boxes? What are the …”
Gonto: Template libraries matter most for user intent data, not utility
“So that's why I think building a template engine and templates library is key, not because of the templates that the user will receive and how fast they will act on it, but because they self-service, self-servicely give you a lot of information about who they …”
Gonto: Traditional SDR outreach to developer users never works
“Like if you are doing a product for developers and you have an SDR reach out, it's never going to work out. The developer doesn't want to talk to that.”
Gonto: Vercel deployed Product Advocates to unblock users before sales
“So for example, one thing we did at Vercell was build a team that's called Product Advocates. That team of Product Advocates is literally people that are sort of technical that will, when somebody gets blocked in the dashboard, you'll contact them and hit them…”
Gonto: Vercel hired Apple Genius and bootcamp grads as Product Advocates
“For example, what we did was we hire people from Apple Genius Bar or Best Buy Tech Support or people who did like Lambda School or some CS Bootcamps.”
Gonto: Drive Freemium Upgrades by Withholding Key Features From Tier Clusters
“Once you see that, what I typically try to do is each of those clusters is one of the plans, but I always remove one or two features that are very useful from one cluster and move it to the next as an excuse for people on like, okay, I start to use the product…”
Auth0 users setting up MFA in week one churned, while month-one setups retained
“Anybody who implemented multi-factor authentication in the first week stopped using OutZero. Anybody who implemented multi-factor authentication between month one and month one and a half Was very likely that they would be retained.”
Startups should map feature adoption timing between retained and churned cohorts
“We actually built a dashboard of people who after one year were still retained in the products and comparing them to people that after a year were not retained in the product. And we built a table that was like all of the features and when did I implement it i…”
Gonto: Personal project adoption drives portable, career-long user retention
“If you get the user to use you for a personal project all the time, that means that whatever company they will go throughout time, they will always remember you and try to use you.”
Gonto: Hard paywalls should not exist; users must try features first
“I don't think that that should exist anymore. Like the idea of a paywall without you using it. I think it's a very, it's very hard to get people to experience things. I'm a big believer on experience. So it's like, okay, if you explain me a feature, but I cann…”
Targeted brand ads build executive availability bias to aid bottom-up sales conversion
“We start doing ads for directors of engineering, directors of product, VPs of engineering and VPs of product. We don't do the ads to them because we want them to click them. I don't care if they click them. It's more so that they see our brand and they see the…”
Gonto: Startups ignore Gartner reports while enterprise buyers rely on them
“If you're selling to an enterprise company, I think what Gartner, Forrester say matter because they are like the analysts. So in those cases, sending something like a Forrester total economic impact, like the ROI white paper is very useful. If you're selling t…”
Gonto: Arming internal users to demo software beats sales-led demos
“If you can arm the user to do a demo internally, I think those things are really, really good. Where instead of doing a demo with a sales team, sometimes what I, what we've done is give them something that they can demo that is already packed up and it explain…”
Gonto: Organize sales playbooks by use case, not industry vertical
“I don't think about industries because sometimes three industries have the same use case. So that's why I think grouping them by use case makes more sense than just grouping them by a specific industry.”
Stebbings: HubSpot built a $28B business through SMBs alone
“HubSpot have built a twenty eight billion dollar company through SMBs alone.”
Gonto: 80% to 85% of top PLG revenue comes from enterprise accounts
“If you look at all of the companies like Atlassians of the world, like the Twilios of the world, Outsido, or even Bursell, most of them, what ends up happening is that 80, 85% of the revenue comes from mid-market and enterprise, 15% come from SMBs, but that 15…”
Gonto: Building a company solely on SMBs is harder than enterprise
“I do agree with you that you can potentially create a business from SMBs. I just think that in a lot of it's a lot harder to be honest, because enterprise or mid market, you don't need that many customers on one side.”
Gonto: Enterprise accounts taking over a year to expand likely never will
“If they take more than a year, I would say to expand to a really big deal, it's likely never going to expand.”
Gonto: Auth0 expanded $30k initial deals using DevRel internal workshops
“At OutZero, a lot of times we had a very small land, like, thirty-k land. But then, we actually asked our champion to invite other people from inside the company that could potentially use us for a workshop. We then sent our DevRel team there to do interviews,…”
Gonto: Enterprise Checklist Bloat Destroyed New Relic's Developer Adoption
“What they didn't realize is by adding all of those features to hit the checklist, they added a lot of bloat in the product. That bloat made the experience be a lot worse. When the experience became a lot worse, developers stopped to use it.”
Gonto: Enterprise buyers care most about career preservation and promotion
“To be honest, I think there's two main things that somebody who is an enterprise cares about. Number one, I want to buy something and not get fired because of this. I want to keep my lovely, beautiful job that is nine to five and I don't need to do much more a…”
Gonto: Analyst reports and compliance serve as enterprise buyer insurance
“Why do they care about security and privacy and compliance and Gardner and Forrester? Because it's ways to prove that if the product doesn't work or if I made a mistake, at least my mistake was validated by this Gartner and this Forester and these things.”
Gonto: PLG Startups Must Align Marketing and Product Over Sales
“The world where we're moving to with a, with more product growth, with more bottoms up, with more people trying new things. I think the links that are the most important right now are marketing and products together, not sales and marketing.”
Gonto: In PLG, the Product's Job Is Proving Marketing Claims
“Because when you think about marketing, What marketing is doing on their website is they are promising what your product will be about. Then because there's no sales, people will click on sign up and then they will use and try your product. That means that the…”
Gonto: Split growth teams between marketing acquisition and product retention
“I try to have a growth team in marketing that focuses mostly on driving activated signups. And I try to get a growth team in product as focus is mostly on retention and conversion.”
Gonto: Product managers often know users less well than marketing teams
“I've worked with so many companies where the product managers don't really know the user as much. Marketing has to know the user because they are in events. They are in trade shows. They are online. And they need to send that information back to products and t…”
Gonto: Early growth at Slack and LinkedIn was not driven by growth loop frameworks
“When you go to the leaders in Slack and LinkedIn at those times, they will tell you they will not, they were not implementing growth loops. They didn't know what a growth loop was. They were trying to find creative ways to get other users and other people to c…”
Gonto: Growth candidate case studies should be intentionally abstract
“What I think is important about the exercises is to make them the most abstract possible. If you make them abstract, you're giving people a chance to Reflect about them, ask questions to better understand, and then refocus on something to try to be creative. I…”
Gonto: Leaders should hire at least one smart direct report they dislike
“On the people that report to me, I always try to get at least one person that I don't like, because if I don't like them, but they are smart and they can get shit done, I will hate everything they say, but I will put my thoughts into believing them, trying to …”
Gontovnikas spent $300K on paid ads that yielded zero revenue conversions
“The easier mistake was we spent a lot of money on ads to drive signups and we spent 300 K. But what we didn't think about then was we got a lot of signups. But none of them activated or very few of them activated. That meant that most of them weren't retained …”
Gonto: Auth0 users receiving no onboarding emails were 20% more retained
“So instead of just looking at open and click rate, what we did was to 10% of the users, we sent no emails. To the other 90%, we sent the emails, and we compared which of them were more retained. What we saw is that the people who got no email were 20% more ret…”
Stebbings: Gamification of metrics is the most dangerous practice in companies
“The gamification of metrics, I think is the most dangerous thing in companies.”
Gonto: Fine-tuned GPT models can fully automate personalized SDR emails
“You can then use fine tuned models of GPT to fully personalize emails in the name of an SDR without a real SDR.”
Gontovnikas built CEO deepfake technology for personalized outbound sales emails
“You can even, and we've created this, you can even create deep fakes of CEOs telling you something on a video, on an email, even though it's not really the CEO.”
Gonto: Off-the-shelf automated AI marketing tools will lose effectiveness monthly
“So to me, what will change on that is you will still think, I think, need engineering and ops because as more things get automated and you get more ease of access to some of those, they will Start working less and less every month, and you will need to switch …”
Gonto: Strong PMF Startups Wait Too Long to Build Growth Teams
“The biggest mistake that I see is that they start very, very late if the PMF is very strong. I've seen a lot of companies that have a very strong product market fit. And because of that, in the beginning, they just start getting a lot of revenue because big co…”
Gonto: Dedicated YouTube sponsorship video for Clerc generated zero growth
“We went to find a really known YouTuber and we paid them to do an entire video about us, an entire video about Clark. When we did that, that had literally no effect. That didn't work at all.”
Gonto: Organic product placement in YouTube tutorial builds outperforms dedicated sponsorships
“Instead of paying them to highlight us as the main thing, as the main focus of the video, what they would do is they would build an app. They would build any type of app. And as they were building it, they would use Clerc without mentioning it and without talk…”