Luke Harries on Building Growth Teams That Don't Need Product Managers

The Heresy of Horizontal Products

When Matty Shulman pitched Luke Harries on ElevenLabs while swimming in the 11-degree Hampstead Heath ponds, the go-to-market plan sounded terrible. Step one: build the world's best audio AI models. Step two: sell it to everyone. Creators making audiobooks, people listening on their commute, developers building voice agents. Harries, then angel investing, passed on writing a check. "That is a terrible go-to-market plan," he told his friend from a Cambridge hackathon seven years prior.

Six months later, Shulman called back. ElevenLabs had hit one million users. The horizontal strategy that violated every growth playbook—tight ICP, focused channels, craft product for one buyer—was working. Harries joined as Head of Growth, inheriting a three-person team and a company now valued at $3.3 billion that sells to radically different customer segments.

The conventional wisdom Harries absorbed at PostHog was elegant: one ICP (product engineers), multiple products (analytics, session recording, feature flags). Each new product created another entry point, and integrations unlocked compound use cases. ElevenLabs had none of that cohesion. Instead, they had foundational AI audio models—text to speech, speech to text, sound effects, voice cloning—packaged as APIs for developers, a reader app for mobile consumers, and a creator platform for audiobook producers.

We don't have any of that in some ways. Because we have the best AI audio models, this is a very horizontal foundational layer which we sell as an API. But then because the team is super ambitious, Matty and Peter literally want to build a bigger company than Google.

— Luke Harries

The ambition isn't theoretical. It demanded a structural solution most growth teams never attempt: complete organizational sharding.

Growth Teams as Startups Within Startups

Harries didn't build one horizontal growth team attempting to serve multiple masters. He built discrete growth organizations for each product vertical, each with dedicated specialists. The consumer app team has its own growth marketer, paid ads, and community manager. Creator growth runs separate paid ads and affiliate programs. Developer marketing operates standalone. By year-end, enterprise marketing alone will be 20 people. The mobile app will command a 5-to-10 person growth team.

This structure resembles a holding company more than a traditional growth org. Each product gets its own "CMO"—a product growth lead responsible for all metrics, activation, and awareness within their domain. They're not account managers coordinating with a central growth team. They're building sub-teams specialized for their vertical.

The horizontal layer exists, but it's inverted. Channel specialists sit above the product teams: a head of performance marketing from Shopify, an SEO expert from Canva. These aren't junior coordinators. They're world-class practitioners whose expertise gets deployed across verticals where it makes sense, not mandated universally.

We have horizontal growth team, which are channel specialists. But then within each team, we have a product growth lead who's responsible—they're basically CMO for their product, and they're thinking out all the metrics, the activation, the awareness.

— Luke Harries

The model works because ElevenLabs started with legitimately superior technology. The audio models created natural inbound demand across segments. That enabled fundraising ($281 million) to hire "incredible founder-type peoples" who could own discrete products. Most companies can't replicate this. They should choose one ICP and build compounding products around it. But if you have horizontal pull from genuinely best-in-class infrastructure, sharding might be the only way to capture it all.

The First Three Hires Are a Generalist, an Engineer, and a Motion Designer

When Harries joined ElevenLabs, the growth team was three junior specialists. He hadn't run growth before. His advice on early hiring is specific: start with one generalist growth marketer responsible for everything from messaging to channel testing. Not a product marketer who perfects positioning while no one hears about it. Not a quantitative performance marketer who drives traffic that doesn't convert.

The generalist must deeply understand the product. For technical products, hire technical. For consumer products, hire someone who intuitively gets virality. The second hire should be a front-end-leaning growth engineer—hacky, metrics-driven, building landing pages for SEO, mini tools, automated outreach. After that: motion designers and backend-focused growth engineers.

Motion designers? The recommendation surprises American founders unfamiliar with the discipline. Harries is emphatic: video is "hugely underrated as a medium for growth," and motion design—animated graphics, not talking heads—is the best format for launches. Every major ElevenLabs launch ships with a motion design video as the keystone marketing piece.

The big mistake that I see with most people's videos is that they put the founder themselves, they do a full 5-minute monologue themselves about the company mission. And unless you have the editing skills of MrBeast, you're just not going to keep people for the launch video.

— Luke Harries

He's not precious about this. Harries calls out Superhuman's recent founder-led launch video—Rahul Vohra walking down a beautiful staircase for five minutes—as too indulgent for the attention economy. Motion design keeps it abstract, focuses on core value props, and captures attention in the first 30 seconds. That's where all optimization effort should go. Building the video out to five minutes after that is fine, but recognize most drop-off happens after the opening.

The other video formats—founder-led narratives, screen-share product walkthroughs—have uses. Founder-led works for lifestyle products, prominent creators, or consumer packaged goods. Screen-share works for complex product education. But for launches targeting broad awareness, motion design wins.

Patience as Competitive Advantage

Before ElevenLabs, Harries co-founded Fella with Richie, a men's health company that evolved from AI browser automation to a COVID testing clinic to digital CBT for binge eating to GLP-1 weight loss medication. They read the semaglutide paper the day it came out, raced to set up a clinic, and hit $300K in revenue. Then growth flatlined.

Harries was impatient. First company, still young, wanting bigger impact. He pushed to pivot to B2B. Richie wanted to stay focused and wait for the wave. Harries stepped back, joined PostHog. Richie kept building. Fella is now doing over $30 million a year in revenue, helping thousands of patients through the Ozempic wave that arrived after they were early.

Big lesson for me is commit for the long term, be patient. And Ozempic had a much, much bigger wave after. We were actually quite early on.

— Luke Harries

Market timing isn't just about being right. It's about being right and staying in the game long enough for the market to prove you right. Harries learned this by watching his co-founder succeed after he left. The lesson stuck. At ElevenLabs, the strategy that sounded terrible in the Hampstead ponds—horizontal everything, no focused ICP—required the same patience. Wait for the AI wave. Hire incredible people. Let discrete teams prove out discrete verticals.

The Cambridge hackathon that connected Harries to Shulman also introduced him to Filip, who organized it and later founded Wordware, which became YC's largest raise ever at roughly $40 million. The hackathon prize was a Microsoft Xbox each, immediately sold on eBay for £300. It also landed Harries his first job at Microsoft. His top tip for young people: do hackathons with your smartest friends.

Seven years of staying in touch with Shulman, catching up every six months to riff on startups, created the relationship that brought Harries to ElevenLabs. The company that builds audio models for everyone is led by a growth exec who learned patience by leaving too early once before.