Matt Lerner on Finding Growth Levers and Hiring Teams That Learn

Harry Stebbings welcomes Matt Lerner, who traces his career from philosophy and rhetoric in Silicon Valley to customer acquisition. He joined PayPal's growth team in 2004 and stayed almost 11 years, then became a venture investor. Seeing startups repeatedly waste money and talent on avoidable mistakes, he left investing to help them find their most powerful growth levers.

At a large company, Lerner learned, having the right answer can be a small fraction of the work; aligning people, changing minds and getting resources can be much harder. He shifted his attention and hiring toward influence. His percentages for that division are rough illustrations rather than a precise measurement.

Looking back at PayPal, he estimates roughly 90% of growth came from 10% of its activities: eBay integration and referral bonuses before his time, network effects, developer engagement, integrations with shopping carts and hosts, and some outbound sales. PayPal also spent heavily on campaigns and products that did little. A cash-rich company could afford that; a startup generally cannot.

The question is how a startup finds its few high-impact levers quickly. Lerner wishes he had understood earlier that growth is primarily discovery, not copying a playbook. Search rankings and ad auctions are crowded by established competitors with stronger economics and years of optimization; the challenge is to learn the details that make a channel work for this business.

The hardest obstacle is often hubris: people assume that the practices from their last job or a podcast will work immediately. Lerner defines growth as finding customers, helping them understand the value offered, and getting them to use and love the product. For a novel product, the customer may not even know the category exists.

Growth is cross-functional. The team must discover what customers are trying to do, what they think they need and where they look, then build relationships across product, engineering and other functions. Unlike a familiar consumer-goods job, there may be no established playbook for selling the product.

Asked whether growth is art or science, Lerner says both matter. Pure optimization can produce unpleasant travel-booking experiences, while entertaining advertising may fail to sell anything. Pressed to choose, he picks science: in his view, science without art can still deliver business results, whereas art without measurement rarely does so reliably.

Stebbings asks whether user experience and conversion can conflict. Lerner says some friction can help: services such as Noom and Calm ask many onboarding questions, taking customers through a purchase journey. Besides gathering information, the questions can show that the service understands a person's circumstances and build intent; sunk cost may also play a part.

Whether to pursue big bets or incremental optimization depends on the company's stage. Early on, Lerner says, it is premature to optimize before finding a lever. Once one or two work, the team can devote more resources to refinement while reserving some capacity to discover the next segment, use case or channel.

Good intuition rests on knowledge rather than the confidence of the loudest person in the room. A leader needs to understand the customer's context, the mathematical growth model, the largest points of leverage and the company's capabilities. Those inputs make intuition more useful when deciding where to focus.

To map a growth model, Lerner starts with a north-star metric that increases when the company delivers customer value and encompasses the funnel. Stress-test it: can the team make the number rise without actually creating value? He says revenue or profit alone can reward early activities that produce money without producing the conditions for durable growth.

A shared customer-value metric also gives functional leaders a common goal. Stebbings suggests going beyond generic weekly active users to a behavior appropriate to the product. Lerner's test is: how would a normal customer behave if they loved the product, and how many customers behave that way? Track that behavior by cohort over time.

Lerner considers the question of whether customers would be upset if a product disappeared a useful product-market-fit test. But six delighted customers becoming seven the next year do not necessarily make a scalable business. The north star can change as a company learns what value customers actually receive or starts a distinct business line.

A founder may know the broad problem the company wants to solve, but customers arrive with particular jobs and contexts. In Lerner's accounting example, the founder might assume expense logging matters most, only to discover that customers who file quarterly VAT returns retain better. A quarterly filing measure might then make more sense than weekly activity.

To align a team around the north star, Lerner asks each person to explain how their work affects it. The answer can uncover missing context or an unexpected contribution. He then asks what that person can do this quarter that would have the greatest impact, even if the connection runs through several steps.

As an investor, Lerner saw three recurring founder patterns. Overthinkers spend too long debating and planning, sometimes bringing habits from large organizations. Underthinkers have a bias for action but keep building features without learning, adding complexity to the product, sales process and organization when speed matters most.

A third group hires functional heads and delegates before it has established a strategy or product-market fit. Each leader can bring a separate set of best practices and resource requests, leaving the founder to resolve competing priorities. A young startup, Lerner argues, needs to excel at a small number of things rather than all of them at once.

Stebbings worries that founders spread themselves across too many channels. Lerner says paid acquisition can be useful early as a quick way to get attention for tests, even if it should not necessarily be the foundation of growth. Trying every channel once is inefficient: most do not work on the first attempt, and early economics may deteriorate as spending scales.

Channel categories also overlap. A podcast can be content, influencer distribution and a network-effect business at once. Lerner lists broad ways people discover products, including sales, partners, search and inbound content, advertising and influencers. The useful first step is to eliminate channels that do not fit the product's price, economics or customer behavior.

Choosing the likely channels is only a small part of the job; making them work takes experiments. Search, ads and outbound all face intense competition. Stebbings notes that content can take a long time to show results, and Lerner agrees that entrepreneurial progress often arrives in irregular bursts after stretches with little visible movement.

Lerner calls a precise moment of customer need a 'locksmith moment.' He once saw a 24-hour locksmith's sticker on the gate to his apartment complex, exactly where someone locked out would need help. A startup should look for the equivalent situation and find a way to be present when the customer recognizes the problem.

For a horizontal product with several possible uses, Lerner starts by interviewing customers who already signed up. Jobs-to-be-done interviews seek the specific problem and outcome behind that decision. A good landing-page headline, he says, completes the phrase 'Now you can' with something meaningful the customer wants to accomplish.

After collecting interviews, he looks for recurring verbs and desired outcomes; a language model can help extract themes from many transcripts. Teams can show mock landing pages to customers or run inexpensive ad tests with several headlines aimed at a relevant audience. The ads need not be profitable: the immediate purpose is to learn which messages attract attention.

On when to hire a growth leader, Lerner challenges the advice he hears from leaders at well-funded companies. Early startups often cannot attract a proven growth star or give that person all the resources required. He believes the founder should be the first head of growth, close to customers and able to mobilize the company.

A senior 'growth wizard' who imports a prior playbook may not understand the new business, and founders may not give them the engineering and product resources to succeed. Once the rough growth approach is clearer, Lerner prefers early hires who can do some of the hard work already and learn the rest. His deliberately playful formulation is '50% of the hard things and figure out the other 90%.'

Requirements change between writing a job description and the person starting, and again months later. Lerner looks for bright generalists with a high rate of learning, hoping one can eventually become head of growth. To find candidates, he recommends removing unnecessary filters such as a particular university, geography or narrow sequence of prior jobs.

Lerner says his best growth hires at PayPal had no marketing or product background: one had been a physicist and another an academic computer science researcher. They joined a supportive organization and brought data skills, first-principles thinking and an understanding of experiments and causation. He values the curiosity and humility to recognize what is not yet known.

In interviews, Lerner begins broadly: what is the candidate great at, what work do they enjoy, and what would they delegate? He then probes the skills still in doubt with open questions about attribution, retention or obtaining resources without authority. He asks candidates to explain their thinking and describe relevant past situations.

To assess learning, he asks what candidates hope to learn in the role, about a mistake they made and what they would do differently, and whether they have questions for him. He looks for comfort with mistakes and genuine curiosity. He also asks what open questions they see in the business to distinguish people who discover a new playbook from those who merely execute an old one.

Lerner does not necessarily prepare candidates for that business question; he is interested in their reasoning, not only a correct answer. He cautions against being dazzled by a prestigious employer on a résumé: having worked at a successful company does not prove the candidate caused its success. He generally probes examples informally rather than using formal case studies.

His own hiring mistakes included overlooking small signs of dishonesty because he liked a talented candidate. In his experience, those hires later put him and the organization in difficult positions and he had to let them go. He treats integrity as a serious hiring consideration, though his account is a personal lesson rather than proof about every candidate.

Lerner favors some in-person time for growth teams: execution such as analysis, customer interviews and setting up experiments can be solitary, but sharing knowledge and generating ideas benefit from coming together. He suggests one or two shared days and a weekly growth meeting that reviews priorities, experiments, customer learning and trends before deciding the next week's work.

Onboarding has two steps. First, the hire becomes an information sponge: listen to customers, study research and old experiments, talk to sales and support, understand the economics and funnel, and meet the team. Second, after roughly 30 days, the founder and hire should debate priorities and what not to do, testing the reasoning even when they agree.

Experienced growth people may have quick wins, such as improving a headline, reallocating ad spend toward higher-value customers or reactivating dormant leads by email. These can be useful, especially when a company needs cash, but are often one-off improvements. A conversion rate that rises from 1% to 10% cannot be multiplied tenfold over and over; fundamentals still matter.

Stebbings likes a paid one-week trial so both sides can see whether a hire fits. Lerner agrees that trying before committing can work and warns against arriving with a fixed playbook instead of the humility to understand the business. His own example of a mistaken assumption was dismissing LinkedIn as a channel for his company.

At a respected person's suggestion, Lerner repurposed his weekly emails as LinkedIn posts. Results were uneven at first, then compounded until LinkedIn became a major source of subscribers. Subscribers are a funnel driver, he says, not his company's north star: that is the number of startups it can positively help.

Both speakers find that posts they expect to do well can disappoint, while human stories and emotion often travel further than abstract analysis. Lerner defines a growth loop as positive feedback that keeps running: referrals, ads whose returns can fund more ads, or a podcast whose guests attract listeners and make it easier to attract future guests.

A company should find and deepen its strongest loops, although optimizing several at once is hard. Lerner says improving ad returns can require a great many experiments. In an early startup, the whole company may temporarily become the growth team under the founder; in a larger company, a distinct cross-functional team needs strong relationships to obtain resources from other groups.

Writing a book was harder than Lerner expected because organizing a full-length work differs from writing short emails and posts. Editors helped him sequence his thinking. Asked about hiring mistakes, he returns to the danger of impressive résumés: established growth stars are scarce, so startups may need to find people with potential and let them grow.

Stebbings observes that many growth specialists began in SEO. Lerner says search work forces people to think about customer intent: what someone is looking for and trying to accomplish. On AI, he believes models can raise the baseline for design and copy, help automate outreach, and quickly identify themes across customer conversations.

Stebbings questions whether automated outbound will become saturated and generic. Lerner thinks the window was still open at the time of the interview, but acknowledges that he has long predicted outbound's death and that some practitioners still make it work. He does not commit to a definitive forecast and expects AI capabilities to change quickly.

Lerner compares the excitement around AI with earlier internet and mobile waves. Although distribution technology evolves, he says familiar fundamentals persist: the right message, good design, testing and experimentation. His goal in his own business is to help founders replace premature optimization and confident assumptions with discovery and learning.

On European technology, Lerner is optimistic about talent but worries people may move to more startup-friendly funding, regulatory, labor and tax environments. He says Europe needs reasons for founders to stay, list and raise later rounds locally. He describes stronger product-market fit for his own business in Germany and expresses concern about the UK; these are his observations, not a comprehensive market assessment.

For a final example, Lerner returns to lengthy onboarding flows at consumer services such as Calm and Noom. He thinks B2B companies with considered purchases could adapt them to guide customers through an information and decision journey without requiring a salesperson for every interaction. The lesson is to design for the customer's decision, not simply remove every step.