Adam Fishman on Growth Loops, Hiring and Learning from Failure

Adam Fishman began as a quantitative marketer, moving toward consumer psychology, analytics and experimentation rather than engineering. He later joined Zimride as head of growth, shortly before its pivot to Lyft, and describes that period as an education in building a company.

At Lyft, he says, speed changed his relationship with risk and failure: many decisions could be reversed, so teams could learn by moving. At Patreon, he learned a different lesson about respecting creators and including their input in product decisions.

Stebbings warns that speed matters only if teams actually learn from failures. Fishman agrees: he documents experiments, conducts post-mortems and shares findings so the organization learns about customers or its growth model, rather than merely repeating unsuccessful work.

Fishman rarely relies on instinct alone, but also warns against treating quantitative data as the whole answer. At Patreon, he says, the team sometimes inferred what creators wanted from numbers without asking them; he advocates pairing measurements with interviews and prototypes.

He defines growth as understanding how a company grows, identifying levers and testing them. The goal is to connect more people to existing value through acquisition, activation and retention, and monetization, not necessarily to invent a new product.

Before hiring, a founder should understand the growth model, the skills needed to test it and how involved they intend to remain. Fishman often favors a director-level player-coach who can build processes, establish data and communicate, rather than a purely junior executor or a senior leader who first needs a large team.

He says the right organizational home depends on the growth model. Viral product growth may belong within product; if paid acquisition is the main lever, a marketing hire may be appropriate. Copying Facebook's standalone team structure is not automatically suitable.

Organization design changes as companies evolve. A standalone team can provide focus and accountability but create tension with other leaders; an embedded team may collaborate more naturally but lose focus when core product or marketing work takes priority.

Fishman teaches growth models at Reforge and advises companies, giving him a professional interest in the framework. He describes a loop as a self-contained cycle in which an output feeds back as an input; mapping it can support alignment, prioritization and organizational design.

His Lyft example is an incentivized referral: a rider invites another person, the new rider receives credit and, after a qualifying ride, the inviter receives credit. The new rider can then invite others, restarting the cycle.

A qualitative model maps steps visually; a quantitative model adds numbers and conversion rates. Even a simple map can reveal levers such as the invitation rate, the number of invites and the proportion of signups who take a first ride.

He recommends keeping loops simple enough to guide decisions, often around three to six steps rather than dozens. This is a practical heuristic, not a requirement that every product's growth can be represented in that way.

Loops can reach a ceiling. Fishman generally recommends focusing on one or two core loops, forecasting when they may slow and then considering new loops or a larger eligible population; user-generated loops may scale differently from company-funded activity.

He sees two common mistakes: calling every channel a loop when the output does not actually produce the next input, and spreading attention across too many loops at once. A partnership process, for example, may still require substantial separate effort to generate each next partner.

At tutoring marketplace Wyzant, Fishman says a Lyft-style referral incentive failed despite a well-built system. Follow-up conversations suggested that some parents were reluctant to tell others their child needed tutoring; his account describes a contextual finding, not a measured claim about every US family.

For hiring, he would ideally observe leaders working alongside a team because traditional interviews provide an incomplete picture. He also recognizes companies need a way to assess candidates before working together.

His pre-work is to understand the business's growth model and hire for the actual gap: marketing if the work requires marketing, product if it requires product. He likes quantitatively capable generalists who prioritize well, communicate and are willing to learn.

He identifies competencies and designs exercises around them: data analysis, prioritization, behavioral questions and a presentation or take-home task. The goal is to approximate the work rather than rely only on a candidate's résumé.

In a quantitative exercise, a candidate might examine a dummy cohort dataset, explain what they observe and suggest next questions. Fishman looks at how they organize information and reason about an apparent drop-off, not just whether they guess a correct explanation.

For prioritization, he gives candidates growth-model information and a list of possible initiatives, then introduces new constraints. He wants to see how they assess impact, confidence and reach and whether they can revise priorities when information changes.

Behavioral questions probe risk tolerance, persistence, humility and response to feedback. He asks for examples of being wrong, changing course or persuading a skeptical CEO, and listens for accountability rather than reflexive blame.

A presentation can test whether someone explains a business and a proposed change in plain language. Fishman may also use a working session and seek references to understand collaboration beyond a polished interview.

He shares a story about Lyft investor Raj Kapoor speaking with Logan's former school principal during diligence. Fishman recalls that the principal remembered a childhood prank involving an automated call; he presents it as an anecdote about unusually thorough references.

For onboarding, he compares a new employee's first months to customer activation: the company should protect early momentum and help the hire achieve useful wins. He recommends a clear plan agreed with both the hire and the colleagues who will work with them.

His 30-, 60- and 90-day plan typically starts with learning and relationships, moves to improving or developing growth strategy, then turns to longer-term planning and alignment. He spends substantial time with the new person instead of leaving them to navigate alone.

At Imperfect Foods, Fishman says he documented milestones and expected deliverables in a shared spreadsheet and circulated it broadly. The specific tool matters less to him than fitting the company's communication culture and making expectations visible.

He generally gives a new hire around 90 days to find their footing, while accounting for a steep learning curve in a new industry. At 45 to 60 days he looks for momentum, thoughtful questions and effort, not an immediate transformation of the company.

His first communication rule is to make the work understandable and repeat it until stakeholders have context. He adapts the format to the audience: some executives may need a one-on-one explanation of the what and why rather than an email.

At Patreon he sent a predictable weekly update with highlights, low points, customer and experiment learnings, and upcoming work. He says colleagues came to expect it, asked questions and offered feedback; a template later helped distribute the effort.

Fishman prefers an imperfect update sent on time to one delayed in pursuit of perfection. Visual aids and prototypes can make an idea easier to understand, while candor and kindness make it easier for others to provide feedback.

At Patreon, Fishman concluded that branding and positioning were constraining future growth. He brought in outside expertise and says the work required persuading skeptical colleagues and the founder rather than assuming everyone already saw the same problem.

The team used customer interviews and blinded positioning tests because a whole-brand change is difficult to test as a simple reversible experiment. Fishman says this built confidence in shifting the positioning from crowdfunding toward membership.

He distinguishes the promise made by a brand's words and visuals from the product experience that must deliver on it. Patreon introduced new functionality alongside the new look, and he prepared colleagues for negative social-media reactions after launch.

In the quick-fire round, Fishman draws parallels between management and parenting, including empathy, difficult conversations and goal-setting. He acknowledges that employees may not like being compared with children; this is his personal analogy.

He says enduring growth principles include understanding motivation and behavioral effects such as scarcity, while tactics change. In his 2022 view, referral mechanisms, content marketing and SEO had become less novel or more competitive, rather than necessarily disappearing.

A founder's mistake is expecting growth to be an immediate fix or using a growth hire to compensate for missing product-market fit. Stebbings adds that an unsuccessful executive hire can also damage relationships and morale across functions.

Fishman wants more companies to treat acquisition, retention and monetization as a connected system. He also urges growth leaders not to assume a tactic that worked at a prior company will transfer unchanged.

His most painful example was Patreon's proposed shift of payment-processing fees from creators to patrons. The team tested conversion quantitatively but, he says, did not speak to a single creator before announcing the change.

Creators reacted strongly; the company first defended the decision and then rolled it back. Fishman says the episode slowed growth and damaged trust for months before the company recovered. The account is his retrospective description, not an independently quantified impact.

Asked for a growth strategy he admires, he names Locket, a photo-sharing app using iOS widgets, and praises its simple product-led sharing. That is his view in 2022, not a current performance assessment.

Stebbings thanks Fishman, and they close the conversation.