Ely Lerner on Startup Advisors, Growth Strategy and Focus

Ely Lerner tells Harry Stebbings that he trained in computer science, worked as an engineer and engineering manager, and spent more than eight years at Yelp. Yelp’s many marketplaces gave him exposure to consumer, small-business and enterprise dynamics. He often worked on new products and later led consumer product at Chime.

Yelp acquired food-delivery startup Eat24 in 2015, and Lerner became its product head. He says Yelp sold it to Grubhub about two years later for a little more than twice what it had paid, along with a partnership that let Yelp keep higher-margin lead-generation revenue while leaving fulfillment. This is his account of the transaction and its economics.

For a genuinely new business inside an established company, Lerner recommends giving a leader profit-and-loss responsibility and the autonomy of a founder. Fund the unit in stages, he says, and review its progress as an investor might. In his experience, putting a zero-to-one effort inside a legacy product organisation reduces its chance of success.

He distinguishes augmenting an existing product with new AI technology from creating a new business serving a different need. A corporate startup has an advantage when it uses assets an independent startup lacks: customers, distribution, data and technology. Stebbings adds that existing distribution can accelerate learning as well as reach.

The lesson Lerner carries from Yelp and Chime is alignment of incentives. In a marketplace, he says, growth is easier when an improvement benefits suppliers, customers and the company’s economics together. He sees similar alignment between Chime’s business model and its customers, a judgment informed by his former role at the company.

Lerner distinguishes company-level advice to a founder, functional advice such as marketing, and deep domain advice such as SEO or logistics. Coaching is a separate activity aimed at personal and leadership development. His own company-level practice focuses on helping founders grow their businesses across functions and organisational questions.

Founders should own the strategy for their primary growth lever, Lerner argues, whether that is product, marketing or sales. An expert can teach and support them, but cannot substitute for the founder’s company-wide perspective and responsibility. He is more comfortable delegating areas that protect the business but are not its principal growth engine.

The right engagement changes with the company’s needs. An adviser can help an early founder learn to make decisions; later a fractional leader may bridge the gap before a full-time functional head is justified. A functional adviser can also support a more junior manager who runs the day-to-day team.

For a functional adviser, Lerner recommends first identifying the lever the company needs to move, then seeking someone who has worked on that lever at a comparable stage and company type. He contrasts traditional occasional advice based on past experience with his own hands-on model: weekly founder meetings plus time studying data, calls and the product.

His aim in hands-on advising is to teach the founder rather than take over execution. Stebbings challenges whether an adviser should instead work directly with a product or marketing leader while the CEO handles fundraising and other demands. Lerner responds that, when the issue is the primary growth lever, the founder still needs to own the strategy.

Stebbings suggests hiring separate advisers for several secondary levers. Lerner argues that early-stage startups are generally better off focusing on one primary lever first, then adding channels as they scale. He says his clients usually range from seed to Series C, with some later-stage leaders building a new business unit under their own P&L.

Traditional functional growth advice can be premature before a startup knows its growth mechanism, Lerner says. He describes his own work at the seed stage as helping founders identify the lever in the first place. This is also a description of the service he sells, not independent evidence of its effectiveness.

Prior experience matters, but it need not be identical. Lerner asks whether an adviser built a playbook or scaled one someone else created, and matches that experience to the problem. Expertise in virality is different from paid acquisition; geography matters more for some areas, such as cultural brand work or local fintech infrastructure, than others.

To test both value and working chemistry, Lerner says he helps founders for free before formalising an arrangement. After roughly three or four calls, both sides usually have a clearer sense of fit. If they do not continue, he tries to leave the founder with frameworks to use independently.

His own engagements operate week to week and month to month, with either side able to stop or pause future work and compensation. For company-level advice he asks whether the founder believes the adviser can materially improve the growth trajectory. For functional or fractional work, he says compensation can be benchmarked against an equivalent full-time role.

A fractional CMO or CTO working two or three days a week might receive a proportion of a full-time executive package, Lerner suggests. He says equity often has no cliff because an engagement may be short. These are his broad practices, not a standard contract or a universal market rate.

For his own hands-on work, Lerner describes a roughly 80/20 cash-and-equity mix, with compensation accruing month by month; he does not supply an equity percentage of the company. The split is his arrangement, and the value of any equity accumulates over a long engagement. He frames the model as exchanging value as the work proceeds.

A productive advisory relationship requires the founder to ask questions of data and customers and ship small experiments quickly. If a team cannot ship, Lerner may temporarily help address that before returning to strategy. If a company has no usable instrumentation, he may ask it to return once it can measure.

One reason teams struggle to ship, he argues, is that engineering owns output rather than outcomes. A team handed a ticket may view its job as building the requested feature, rather than helping achieve the customer and business result. He also sees confusion about which work advances the strategy and which protects the existing business.

Lerner calls the small number of bets aimed at taking a business to the next level “offense.” “Defense” includes necessary work to protect against downside, such as reliability or risk controls; some work belongs to neither category. In an early business, offense often means improving a compounding growth loop rather than a one-off metric bump.

Defensive investment has diminishing returns, in his framework: reducing risk toward zero can absorb resources that could otherwise support growth. Stebbings offers a hypothetical comparison of 95% and 100% security or reliability; those numbers are illustrative, not a recommended safety threshold. Lerner stresses evaluating the specific risk and opportunity cost.

For technical capacity, he gives a hypothetical: if infrastructure may fail around ten million users, plan conservatively around eight million, estimate the work required, add time buffer and use growth projections to schedule it. Revisit those projections as growth changes. His aim is to avoid both an avoidable failure and premature work that crowds out higher-leverage investment.

Lerner suggests tagging the roadmap as offense, defense or neither, and questioning the last category. A core-experience organisation can prioritise defensive work against other defensive work; separate teams can focus on each offensive bet. He suggests at most three major offensive bets, with focused teams for each.

Categories can change. An established growth loop may become defensive once additional investment produces diminishing returns, even though it still needs maintenance. Trust and safety, usually defense, can become offense when greater trust unlocks new demand; Lerner cites Airbnb as an example of that possibility.

Defensive work is just as important and may be existential, Lerner says: without it, the company may not survive to invest in growth. Leaders should celebrate a team that safely reduces the resources needed for defensive work as well as a team that moves a growth metric. The saving creates opportunities elsewhere.

Organising teams around outcomes can mean several groups touch the same code, creating coordination costs. Lerner sees that as a trade-off against teams organised solely around features. He says a frequent strategic error is calling too many initiatives offensive; another is investing so much in defense that diminishing returns consume resources needed elsewhere.

An offensive bet should identify a leverage point and desired outcome, not prescribe one feature. If a proposed solution fails, the team should learn whether the underlying bet is still sound. Lerner advises surfacing the riskiest assumptions early and testing them quickly rather than treating the initial strategic bet as settled.

Asked about retention, Lerner argues against focusing a team narrowly on people who are already churning. He often finds a stronger lever in activation: helping suitable new users experience the product’s core value and build a habit. Some departing users were never a good fit; his claim is a strategic view rather than proof that activation always dominates.

Advice can fail when an adviser simply repeats what worked at a different company without accounting for context. Lerner favours a close working relationship that applies experience to the actual business. It can also fail if a founder is too busy to engage; regular work and shared objectives are essential in his model.

Lerner says he was working closely with about six clients at the time, usually an hour a week each plus preparation, and might take a couple more. A less hands-on adviser could handle more. As a company’s problems change, he says, the founder should reassess whether the adviser’s strengths still match its needs.

The hardest part of solo advising for him is loneliness: unlike an operator, he has no team around him every day. He seeks community with other independent practitioners. On organisation, he says a separate growth team can help a company develop growth capability, but the longer-term aim should be to integrate growth thinking throughout the business.

The highest-impact growth opportunities he has seen often involve activation, though he says the top lever is not always activation. Strong retention compounds, supports engagement and can contribute to referrals. Teams should investigate their own bottlenecks rather than assume a universal answer.

A common activation mistake is to optimise aggregate conversion without separating ideal customers from people unlikely ever to benefit. Making a button bigger may improve a metric while missing the highest-leverage gap. Lerner recommends identifying the ideal customer profile, measuring its progress through activation and targeting the biggest obstacle for that group.

Even horizontal products have an ideal customer profile, Lerner says, but the shared attributes may span different industries and job titles. He offers a hypothetical Airtable user who needs people to collaborate around shared data. Early focus can be narrow around an acute common problem without being limited to a single profession.

Lerner’s own early product mistake was focusing too much on features and solutions rather than customer and business outcomes. Running zero-to-one businesses with P&L responsibility helped him see the whole system rather than one small part of a mature product. His offense-and-defense framework emerged from mistakes he says he made himself.

He argues that generic growth hacks have lost their appeal, but does not say channels such as push notifications or paid acquisition are dead. They can still work in specific contexts; simply adding a channel without identifying leverage is the weaker approach. He advises incoming growth leaders to assess the opportunity and whether founders will support a growth mindset.

A startup need not have perfectly clean and structured data before its first growth hire, Lerner says, but it needs some information and the ability to instrument important questions quickly. Stebbings mentions Canva’s TikTok and Notion’s YouTube work; Lerner says marketplace network effects most excite him.

He points to early-stage Cash Flow Portal as an example of a company exploring network effects in real-estate investing, and says he knows a founder. He asks whether a change in technology or market structure makes a new model possible. The example is his assessment of potential, not evidence that the company has achieved network effects.