Adam Grenier on Building Growth into the Business

Adam Grenier started in data-oriented digital marketing, helping clients reach Java developers on early mobile phones and adapt their products to social media, video and mobile. Uber was the first company where he formally worked on a growth team, recruited by Ed Baker, who brought experience from Facebook.

Grenier sees growth as both art and science: understanding customer psychology and the moment when a product's value is most compelling, then using technology, data and systems to repeat what works. Neither customer insight nor measurement is enough on its own.

At HotelTonight, hotel agreements restricted its advertised rates to mobile devices, ruling out the search engine strategy used by established travel companies. The team explored new mobile channels and referral programs; Grenier recalls being an early customer for Facebook mobile ads.

At Uber, he says, the business grew from roughly a dozen cities to more than a thousand during his tenure. Growth meant balancing both sides of a highly local marketplace whose conditions could change from moment to moment, not simply acquiring more riders.

Lambda School's central challenge shifted from enrolling students to helping them get jobs. Grenier says the company had to design product and marketing around that outcome and the needs of students, teachers, coaches and hiring managers.

At MasterClass, he encountered changes to ad targeting, the return to life outside after COVID, and an audience increasingly accustomed to famous instructors. The marketing challenge became communicating the value of the whole subscription and increasing engagement, rather than selling one instructor at a time.

Talent networks helped create a burst at a class launch, and the company could market through instructors' channels. Grenier says the mix was nevertheless broader, with much of a class's value accumulating after launch and the majority of acquisition still coming from MasterClass itself.

Grenier argues that a growth-minded marketing leader should apply experimentation, data, product and engineering to brand as well as performance advertising. Individuals may have different strengths, but he sees less reason to operate the two sides as entirely separate disciplines.

Harry Stebbings asks whether leaders who combine brand and analytics are too rare. Grenier says a leader can start from either specialty, build breadth and assemble complementary skills; the important change is how the team operates rather than finding one person who is expert at everything.

Most companies, Grenier argues, do not need a separate growth team. Uber created one after it already had strong product-market fit and a substantial organization; its specialized growth functions eventually merged back into the core business.

Product and marketing teams already care about growth, he says, though urgent work can crowd it out. A cross-functional strike team can attack a particular constraint, but successful acquisition, activation or referral work still needs an enduring owner after that team moves on.

In Grenier's experience, a focused team is most useful when a company already has rapid growth and must understand and sustain it. Hiring a head of growth because growth is absent can mistake a product-market-fit problem for a hiring problem.

When advising on hiring, he starts by mapping the current team and what it might need a year later. That exercise can reveal that the real need is a product manager focused on activation or a performance marketer embedded in marketing, rather than a permanent independent team.

A growth hire needs to be comfortable using data and asking good questions, Grenier says, but need not personally build every model. The other skills depend on the company: customer understanding may matter most in one situation, technical building in another.

One interview exercise asks how much a marketplace should pay to acquire a rider or driver. A trip creates value for both, making it difficult to assign lifetime value without double-counting; different local shortages also change which side deserves investment.

Grenier says Uber's data scientists worked on that allocation without finding a perfect answer. He values the candidate's reasoning and ability to challenge assumptions more than a memorized formula.

A second exercise asks candidates to improve engagement: how would they bring in new users, increase frequency or improve the quality of engagement, and how would they prioritize and test those ideas? This exposes how they think about customers beyond acquisition.

His red flags include talking only about data without the customer problem, or copying a playbook from a previous company without adapting it. A useful growth leader, in his view, must explore an unfamiliar problem rather than simply execute a prescribed task.

Grenier discusses the marketplace question live to hear the thought process, while an engagement case can be a take-home exercise. For compensation, he suggests product compensation bands as a starting point, adjusted for how scarce and relevant a person's experience is.

A north star should connect as much of the company as possible and reflect the customer's problem. Uber's trips required both a rider and a driver, so a slowdown prompted the team to investigate pricing, supply, demand and the needs of individual markets.

Generic daily or monthly active-user measures can mislead when they do not match the natural frequency of the job being done. A MasterClass customer may engage daily during a course but take only a few courses in a year, Grenier observes.

For early-stage companies, he recommends beginning with a simple growth model rather than making a complex one an end in itself. A loop might connect customer acquisition, content creation, sharing and the arrival of another customer.

If the steps do not feed back into one another, the business may grow linearly and need sales or paid acquisition; Grenier says that can still be successful. A simple model also gives teams concrete elements against which to judge their priorities.

Stebbings asks about fragmented versus concentrated marketplace supply. Grenier finds concentrated supply easier to operate, but says the distributed long tail is more fun because of the variety of people and problems involved.

Uber's driver acquisition required many channels, not a single trick. The team studied how local operations teams used Craigslist, then looked for ways to scale job-board acquisition and adapt services such as Indeed and Monster to much larger volumes.

Research also revealed different driver motivations: professional drivers, people seeking flexible extra income, parents and retirees did not all respond to the same promise. The growth team brought customer research in early to understand these fragmented groups.

An experiment showed why the timing of an insight matters. Existing drivers often ranked flexibility above income, so the team shifted ads toward flexibility; Grenier recalls the ads performed poorly. Earnings attracted initial interest, while flexibility mattered more to engagement once someone was active.

His lesson is to know whom a research sample represents before applying its findings. What retains an existing customer is not necessarily what persuades someone unfamiliar with the product to try it.

For broad audiences, Grenier describes a messaging pyramid: a general promise such as earning money can run in mass channels, while more specific messages about safety or a local market can appear where targeting and economics justify them.

The time needed to evaluate a message depends on scale and the outcome being measured. Grenier says Uber could test many ads quickly at its spending volume, while Lambda School's smaller cohorts and long path from enrollment to employment required longer tests and more qualitative signals.

To define retention, he favors cohorts and the natural frequency of the use case rather than simply watching monthly active users rise. For a commuter segment, for example, a drop from an expected three trips a week to one or two can flag declining engagement before complete churn.

Where there is no growth team, Grenier wants data staff to provide product and marketing with proactive insight. The same early-warning cohort might lead marketing to try email or an event, while product investigates discovery, arrival times or other experience changes.

He thinks post-mortems should ideally happen regularly, including after successes, although early-stage teams often only conduct them after something breaks. The right stakeholders and a blameless effort to understand what happened matter more than a perfect template.

At Lambda School, his team introduced a student chat tool without involving customer service, which already had its own system. The lesson was to own the collateral effects of a change, bring affected teams in and learn how to repeat the useful work more effectively.

At Uber, local teams sometimes bought Facebook ads using personal credit cards. Grenier initially tried to stop them because they could bid against central campaigns, but came to see that local teams understood events and marketplace conditions better than a global buyer could.

A local team might care intensely about a festival that looked insignificant against a global budget. Grenier's team built ways to support city marketers, including reusable approaches for recurring local events, reducing friction while preserving local knowledge.

Paid acquisition should be a tool, not a crutch, Grenier says. He uses 30 percent as a rough anchor to prompt scrutiny of dependence on paid channels, not a universal ceiling: a brief market opportunity might warrant spending much more, while zero spend can leave a team unprepared.

Grenier recalls that Uber at one point spent about $1 billion a year on advertising, including China. He estimates that $600 million to $700 million spent in the rest of the world accounted for roughly 40 to 50 percent of growth at that time; these are his retrospective estimates, and the video title's monthly figure is not established by this passage.

China required learning different channels and a more fragmented app-store landscape. Grenier also alleged that competitors, including Didi, and the Chinese government had access to Uber data; the interview supplies no independent evidence establishing those claims. He describes unexpected participants on calls and a different operating environment as his experience.

Grenier says complete attribution across channels is impossible, especially with the limits imposed by large platforms. Customer acquisition cost can still help a media buyer make decisions when informed by a media mix model; the aim is greater confidence in a working estimate, not perfect attribution.

Enterprise purchases can be easier to reason about than noisy consumer marketplaces, he suggests, but consumer media buyers still need practical targets. Stebbings illustrates his own way of comparing a campaign's cost with customer lifetime value and the probability of a conversion.

As an angel investor, Grenier says he is drawn into a company's existential questions as well as its immediate channel choices. He often finds that the issue is not an undiscovered channel but product-market fit or the direction and scope of the product.

In a quick-fire round, Grenier says he no longer believes every company needs a growth team. He also criticizes Twitter's recommendations at the time for overemphasizing a user's apparent interests and argues for highlighting the timeliness that makes the service distinctive.

He interprets Clubhouse's early appeal as a strong fit for a moment when people wanted dinner-party-like conversation, followed by a mismatch as behavior and the market changed. This is his assessment, not a definitive explanation of the company's trajectory.

Asked in 2023 about TikTok's future, Grenier predicts it will endure and does not expect a ban, while acknowledging uncertainty. On hiring, he returns to his main warning: companies should not treat a growth hire as a substitute for a product people want.

Grenier credits his ability to connect ideas across experiences partly to studying improvisation and comedy. He admires Epic Games' combination of a platform in Unreal Engine and games built on it, seeing a possible parallel with early app platforms and emerging AI companies.

Stebbings thanks Grenier for the discussion, and Grenier closes by saying he enjoyed the conversation.