Phil Carter on Growth Loops and Consumer Subscription Economics

Harry Stebbings speaks with growth advisor and angel investor Phil Carter, whose background includes consumer investing and product and growth work at Ibotta and Quizlet. Carter focuses on consumer subscription businesses. He defines growth as getting a product into more users’ hands and helping them understand its value, so that conversion, retention and monetization improve.

Whether growth belongs in a separate team depends on how a particular business grows. Carter says Quizlet relied heavily on word of mouth and search, while a higher-priced product built around paid advertising would need a different team. Founders should understand their customers and channels before hiring a leader based on a generic template.

Stebbings raises the argument for hiring growth specialists before product-market fit to generate more data. Carter generally prefers hiring a growth leader around the point that fit is demonstrated, while acknowledging exceptions. A curious early engineer or product person can naturally become a growth leader by investigating unexpected patterns and running experiments.

The size of experiments should depend on company maturity. Early-stage teams often lack enough users to measure small changes and should pursue larger improvements; mature teams at scale can justify small conversion gains worth substantial revenue. Carter uses the S-curve as a way to distinguish early adoption, rapid growth and eventual saturation.

Carter says his own hiring mistake was overvaluing narrow specialist backgrounds when recruiting product managers. He now favors smart, curious generalists who move quickly and take considered risks. The ideal first growth hire still depends on the actual acquisition and product dynamics of the business.

For take-home assignments, he tries to avoid bias from prior knowledge of the company or unlimited free time. He prefers a hypothetical or familiar product and a problem that should take only one or two hours, testing first-principles reasoning and quantitative thinking rather than a polished presentation.

He asks candidates which growth teams they admire and about an unexpected experiment: what happened, what they learned and how their roadmap changed. A failed A/B test can be valuable when it reveals a mistaken assumption and leads to a better next experiment, he says.

At Quizlet, Carter’s team initially assumed UK search growth was constrained by insufficient locally relevant content. After failed attempts to encourage more creation, it found UK content-creation rates were actually higher than US rates. The bottleneck was technical SEO: too little content was crawled, indexed and ranked.

The team shifted toward changes such as country subfolders and authoritative links, which Carter says improved UK search growth. The lesson is to inspect the entire growth loop and its inputs before choosing a solution. His account does not isolate the measured effect of each change.

A common metric mistake is focusing only on outputs such as ARR or subscriber count. Growth teams should connect those outputs to inputs they can affect, such as activation, trial starts, conversion and retention. Carter recommends looking at the company’s historical baseline and using external benchmarks as a starting point, not a substitute for understanding the product.

Consumer subscriptions can be relatively easy to launch: many do not require a sales team or a two-sided marketplace, and app stores provide distribution and payments. Carter says scaling is harder because platforms control part of the customer relationship, paid acquisition becomes crowded, and many products face low revenue per user and high churn.

Citing RevenueCat data, he says more than half of annual subscribers typically leave within the first year and more than half of monthly subscribers within three months. These are figures attributed to a dataset he references, not a guarantee for every category. Low retention can make lifetime-value estimates fragile and limit how much a business can pay to acquire customers.

Stebbings asks whether customer-acquisition cost falls as a brand grows or rises as a market saturates. Carter identifies viral outliers where blended cost may fall, but expects most companies to face higher costs as they move beyond the highest-intent users and spend more on paid channels. His view is a generalization with exceptions.

Very small ad budgets may not produce enough signals for platform algorithms, he says, while rapidly increasing spend can bring lower-intent users and weaker inventory. For early companies with limited capital, he recommends pursuing organic product-channel fit first and running modest paid tests, rather than assuming money alone can buy sustainable growth.

Once a primary channel works, Carter suggests concentrating on it while tracking conversion and efficiency at each step. Stebbings worries that waiting too long leaves a company exposed to saturation. Carter agrees: if leading indicators suggest a bottleneck six to twelve months away and it cannot be solved, start developing a second channel before the first stalls.

He allows that some marketplaces with strong network effects can subsidize growth until they reach a tipping point, but says that is less often viable for consumer subscriptions. For paid acquisition he calls six-month payback good and one month great, and generally aims for three months or less when advising clients. These are practitioner heuristics, not universal benchmarks or guarantees.

Carter cites a 2024 RevenueCat report that more than 75% of trial starts occur within 24 hours of installation. Together with high churn, he says, this makes early conversion and payback important. The figures depend on the dataset and its definitions; the interview does not establish the same rate for every app.

He separates monthly and annual subscriber retention. For monthly plans he considers retaining more than half for six months strong; for annual plans he examines the first two renewals and two to three years of retention. He also looks for a curve that stabilizes or recovers through reactivation, as users may return to a product after a dormant period.

Stebbings argues that narrow consumer subscription markets can be difficult because churn exhausts a small audience. Carter agrees that venture-scale outcomes often require a large market, but says a smaller profitable company can still be attractive without venture-scale ambitions. Revenue, valuation and investor returns should not be treated as interchangeable.

He estimates that a list he compiled included roughly 30–40 consumer subscription companies with valuations above $1 billion, including both public and private companies. Stebbings notes private secondary prices may be lower than headline valuations. Carter acknowledges he has not seen most of those companies’ underlying numbers; his explanation for why some remain private is speculation.

Carter considers lower revenue multiples understandable given historical outcomes, churn and limited expansion revenue. He expects potential changes in app-store fees and AI distribution to create opportunities, but these are forecasts. He discloses that he is an investor in fitness app Ladder, one of his examples of TikTok-led growth.

A core value promise is the distinct, enduring benefit a customer receives. For subscriptions, differentiation supports pricing and an enduring need supports retention. Carter says some tracking apps face diminishing value once users understand their habits; that is a product observation, not a medical conclusion about any device.

He believes hardware can increase commitment and upfront revenue for wearables, while frequent new features and insights can reinforce ongoing value. These are his interpretations of companies such as Oura and Whoop, not independently verified retention experiments or claims that the devices guarantee health outcomes.

Streaks, badges and leaderboards are tactics, not a substitute for understanding motivation. Carter describes Duolingo in terms of accomplishment, ownership, social influence and avoiding the loss of a streak, drawing on the Octalysis framework. Different products should use different mechanisms based on why their users return.

Notifications can increase short-term engagement but damage the channel if users feel overwhelmed, he says. He wants each notification to improve the experience and produce a meaningful benefit, not merely a statistically detectable lift. Stebbings counters that sending fewer messages may lose attention in an already crowded environment.

Carter responds that excessive reminders can lead users to disable notifications or delete the app. He describes Duolingo throttling messages when a user is inactive and increasing them after the user returns. Both speakers reject using fear or false urgency as a habitual tactic because short-term conversion can erode credibility.

Paywall view rate is the share of app installers who see the paywall at least once, ideally within a specified period. Carter cites Superwall founder Jake Moore’s work and suggests more than 80% visibility in the first session or week as a goal for many early subscription apps. This is a heuristic to test alongside user experience and retention, not a universal requirement.

The choice between a hard paywall and freemium depends on alternatives, willingness to pay, price and acquisition strategy. Quizlet benefits from free users who create content and bring search and word-of-mouth traffic, so an aggressive paywall could hurt its growth loop. A differentiated product with paid acquisition and a quick purchase decision may justify an earlier paywall, he says.

Carter separates subscription tiers from durations. He believes most consumer products should start with one paid tier, reserving multiple tiers for mature products with genuinely distinct value propositions. Monthly and annual plans are common; education may justify terms aligned to a school calendar, while weekly plans can have high churn.

He recommends revisiting price at least annually through analysis or testing, while keeping packaging simple enough for users to understand. Multiple tiers make sense only when the audience and benefits differ sufficiently. His examples are advice based on experience, not a promise that a higher price will improve every business.

Carter cites Tinder as a case of differentiated tiers and additional purchases for highly engaged users, drawing on former product leader Ravi Mehta’s explanation. Stebbings disagrees with the clarity of its packaging and says he found the plans confusing. Carter acknowledges he has not personally used Tinder as a typical customer, limiting his firsthand assessment.

Onboarding should get users to value quickly, but Carter has changed his mind about always making it short. He describes Noom’s lengthy, personalized questionnaire as an example where a considered health purchase may benefit from building confidence and intent over multiple screens. Neither speaker has the abandonment data; Carter’s proposed pattern of early drop-off and later commitment is a hypothesis.

Targeted discounts can match different willingness to pay, but indiscriminate promotions can erode brand and revenue. Carter describes activity-based offers after a user has not converted, family or student plans, and seasonal promotions. He also discusses web checkout economics; the feasibility and fees depend on platform rules and circumstances, which the interview does not establish.

For seasonal products, companies can use quiet periods to improve the product, reduce technical debt or prepare a launch. Larger businesses may diversify geographically across school calendars or add use cases that are relevant in other seasons. He says Quizlet saw some offset as it gained users in Asia and Latin America.

In the quick-fire round, Carter warns against raising too much or paying for acquisition before understanding the product, customer and unit economics. He also rejects both “build it and they will come” and the idea that every company can buy its way to success. Distribution and a sustainable growth loop still need deliberate work.

He says the old playbook of raising a Series A and scaling only on Facebook has become harder because of competition, funding and attribution changes, though he sees signs of improvement. On SEO and AI answer engines, he expects disruption but believes relevant content and authority remain useful. Those are his views and predictions from 2024.

Carter reiterates his changed view on onboarding: longer flows may work when each step adds useful personalization and builds intent, especially in considered categories. Length alone is not the goal; unnecessary steps can still lose users.

His example of a recent growth strategy is Ladder, in which he is an investor. He says coaches with existing audiences create organic social content, the strongest content is promoted, and web onboarding helps the company understand users. He reports sixfold subscriber growth during 2023 and more than 100,000 subscribers, figures not independently verified in the interview.

Stebbings questions whether even $100 million in revenue would produce a venture-scale return if the market applies a low multiple. Carter concedes that time will tell. His investment thesis rests on company metrics he believes exceed benchmarks and a team willing to change channels after paid Facebook and Google efforts failed; his financial interest should be considered when assessing the case.