Adam Gross on Building Product-Led Growth as a Business Model

Adam Gross argues that product-led growth (PLG) is a business model, not simply a marketing channel. In his experience, good PLG companies find a distinctive acquisition mechanism beyond paid advertising; he is skeptical of early-stage companies relying on paid acquisition while calling themselves PLG.

Harry Stebbings begins by asking Gross what he learned from the companies where he worked. Gross joined Salesforce in 2003, when it was far smaller than the enterprise software company people know today. Its growth, he says, was built deliberately through the energy of its team rather than simply arriving through luck.

At Salesforce, a growth mindset meant taking ownership of growth every day. Gross contrasts deliberate marketing, messaging and a story about industry transformation with simply experiencing viral adoption. Those capabilities helped Salesforce keep finding growth as its scale increased.

Gross joined Dropbox in 2010 as employee 11. Even then, its unusual ability to acquire customers was apparent. He encourages founders to pay as much attention to innovation in customer acquisition as they do to new databases, application frameworks and engineering practices.

Asked where new acquisition opportunities might come from, Gross distinguishes enterprise software from consumer apps. He points to AI's potential in the customer journey after acquisition. As interim CEO of Vimeo, he found tools such as Gong useful for getting up to speed through recorded and analyzed sales calls.

More sales conversations now occur online and can be captured as data, creating opportunities for AI-assisted sales development. Gross treats this as an early possibility, not a solved acquisition channel: such tools may improve what happens after a prospective customer reaches the company.

Stebbings asks whether Dropbox has lost its acquisition advantage as growth slows. Gross, who had left more than a decade earlier, avoids diagnosing its present operations. Growth is difficult at scale, he says, and companies need new acts to grow from a large revenue base.

Stebbings cites the proposition that one channel gets a company to $50 million in annual recurring revenue and two get it to $100 million. Gross resists a rigid rule: he has seen product-led businesses, including Dropbox, go well beyond $50 million on one channel, and hopes to get more from a channel because building one is hard.

At Heroku, Gross saw the whole PLG model come together: a high-quality free product for individual developers and a strong community could coexist with enterprise customers paying millions of dollars a year. When his tenure began in 2013, he says, many people saw the two approaches as being in tension.

Gross distinguishes optimizing a particular flow or journey from setting a broader growth vision. He respects growth teams and has run them, but thinks local optimization alone can miss how the company creates transformative value and communicates it to customers at scale.

He asks founders for their theory of the business: where customers come from, how they become successful, and how both the customer and the company gain value. A free individual may become a paying team and later an enterprise customer; each transition belongs to an interconnected business architecture.

A founder may focus on a prototype on day one, but Gross thinks the economic model needs clarity early, perhaps in year two. Reaching the next stage through customers paying $50,000 each requires a very different system from reaching it through many customers paying $5,000 each.

Founder-led sales can make the first few large customers look deceptively repeatable. Getting the next 50 requires building a customer acquisition machine. A lower-priced model has its own difficulties and calls for a different acquisition expertise, so trying to build both too early can stretch a startup thin.

Stebbings sees companies selling at a $5,000 annual contract value while carrying the sales and customer-success costs of a $50,000 customer; expansion often fails to make up the difference. Gross agrees that PLG is not for everyone: it must shape the product, organization and economics, not be bolted on as an extra distribution channel.

Gross's rule of thumb is that a good PLG business has some specialized acquisition mechanism other than paid advertising. He says he would question a company relying on paid acquisition below $100 million in revenue. At $5 million to $10 million, he often sees one or two mechanisms generate a disproportionate share—perhaps 50% to 70%—of acquisition.

On diversification, Gross recommends caution. As a rule of thumb, he would build either a self-serve or an enterprise motion until at least $10 million in revenue, rather than introducing a second business too soon. Enterprise sales can be tempting but building the full organization and its supporting machinery is distracting and expensive.

Stebbings asks how to win an individual champion who can bring a product into an organization. Gross says the journey is not merely selling the same product to progressively larger groups. Each stage typically needs a distinct but complementary product experience, value proposition and go-to-market motion.

Gross describes a layer cake: an individual user's free product often delivers creation, as Heroku lets a developer build and deploy an app. A paying team needs collaboration; GitHub, for example, becomes part of a team's shared workflow rather than just a place for an individual to find code.

At the enterprise layer, compliance becomes central: security, auditability and observability matter to the organization. Those layers connect, but they are distinct products and propositions. A founder must deliberately build a reason for the individual user to become a team customer.

A common failure is to assume that a successful single-player product will automatically become a collaborative one. Gross says this effectively requires finding product-market fit twice. He has seen companies succeed with individual users but fail to translate that value into something a team needs.

How can a homepage communicate both creation and collaboration? Gross usually aims the website at the top of the funnel and the initial free customer relationship. Once someone is using the product, the company can communicate team and enterprise value through the product, events and other channels.

During Gross's 2013–2018 Heroku tenure, the main site focused on getting developers to sign up for free. Some enterprise products on which customers spent millions were barely visible to the typical developer audience. Enterprise customers instead learned about them through more targeted communication, though Gross acknowledges the separation may have gone too far.

Stebbings asks how a horizontal product can market itself when its users range across professions and household contexts. Gross starts with the company's narrative: how customers will imagine the change the product makes in their lives or work. He says the message needs to be both emotive and strategic.

Being emotive means giving people a reason to care about the experience. Gross points to developer brands such as GitHub, Heroku and Vercel, where craft and quality matter. Companies have to decide that user experience matters and attend even to apparently small details; at Heroku, executives debated command-line switches.

Emotion alone can make a product interesting without giving it the strategic impact needed to command a budget and grow. Gross argues for enterprise empathy: understanding how the product fits into the customer's organization, politics, business processes, successes and dysfunctions, rather than seeing everything through the vendor's own perspective.

Why do fast-growing companies plateau? Gross says an early idea can produce extraordinary success, but the beliefs and practices associated with it may not support the next stage. Serving customers more deeply can require challenging the very principles that helped the company break through.

A company can become too inward-looking and treat its own culture as the source of all growth. Gross warns against elevating semi-arbitrary habits into articles of faith. For him, a growth mindset means staying customer-focused enough to question cherished values and strategies when the next opportunity demands it.

Stebbings asks about Gross's interest in names. Gross sees positioning and naming as an invitation to understand and inhabit a product's vision. Customers have very little time to work out why a product matters, so the name should convey its essence and generate interest.

While working with GitHub, Gross says he helped name GitHub Actions. Alternatives centered on flows or workflows, but 'Actions' captured a shift from GitHub as a relatively inert store of code to a platform that could do things with that data. The name expressed both the change and why a customer might care.

Stebbings asks how sales teams avoid feeling second-class in a product-led business. Gross says trouble starts when self-serve and sales-led activities blur. Each needs clarity about its customer, product, value proposition and motion; otherwise teams interfere with one another and resentment follows.

The two businesses should be distinct but connected. Gross then considers how AI might change customer experience, beginning with familiar applications such as support triage and routing. The more interesting question, he suggests, is how far AI can go in educating customers and helping them through a journey.

Gross expects some AI value to accrue to companies that already have distribution. He contrasts AI with the shift from on-premises software to cloud services, which forced incumbents to change licensing, sales, finance, engineering and operations all at once. He sees AI as easier for an existing organization to adopt without overturning its entire structure.

That does not rule out new companies: Gross thinks genuinely new AI-native experiences and engagement models may come from elsewhere. He is also interested in selling outcomes rather than tools, such as support providers moving from charging per seat toward charging for resolved incidents. He wonders what equivalent pricing might look like in other industries.

Asked for advice to startup CEOs, Gross starts with hiring. Be clear whether the role needs a 'poet' or a 'librarian': someone with deep subject-matter and creative expertise, or someone who can build and run the processes that scale an operation. Most roles combine both, but confusion over which capability is missing leads to poor hires.

In marketing, the poet may know the space, use the product and speak fluently with customers; the librarian may know demand generation, systems, analytics and acquisition practices. Engineering has a similar distinction between a technically ingenious leader and a process leader who can help a growing organization run reliably.

His second recommendation is alignment. Planning frameworks can become bureaucratic, but their purpose is to focus an organization's scarce time and attention. Gross has used Salesforce's V2MOM, while stressing that the particular method matters less than whether people actually align around priorities.

For a simple starting point, Gross suggests bringing people from across the organization into a room and spending the time to agree on a ranked set of priorities. He describes using this approach with Tailscale's self-serve business. When it works, the result feels like clarity about what to do, not a spreadsheet exercise.

Gross recommends an annual planning cycle, with the same alignment exercise available whenever a new project or area needs it. His third recommendation is to establish a common operating cadence, which he calls the seasons of software: a rhythm that helps different teams coordinate their work.

One practical way to establish that rhythm is to announce a release every quarter, regardless of the underlying development schedule. A blog post or page describing the spring or Q1 release gives everyone a predictable moment to look up, communicate progress and align around the next period.

Turning to investing mistakes, Gross says he has sometimes underweighted team dynamics and assumed he knew things he had not verified. In one case he learned after writing a check that the founding team had not yet decided who the CEO would be. He should have asked explicitly.

Gross recalls seeing Twilio at the seed stage and not thinking big enough about its potential. He overestimated how commoditized its market would be, while acknowledging that the company then was a tiny team and that he cannot claim he definitely could have invested.

He also had early visibility into Stripe and thinks stronger effort through his relationships might have created an opportunity to invest, though he does not claim a deal was assured. His platform work at Salesforce put him close to early developer companies; he had backed Heroku from the beginning and was an early investor in Docker.

Gross says Docker's recapitalization left him unhappy as an early investor, but he is pleased that its team had a major impact through its vision. Stebbings then moves to quick-fire questions about the market, growth and Gross's career.

In April 2024, Gross says the macroeconomic environment has proved more uncertain than he expected and software growth is taking longer to recover. Stebbings describes his bet that IPO markets would not reopen until 2025 at the earliest; Gross chooses the more optimistic side of the bet. Their remarks are forecasts made at the time, not a report of what later happened.

Gross would like companies to treat growth as something that pervades the organization. He is cautious about specializing a growth team too early because optimizing local steps can distract from larger early-stage problems. He also describes enterprise software cultures as more customer-centric, predictable and controllable than consumer cultures.

His advice to someone leaving university is to become good at two things rather than the best at just one. Looking ten years ahead, Gross hopes to remain as excited and energized by technology as he has been throughout his career. He and Stebbings close by thanking each other and the listeners.