Hila Qu
Executive in Residence at Reforge, growth advisor, former Director of Growth at GitLab and VP of Growth at Acorns
Where to hear and read about Hila Qu1 episode or test
Hila Qu is Executive in Residence at Reforge and a growth advisor who was Director of Growth at GitLab and VP of Growth at Acorns. The thing most worth learning from her is to treat product-led growth as data-led growth: build usage data first, then fix activation, then checkout, and staff a committed growth team.
PLG, I always say, is actually fundamentally DLG, data-led growth.
The ultimate guide to adding a PLG motion | Hila Qu (Reforge, GitLab)
Hila Qu: the story and the lessons
1Who Hila Qu is and where her growth experience comes from
Qu scaled PLG at GitLab and took Acorns to 5 million users. She is Executive in Residence at Reforge and a growth advisor. She was Director of Growth at GitLab and VP of Growth at Acorns, where users grew from 1 million to 5 million.
GitLab had an enterprise sales team and an open source product from the start, so she learned PLG by running it beside a strong sales funnel.
Qu writes long, careful pieces because she believes in leverage. Her four-month series of guest posts on adding PLG for Lenny's Newsletter brought in readers, contacts and clients. Both parts rank among the 25 most read posts in the newsletter's history, and one advisory client arrived because of the series.
She learned about leverage from Naval, who names four types: writing, code, capital and team. That is why she invests in writing.
2When Qu says PLG fits and why it fails
Qu says most B2B companies need both product-led and sales-led motions. PLG gives volume because a free product lowers the barrier. Sales gives big, targeted deals. A sales-led company that ignores PLG loses end-user advocates to rivals, and a PLG-only company loses big customers.
She never believed the "PLG is the future, no sales needed" view. Slack, Figma and LastPass all added sales teams, and sales-led enterprise products now add a product-led layer.
Qu tests fit by product complexity and the number of possible users. The product should show value without heavy customization, or its time to value must be shortenable. There should be many end users or small businesses who want to try it.
Software for Boeing or defense firms, with three buyers worldwide, should stay sales-led. Most B2B software sits mid-spectrum.
Qu says a basic free trial is not a PLG motion. Some companies spend three months on a simple trial and expect leads, conversion and self-serve revenue to follow. She advises a one to two year roadmap, built after thinking and data-gathering have created conviction.
"You need to commit to at least a year or even 2 years kind of road map to build this entire thing out."
Qu names three red flags that a company is not committed. The flags are assuming a free version or trial is the whole job, handing the work to one person who must borrow help from sales, marketing and product, and doing PLG because it is trendy.
She says that one person "needs to be a magician". Strong sales skills do not carry over to designing a PLG journey, so she advises hiring someone who has done it or working with an advisor.
Qu says a book-a-demo button cuts off the PLG entry point. The product needs a free version, a free trial or open source. If none is possible, a realistic interactive demo gets close.
Amplitude was hard to try because users had to install code, so it built an interactive demo. She notes the demo is not PLG, but it lets users see value.
3Qu's order of work: data, audit, activation, checkout, then PQLs
Qu builds usage data before anything else. A free product should return broader reach and an understanding of which features correlate with conversion and retention. Usage data is the most essential piece, and sales-led B2B companies usually lack it because they never needed it.
"If you don't have a foundation of data and an understanding of how to analyze those data, you are giving away a free product for nothing."
Qu audits tracking and writes a data dictionary before choosing tools. She walks the product, lists the key actions and checks each is tracked correctly. The audit ends in a dictionary of event names and properties that everyone shares.
Bad data leaves analysts unsure what to trust: "It's garbage in, garbage out." She says picking the analytics tool carries little risk, since it is hard to get completely wrong.
Qu audits the funnel as a semi-interested buyer. She goes through the website, sign-up, first use, aha moment and checkout, reads the first emails, then asks for high-level numbers at each step. Audits almost always turn up low-hanging fruit.
One client's checkout form asked questions only UK customers needed, and a US user dropped off. She finds that excited new users left lost on first entry is usually a big opportunity.
Qu starts with activation and gives users a warm start. Most B2B software was never built for quick first use. She sets the aha metric as the goal post, removes friction and offers sample templates, with emails to bring back users who stall.
Her rule is that doing beats showing, and showing beats telling. Miro is her model: a few targeted use-case questions, then a template, so users are working within about 5 minutes.
Qu wants checkout as easy as Lululemon or Amazon. She keeps pricing simple, makes the buy button easy to find and localizes payment options. She tests heavily here because the revenue is so close to being booked.
One company had a low payment success rate in India because its provider supported the market poorly. Adding a second provider improved it right away.
Qu holds PQL work until activation and checkout are decent. Low-priced products can convert by credit card. When an account fits the ideal customer profile and usage is high, it gets white glove outreach from sales to close a bigger deal.
The PQL effort needs data signals, customer criteria and a sales handoff, so it is much bigger than the two low-hanging fruits. At GitLab, a large company's usage triggered a sales email that became a contract.
4What Qu did at GitLab and Acorns
Qu found GitLab's activation milestone by testing 10 actions. She compared each high-value action against 90-day conversion and 30-day retention, because one metric alone misleads. The result was two users using two features within 14 days.
The first user inviting a coworker signals trust, and two features signal platform value. She used 14 days because "we're not Facebook". She ran experiments afterward: "In data you are only isolating correlation. You are not proving causation."
Qu turned Acorns retention into an adoption problem. Recurring investment correlated strongly with retention, so experiments pushed more people to set it up. Acorns then added a retirement account and a spending account with a debit card, which are used more often.
Nobody cared about recurring investment at the time, since the round-up feature got the attention. An IRA carries tax consequences, so users find it very hard to leave. The question became how to drive adoption of higher-frequency use cases.
Qu treats expansion as three levers triggered by usage data. Users can move to a higher tier, buy more seats or consume more add-ons. Usage data shows the right moment and the right person to prompt.
GitLab showed all three: bronze to silver tier, more seats and more CI minutes. The same activation and conversion tactics apply.
5How Qu staffs and runs a growth team
Qu hires the analyst before or alongside the growth PM. Without insights and a solid foundation, experimentation is directionless. A small company can't expect to find one unicorn who does the analysis end to end.
The core squad is a growth PM lead, an analyst, a dedicated engineer and a designer. The designer can be part-time early on, but the engineer must be dedicated.
Qu prefers a growth lead from inside the company. A PM who wants growth, or an analyst moving into product, can fill the role. An outside hire should match the starting focus, such as activation experience if activation is the priority.
One client's head of growth was the head of investor relations. He reported to the CEO, was analytical and could rally resources, and she advised him on what he didn't know.
Qu gets early wins first, then builds a PLG org. Wins build confidence across the company, and that is the time to add heads of growth product, growth marketing and product-led sales. Each owns different metrics, such as activated teams, quality sign-ups and PQL conversion.
Each head usually reports into their own function and collaborates closely because they share one funnel.
Qu uses a tiger team when the first focus is PQLs. A PQL project needs data, sales and marketing counterparts that one growth PM struggles to pull in. Activation and conversion stay inside product, so a dedicated squad fits.
She combines a dedicated growth PM with a temporary cross-functional tiger team before full commitment.
Qu interviews growth hires on an unexpected experiment result. She asks for an experiment with a very unexpected result and what the candidate did next. It shows they run many experiments, and why it surprised them shows how well they know customers.
She avoids asking for a success story because most people prepare that answer well.
Qu requires a success metric and growth lever in every ticket. She added a section to the ticket spec where PMs state the metric up front and whether the work helps acquisition, activation, retention or monetization.
Writing it down made some PMs drop the work, because they couldn't say why they were doing it.
The ultimate guide to adding a PLG motion | Hila Qu (Reforge, GitLab)
A step-by-step guide to adding a product-led growth motion on top of a sales-led business: audit the funnel, start with activation, build a data foundation, and staff a growth squad. Hila Qu draws on scaling PLG at GitLab and taking Acorns from 1 million to 5 million users.
Qu scaled PLG at GitLab and took Acorns to 5 million users. She is Executive in Residence at Reforge and a growth advisor. She was Director of Growth at GitLab and VP of Growth at Acorns, where users grew from 1 million to 5 million.
GitLab had an enterprise sales team and an open source product from the start, so she learned PLG by running it beside a strong sales funnel.
Qu writes long, careful pieces because she believes in leverage. Her four-month series of guest posts on adding PLG for Lenny's Newsletter brought in readers, contacts and clients. Both parts rank among the 25 most read posts in the newsletter's history, and one advisory client arrived because of the series.
She learned about leverage from Naval, who names four types: writing, code, capital and team. That is why she invests in writing.
Qu says most B2B companies need both product-led and sales-led motions. PLG gives volume because a free product lowers the barrier. Sales gives big, targeted deals. A sales-led company that ignores PLG loses end-user advocates to rivals, and a PLG-only company loses big customers.
She never believed the "PLG is the future, no sales needed" view. Slack, Figma and LastPass all added sales teams, and sales-led enterprise products now add a product-led layer.
Qu tests fit by product complexity and the number of possible users. The product should show value without heavy customization, or its time to value must be shortenable. There should be many end users or small businesses who want to try it.
Software for Boeing or defense firms, with three buyers worldwide, should stay sales-led. Most B2B software sits mid-spectrum.
Qu says a basic free trial is not a PLG motion. Some companies spend three months on a simple trial and expect leads, conversion and self-serve revenue to follow. She advises a one to two year roadmap, built after thinking and data-gathering have created conviction.
"You need to commit to at least a year or even 2 years kind of road map to build this entire thing out."
Qu names three red flags that a company is not committed. The flags are assuming a free version or trial is the whole job, handing the work to one person who must borrow help from sales, marketing and product, and doing PLG because it is trendy.
She says that one person "needs to be a magician". Strong sales skills do not carry over to designing a PLG journey, so she advises hiring someone who has done it or working with an advisor.
Qu says a book-a-demo button cuts off the PLG entry point. The product needs a free version, a free trial or open source. If none is possible, a realistic interactive demo gets close.
Amplitude was hard to try because users had to install code, so it built an interactive demo. She notes the demo is not PLG, but it lets users see value.
Qu builds usage data before anything else. A free product should return broader reach and an understanding of which features correlate with conversion and retention. Usage data is the most essential piece, and sales-led B2B companies usually lack it because they never needed it.
"If you don't have a foundation of data and an understanding of how to analyze those data, you are giving away a free product for nothing."
Qu audits tracking and writes a data dictionary before choosing tools. She walks the product, lists the key actions and checks each is tracked correctly. The audit ends in a dictionary of event names and properties that everyone shares.
Bad data leaves analysts unsure what to trust: "It's garbage in, garbage out." She says picking the analytics tool carries little risk, since it is hard to get completely wrong.
Qu audits the funnel as a semi-interested buyer. She goes through the website, sign-up, first use, aha moment and checkout, reads the first emails, then asks for high-level numbers at each step. Audits almost always turn up low-hanging fruit.
One client's checkout form asked questions only UK customers needed, and a US user dropped off. She finds that excited new users left lost on first entry is usually a big opportunity.
Qu starts with activation and gives users a warm start. Most B2B software was never built for quick first use. She sets the aha metric as the goal post, removes friction and offers sample templates, with emails to bring back users who stall.
Her rule is that doing beats showing, and showing beats telling. Miro is her model: a few targeted use-case questions, then a template, so users are working within about 5 minutes.
Qu wants checkout as easy as Lululemon or Amazon. She keeps pricing simple, makes the buy button easy to find and localizes payment options. She tests heavily here because the revenue is so close to being booked.
One company had a low payment success rate in India because its provider supported the market poorly. Adding a second provider improved it right away.
Qu holds PQL work until activation and checkout are decent. Low-priced products can convert by credit card. When an account fits the ideal customer profile and usage is high, it gets white glove outreach from sales to close a bigger deal.
The PQL effort needs data signals, customer criteria and a sales handoff, so it is much bigger than the two low-hanging fruits. At GitLab, a large company's usage triggered a sales email that became a contract.
Qu found GitLab's activation milestone by testing 10 actions. She compared each high-value action against 90-day conversion and 30-day retention, because one metric alone misleads. The result was two users using two features within 14 days.
The first user inviting a coworker signals trust, and two features signal platform value. She used 14 days because "we're not Facebook". She ran experiments afterward: "In data you are only isolating correlation. You are not proving causation."
Qu turned Acorns retention into an adoption problem. Recurring investment correlated strongly with retention, so experiments pushed more people to set it up. Acorns then added a retirement account and a spending account with a debit card, which are used more often.
Nobody cared about recurring investment at the time, since the round-up feature got the attention. An IRA carries tax consequences, so users find it very hard to leave. The question became how to drive adoption of higher-frequency use cases.
Qu treats expansion as three levers triggered by usage data. Users can move to a higher tier, buy more seats or consume more add-ons. Usage data shows the right moment and the right person to prompt.
GitLab showed all three: bronze to silver tier, more seats and more CI minutes. The same activation and conversion tactics apply.
Qu hires the analyst before or alongside the growth PM. Without insights and a solid foundation, experimentation is directionless. A small company can't expect to find one unicorn who does the analysis end to end.
The core squad is a growth PM lead, an analyst, a dedicated engineer and a designer. The designer can be part-time early on, but the engineer must be dedicated.
Qu prefers a growth lead from inside the company. A PM who wants growth, or an analyst moving into product, can fill the role. An outside hire should match the starting focus, such as activation experience if activation is the priority.
One client's head of growth was the head of investor relations. He reported to the CEO, was analytical and could rally resources, and she advised him on what he didn't know.
Qu gets early wins first, then builds a PLG org. Wins build confidence across the company, and that is the time to add heads of growth product, growth marketing and product-led sales. Each owns different metrics, such as activated teams, quality sign-ups and PQL conversion.
Each head usually reports into their own function and collaborates closely because they share one funnel.
Qu uses a tiger team when the first focus is PQLs. A PQL project needs data, sales and marketing counterparts that one growth PM struggles to pull in. Activation and conversion stay inside product, so a dedicated squad fits.
She combines a dedicated growth PM with a temporary cross-functional tiger team before full commitment.
Qu interviews growth hires on an unexpected experiment result. She asks for an experiment with a very unexpected result and what the candidate did next. It shows they run many experiments, and why it surprised them shows how well they know customers.
She avoids asking for a success story because most people prepare that answer well.
Qu requires a success metric and growth lever in every ticket. She added a section to the ticket spec where PMs state the metric up front and whether the work helps acquisition, activation, retention or monetization.
Writing it down made some PMs drop the work, because they couldn't say why they were doing it.