LLenny's Podcast
← All episodes
Hila Qu (Reforge, GitLab)02 April 2023

The ultimate guide to adding a PLG motion

8Frameworks
14Insights

Frameworks in this episode

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Myth Buster· 1

Myth Buster18:00

PLG Isn't for Everyone — The Defense Contractor Exception

Countering the 'every company should do PLG' assumption, Hila argues fit depends on the business. PLG needs a low-complexity product with short time to value and a large pool of self-interested end users or SMBs. A company selling to Boeing or defense buyers — where only a handful of customers exist worldwide — probably shouldn't do PLG at all.

  • Fit requires low product complexity and little customization to see value
  • Time to value must be short (or made short)
  • You need a large pool of potential end-user/SMB buyers
  • Extreme case: software with only three possible customers worldwide shouldn't pursue PLG
  • Most B2B software sits in the middle of the spectrum, not at the defense extreme

Do you have a product that's relatively low complexity, doesn't require a lot of customization for the customer to see value?

Hila Qu · 18:00

if you, for example, develop a software for the air fly companies, like a Boeing Boeing, or defense companies, like only three target customer exist…

Hila Qu · 18:30
#plg#strategy#product-market-fit#b2b

Hot Take· 2

Hot Take00:00

PLG Is Really DLG — Data-Led Growth

Hila's core reframe: product-led growth only works if it's built on a data foundation. Giving away a free product buys you two things — broader reach and visibility into usage behavior — but the usage data is worthless if you can't collect and analyze it. Companies that launch a free product without instrumentation are giving it away for nothing.

  • PLG's real value is the usage data you get in exchange for a free product
  • Two returns from a free product: broader reach and understanding of usage behavior
  • You must know which features correlate with higher conversion and retention
  • Without a data foundation, a free product generates no leverage

PLG, I always say, is actually fundamentally DLG, data-led growth.

Hila Qu · 00:00

If you don't have a foundation of data and an understanding of how to analyze those data, you're giving away a free product for nothing.

Hila Qu · 00:30
#plg#data#growth#conversion
Hot Take05:30

You Don't Need to Be a PLG Purist — Most Companies Need Both

Hila rejects the idea that PLG replaces sales. In practice most startups run both motions: PLG lowers the barrier for volume and reach, while a sales team goes after a targeted list of big customers for strong revenue. Eventually every company needs both, and it's far easier to add sales on top of PLG than to bolt PLG onto a pure sales-led company.

  • PLG is a volume game; sales is targeted at big, high-revenue accounts
  • Sales-led incumbents lose end-user advocates if competitors add PLG
  • Pure PLG companies lose time reaching and closing big customers
  • Adding PLG to a pure sales-led org is the harder direction of change

PLG motion is perfect for lowering the barrier for more people to try, broader the reach.

Hila Qu · 05:30

But, I would say it is easier if you have PLG from early on. If you're pure sales-led, you try to add PLG, that's the…

Hila Qu · 07:30
#plg#sales#strategy#b2b

Explainer· 6

Explainer10:00

What Actually Makes a Product Product-Led (The Zoom Test)

Using Zoom as the example, Hila lays out the defining properties of a PLG product. Users discover and start using it without a sales call, then upgrade themselves when they hit a limit. The key traits: a very low barrier to entry (free version or trial), no need for boss approval, a self-service checkout, and a product that spreads on its own.

  • Users experience value before they ever know they're buying (Zoom webinar example)
  • Low barrier to entry: free version or free trial
  • No approval needed — you can start using it today
  • A self-service checkout flow lets users upgrade themselves
  • The free product spreads organically

So, I think the key properties of a PLG product, think about it should have a very low barrier to entry.

Hila Qu · 10:30

You don't need get get approval from your boss to use it. You can use it today.

Hila Qu · 10:30
#plg#onboarding#self-serve#product
Explainer25:30

PLG Funnel vs Sales Funnel: Usage Is the Leading Indicator

Hila contrasts the two funnels. In the traditional sales-led funnel, marketing scores leads by how much they interact with campaigns (opening emails, reading whitepapers) before handing MQLs to sales. In the PLG funnel, the biggest difference is that product usage — not campaign engagement — becomes the leading indicator of success, because users can now try the product before any contract is signed.

  • SLG: marketing qualifies leads by campaign interaction, then hands MQLs to sales
  • PLG: users sign up for a free version and the goal is to get them using the product
  • Product usage is the leading success indicator in PLG
  • The old 'no access before contract' barrier is now artificial
  • Two PLG conversion paths follow from usage: self-serve credit-card purchase, or a PQL/PQA sales-assisted path for ICP-fit accounts

The most important thing, the biggest difference is now you want them to use the product.

Hila Qu · 27:00

And the usage, product usage is the almost like the leading indicator for success for PLG.

Hila Qu · 27:30
#plg#funnel#sales#metrics
Explainer38:30

GitLab's Aha Moment: Two Users, Two Features in 14 Days

Hila defines the aha moment as the first time a user experiences a product's value, referencing Facebook's classic '10 friends in 7 days.' For GitLab, analysis landed on two users using two features within the first 14 days — the two-user signal captures the team/collaboration nature of the product, and two features signals the platform value. The window is quick but realistic for a complex product.

  • Aha moment = the first time a user experiences the product's value
  • Facebook benchmark: add 10 friends in 7 days
  • GitLab's milestone: two users, two features in the first 14 days
  • Two users signals a user confident enough to invite a coworker (team value)
  • Aha moment and activation are often used interchangeably

I think it as a moment as a first time a user experience the value of your product.

Hila Qu · 38:30

We ended up have something along the line of two users, two features used in the first 14 days.

Hila Qu · 40:00
#activation#aha-moment#gitlab#metrics
Explainer41:30

Finding Your Aha Moment — And Why Correlation Isn't Causation

Hila explains the method for finding an activation metric: brainstorm 10 candidate high-value actions, run correlation analysis on each against 90-day conversion and 30-day retention, and compare against the average to find the standouts. But the crucial caveat — correlation isn't causation. You must run experiments to actually drive those actions before you can trust that they cause conversion.

  • Brainstorm ~10 candidate high-value actions (e.g. merge a PR, run a pipeline)
  • Measure each against 90-day conversion AND 30-day retention — look at both, not one
  • The standouts that lift conversion and retention are aha-moment candidates
  • Data only isolates correlation, not causation
  • Experimentation is the step that finally validates the metric

Because in data you are only isolating correlation. You are not proving causation.

Hila Qu · 44:00
#activation#data#experimentation#metrics
Explainer36:30

The Full PLG Funnel Audit Finds Surprising Low-Hanging Fruit

Before picking where to invest, Hila runs a full-funnel audit for her advisory clients: she plays the end user from website through sign-up, activation, and checkout, noting every point of confusion. Paired with high-level step-by-step data (how many reach each stage), the audit reliably surfaces cheap, high-impact fixes — like a checkout form asking UK-only questions of US customers, or users landing in the product with no idea what to do.

  • Play the end user through the entire journey: website, sign-up, activation, checkout
  • Also read the first few onboarding emails — they can rescue a frustrated user
  • Overlay high-level funnel data on the experience to find the biggest drop-offs
  • Real examples: a confusing checkout form, and users lost at first login
  • Activation and conversion are the two most common starting places

you you you will you will never believe like when I do this audit, there are so many low-hanging fruits usually in this process.

Hila Qu · 37:00
#audit#funnel#conversion#activation
Explainer56:00

Why Retention Is 'The Messy Middle' — Build Habit and Frequency

Hila explains why she doesn't start with retention: unlike fast, high-leverage activation and conversion, retention plays out over a long, messy period where customers can churn at any moment. The way in is to build a habit, which first requires the product to have high enough frequency — a once-a-month product is very hard to turn into a habit. At Acorns, passive 'set it and forget it' investing made this especially hard.

  • Activation and conversion are fast and high-leverage; retention is slow and messy
  • Customers can be retained or churned at any given moment
  • Habit-building is the key to retention
  • A product must have high enough frequency to become a habit
  • Acorns' passive investing meant users didn't need to return — a retention nightmare

I call retention the messy middle.

Hila Qu · 56:00

Like the the key to do that is first of all, your product need to have a high enough frequency.

Hila Qu · 57:00
#retention#habit#frequency#acorns

Story· 2

Story31:30

How GitLab's PLG Funnel Actually Works

Hila walks through GitLab's real PLG motion. A developer discovers GitLab, signs up for a free account, and uses it for a personal side project — usage with no connection to their employer. Later, when the company looks to consolidate DevOps tools, that engineer raises their hand, runs a 30-day free-trial proof of concept, and either self-serve buys a few seats or triggers a sales conversation for a larger deal.

  • Individual developers generate usage on personal projects first
  • That bottom-up usage is disconnected from the employer initially
  • An internal champion surfaces GitLab when the company evaluates tools
  • A 30-day free trial unlocks advanced features for a proof of concept
  • Small teams self-serve checkout; bigger accounts trigger a sales reach-out on the usage signal

And so, you begin to see this individual users having some usage. But nothing to do with his company.

Hila Qu · 32:00
#plg#gitlab#bottom-up#sales
Story1:01:00

How Acorns Flipped Retention Into an Activation Problem

At Acorns, Hila found the biggest retention lever was actually activation. Data showed 'recurring investment' — an overlooked feature — correlated highly with retention, so she ran experiments to get more users setting it up, with fast success. Then she added higher-frequency use cases like IRA retirement accounts and a debit/spending account, reframing 'how do I improve retention' into 'how do I drive adoption of higher-frequency use cases.'

  • Recurring investment quietly correlated with retention where flashier features didn't
  • Experiments to drive recurring-investment setup succeeded quickly
  • Added higher-frequency use cases: IRA retirement accounts and a spending/debit account
  • IRA accounts have tax consequences, making users hard to leave
  • Retention became an adoption/activation problem for high-frequency features

So now you change the problem from how do I improve retention to how do I I drive adoption of higher frequency use cases.

Hila Qu · 1:02:30
#retention#acorns#activation#experimentation

Tool· 1

Tool1:05:00

Hila's PLG Tool Stack: Data, Experimentation, Lifecycle Marketing

Hila lays out the essential PLG infrastructure and the add-on tools by funnel stage. The three infra pieces: a data hub (Segment), a product analytics tool (Amplitude, PostHog, Mixpanel, Pendo), an experimentation tool (Optimizely, Appcues), and a lifecycle marketing tool distinct from lead-nurturing tools like HubSpot. On top: data enrichment for acquisition (ZoomInfo, Clearbit), onboarding builders for activation (Appcues, Userpilot), and PLS tools for conversion (Endgame, Pocus, Toplyne, Pace).

  • Data hub / collection: Segment
  • Product analytics: Amplitude, PostHog (open source), Mixpanel, Pendo
  • Experimentation: Optimizely, Appcues; Amplitude has experimentation components
  • Lifecycle marketing tool — behavior-triggered, distinct from HubSpot lead-nurture email
  • Acquisition enrichment: ZoomInfo, Clearbit; Activation: Appcues, Userpilot; Conversion/PLS: Endgame, Pocus, Toplyne, Pace

So data tool, experimentation tool, life cycle marketing tool, those are the infra.

Hila Qu · 1:07:00
#tools#data#experimentation#stack

Takeaway· 2

Takeaway1:11:00

Audit Your Data Instrumentation Before Buying a Tool

Hila's counterintuitive advice: don't start with the tool. Because it's garbage-in, garbage-out, a product analytics tool is useless without clean, correctly-formatted data. The first step is auditing your data instrumentation — which key actions are tracked, is the format right, where are the gaps — often requiring re-instrumentation. The end goal is a 'data dictionary' documenting every key action, its event name, and properties so everyone shares one definition.

  • A product analytics tool is meaningless without collected, clean data
  • Garbage data in makes analysts trust the data even less
  • First step: audit which key actions are tracked and whether the format is correct
  • You may need to re-instrument and re-format before plugging in a tool
  • Goal is a data dictionary: key actions, event names, properties — one shared definition

If you send a bunch of garbage data into your product analytics tool, your analyst will be just even more confusing, right?

Hila Qu · 1:11:30
#data#instrumentation#analytics#audit
Takeaway1:22:30

Building the First PLG Team: The Analyst Is the First Hire

Hila's advice for the initial PLG team: lead with a head of growth or growth PM (a PM with much stronger analytics and experimentation skills), then add a data analyst — which she'd argue should be the very first hire, even before the growth PM, because without insights your experimentation is directionless. Round it out with a dedicated engineer and a designer. Beyond the team, she urges thinking about a PLG org — cross-functional counterparts in product, marketing, and sales.

  • Lead with a head of growth / growth PM — strong in analytics, experimentation, metrics
  • A data analyst should be the very first hire, even before the growth PM
  • Without insights, experimentation effort is directionless
  • Add a dedicated engineer and a designer to form the core growth squad
  • Think PLG org, not just team: head of growth product, growth marketing, and product-led sales counterparts

a data analyst needs to be the very first hire.

Hila Qu · 1:23:00

I think you should should not only think about a PLG team, you should think about a PLG org.

Hila Qu · 1:20:00
#hiring#team#growth-pm#org-design