Product-Led Grind
The real journey to Product-Led Growth
Most of us learn about product-led growth through books, blogs, and people’s perspectives on the Internet.
We hear why it’s important and the tactics we should use, but the grind is rarely discussed.
That’s what this article is for.
By calling it “product-led grind” we can reset everyone’s expectations.
For starters, implementing PLG is an attempt to change company-wide behavior.
Think about how hard it is to change your behavior or the behavior of someone you love. Now, project this out to hundreds and thousands of people across a company.
It’s the toughest challenge in business. It’s a grind.
At Matillion, we’ve been implementing PLG for over two years now and anticipate at least two more.
Most PLG companies are in the same boat — a multi-year grind.
Resetting expectations
Surrounded by numbers, we expect numbers (the character in the middle is the number 13).
Surrounded by letters, we expect letters (the character in the middle is the letter B).
With product-led growth, we are surrounded by stories, tactics, and advice.
One story that clouds everyone’s expectations is Slack. They are the quintessential example of product-led growth.
They grew to 3M daily active users (DAUs) without a salesperson.
Their CEO even said,
“I think we can get away without having a sales team in any kind of traditional way probably forever.” -Stewart Butterfield, SXSW 2016
But, Slack also went through a grind.
During that same year, they hired their first salesperson.
And, by 2019 40% of Slack’s revenue came from sales.
Something else happened in 2016 — Microsoft Teams.
Within three years of Team’s arrival, they surpassed Slack’s DAUs.
By 2022, they almost 10X’d Slack (image below).
Slack is still a great story and a very successful product, but they weren’t immune from the grind.
The grind is not dissimilar to the “initiation” phase of the hero’s journey where we experience many trials and tribulations.
“But, what about xyz company? It wasn’t a grind for them.”
Even the top PLG companies faced a grind.
For Rows.com, it took 4 years to build their self-serve motion.
It took GitLab ~7 years to fully realize the success of PLG.
For MongoDB, it took 5 years, and it wasn’t always smooth.
For Snyk, it was 3 years.
Many other great companies fall into this category — HotJar, ClickUp, and Amplitude.
There are even companies in the PLG industry, like Pocus and Navattic, who have yet to launch a full self-serve experience.
(Since writing this post, Navattic launched a self-serve motion)
In the book ‘Product-Led Growth’, author Wes Bush actually advises startups to begin with a sales-led motion because it forces you to talk to your customers and learn faster.
“It often makes sense to start with a sales-led approach to better understand the customer’s pain points, objections, and core problems implementing your solution. If you jump too quickly to a product-led model with a new category, you risk a high churn rate because you simply don’t understand what it takes for customers to succeed.”
In a self-serve motion, prospects don’t like to talk to you.
Think about the last time you went through a self-serve experience, you didn’t want to talk to anyone either.
Imagine going through the self-checkout lane at the grocery and getting interrupted while checking out.
Then, once again on your way to your car.
Then, when you get home, you receive a phone call and an email.
You might go to a different store next time.
Some companies, like Snyk, never took Wes Bush’s advice. They started with a product-led motion.
But, they still faced a grind. Their challenge came when they started selling upmarket to buyers and big companies.
Then, they rushed to hire salespeople and pivoted their go-to-market strategy just like Slack.
Snyk just got lucky that Microsoft Teams wasn’t lurking around the corner.
Teams excelled at moving upmarket because they already had a large sales fleet with built-in Enterprise relationships.
In the end, Slack had a superior product. Teams had superior distribution.
“First time founders are obsessed with product. Second time founders are obsessed with distribution.” -Justin Kan, Founder of Twitch
How do I do this?
Clay Christensen, the famous business writer, once met Andy Grove, the CEO of Intel:
“As Clay Christensen recalls, Grove asked him after a meeting,
“How do I do this?”
Christensen responded with a discussion of business strategy, explaining how Grove could set up a new business unit and so on.
Grove cut him off with a gruff reply: “You are such a naïve academic. I asked you how to do it, and you told me what I should do. I know what I need to do. I just don’t know how to do it.”” (source)
Like Grove, we already know what to do (PLG) and why it is important. If you don’t, there’s plenty of content out there.
The challenge is how.
First, we need buy-in from leadership. This includes investment in resources but most importantly strategic goals.
If you don’t have this investment you must lead your company there.
How do you lead your company there? Start small, prove value, and align the rest of the organization.
Essentially, be an intrapreneur.
At Matillion, we had buy-in from product leadership but needed to align the rest of the organization so that PLG could become a strategic goal.
The first experiment we ran was to change the website’s primary CTA from ‘Get a demo’ to ‘Get started.’
It took another year and a half until PLG became a strategic goal, but that’s when things took off.
The side effects of strategic goals are funding, resourcing, and cross-functional alignment. These are what you need to build an optimal self-serve experience.
If you are a new startup or don’t have a great self-serve experience you may have to start smaller.
This is what Rows.com and Superhuman did early on.
During Row’s first three years of existence (2018–2021), they had a waitlist. For every sign-up, they did individual qualification and onboarding calls.
Superhuman also did individual onboarding calls.
In retrospect, it was ingenious for growth. It solved the problem I mentioned about self-serve prospects being very hard to get ahold of.
By onboarding every prospect, it created rapid learning, word-of-mouth virality, and created momentum-building first impressions that turned prospects into retained customers.
The tradeoff, of course, is higher customer acquisition costs (CAC).
Thus, building an optimal self-serve motion is required when you start to scale. But, this takes time.
As mentioned before, it took 4+ years for Superhuman to build a self-serve motion.
It was the same for Rows.com:
“A spreadsheet is a complex product and building all the necessary features to provide a compelling experience on a self-served model took time (~4 years).” -Henrique Cruz, Head of Growth
A lot of us do not have the luxury to grow this slowly. We may be at later growth stages or being pulled towards quick ROI from investors.
“Growth takes time. But, don’t take your time.” -productology.substack.com
Building an optimal self-serve experience
Hacks and playbooks won’t work here. Your focus needs to be on systems.
“Growth isn’t something we do. It is a side effect of our systems.” -productology.substack.com
Here are some systems for building an optimal self-serve experience:
- Treat self-serve like a product
“When the executive team calls me and asks why they aren’t converting users into customers, I tell them to buy a plant. Seriously. If they don’t water it, it’s going to wither and die. If you water the plant and give it sunlight, it’ll grow. Everyone knows how the system works. Yet, even though we know what to do, millions of plant still die.
Why? Nobody takes ownership.
The problem is that most businesses don’t appoint a person or team to take ownership, nor do they give that person the resources or time it takes to thrive. “ (Wes Bush, Product-Led Growth)
You have to treat your self-serve experience like a product. This means you need the commitment of a full product team — engineers, a product owner, a product manager, UX designer.
Your team also needs to be empowered to experiment across many different code surfaces. This means you’ll naturally rub shoulders with most product teams.
I have been the inaugural Growth team member at every startup I’ve joined and had to beg and borrow resources and backlog space from product and marketing.
This isn’t a recipe for success, but it’s the natural progression of most organizations.
“It’s hard to develop new things in big organizations, and it’s even harder to do it by yourself. Bureaucratic hierarchies move slowly, and entrenched interests shy away from risk.” -Zero to One
The biggest difference between a self-serve product team and other teams is that — your primary user is the prospect, not the customer.
This means it’s likely that the product team has rarely done any product discovery with prospects.
Prospects come with radically different contexts and needs.
Customers love using features like Git or automation, but someone who is using your tool for the first time could care less about those.
Prospects are usually evaluating more than just your tool and have a list of evaluation criteria.
In a self-serve motion, you don’t have humans (e.g. sales) intervening to ensure your tool is differentiated and meets their evaluation criteria.
So, without human intervention, it’s critical to understand the psychology of the human.
2. The psychology of self-serve
Instead of explaining the complexity of psychology in writing, I created a comic strip inspired by the guys at Growth.Design using Matillion as an example.
3. Increase your rate of learning
When we talk about the rate of learning, we immediately think of experimentation (aka quantitative learning).
This is a crucial aspect of learning but it also needs to be paired with qualitative learning (aka talking to humans).
Unless you are onboarding every prospect like Rows or Superhuman, you don’t have a system in place for continuously talking to humans.
This means your rate of learning is much slower in a self-serve motion and that’s a problem.
“The only way to win is to learn faster than anyone else.” -Eric Ries
You have to find ways to talk to your prospects weekly. At a minimum, monthly.
If you browse the web you can find many tactics and I’ve tried them all — $100 survey incentives, in-product feedback, offers of 1on1 help, personal emails from the CEO, and more — with very limited success.
The best method, by far, has been in-product chat.
Some in-product surveys work but they need to require only one click. Doing it this way defeats the whole purpose and turns it into quantitative learning.
The other method that’s been successful is to source high-quality candidates from Respondent.io who match your ideal customer profile. This isn’t 100% ideal because context is king and they are not currently in-context evaluating your tool.
What else could you do?
Let’s think outside the box for a second, what if we individually onboard some prospects?
For example, new personas who are less likely to be successful via self-serve.
Or, if we normally target engineers but an analyst or IT signs up, 1:1 onboard them.
If this idea wouldn’t fly at your company, then propose 1:1 onboarding for 15% of new prospects.
Or, just 1:1 onboard your buyers and not your users.
If none of this would fly at your company, then mimic Superhuman and create a canned self-serve experience.
They force their prospects into a 3-minute sandbox experience to learn how to use their product.
This reduces complexity and helps shift the prospect’s mental model away from how they used to do things to how your product works.
This forces a good first impression and first impressions are paramount.
4. Have an overarching theory
I have used the scientific method my entire career but I forgot a critical component of it — theory.
Most of us primarily use ‘hypotheses’ to drive our thinking and experimentation. This is great but a hypothesis is insufficient without an overarching theory.
- A hypothesis is a proposed explanation for something that can actually be tested
- A scientific theory is an explanation that has been substantiated through repeated experiments or testing
Thus, a theory is an in-depth explanation of a wide range of customer behavior.
A hypothesis would be a proposed explanation of a more specific piece of customer behavior within the confines of that theory.
The importance of creating a theory is that it creates a shared understanding of the perceived facts.
Hypotheses, then, are how we learn and update those perceived facts.
Here’s how this terminology changes behavior for your teams. A theory creates objectivity, and inevitably when someone disagrees with your hypothesis or your design decisions, you can point back to the theory.
If they continue to disagree with the theory, you can tell them to propose a new one. You can say that you’d be happy to work from a better version if they propose one (This is where their disagreement usually ends).
What does a theory look like in this context? Here is an example of one we used at Matillion for the prospect’s self-serve journey:
The fact of the matter is — it is incomprehensibly difficult to build and sustain a great business. Even our most famous success stories were on the brink of ‘failure’:
- A year after founding Google, they were willing to sell their company for less than $1 million, but the buyer said the price was too high. -TechCrunch
- Both Netflix co-founders said they offered the company to Blockbuster for $50 million. -Newsweek
So, next time you are in the midst of a Product-Led Growth discussion, reset everyone’s expectations and start calling it Product-Led Grind instead.