How Startups Amplify Growth

This is a revised version of a Matillion talk that four Matillioners presented at our February 2024 annual Super All Hands meeting in Manchester.

"If you want to understand startups, understand growth. Growth drives everything in this world.” -Paul Graham Y Combinator, 2012

Growth comes in all formats 

Top-down, bottoms-up, and everything in between.

But, our impulse is to organize growth top-down, and our children are the best example. 

Below are images that demonstrate this top-down “education.” Teachers taught us through typing activities and typing games.

In 1999, this was flipped on its head in a remote village in India. 

A “hole in the wall” as Sugata Mitra called it, sat a computer with an internet connection.

The children in the village where this computer was placed had never even seen a computer before let alone the internet. They didn’t even know English and most were illiterate. 

As Sugata observed from his lab, he noticed the mouse began to move. Then, a document showed up with a child’s name on it (which was interesting because they didn’t even have a keyboard at first).

And, before long the children were sending emails, recording their own music and playing it back to each other, and learning everything from DNA replication to the English language—with no adult supervision.

These experiments were replicated in many 1st and 3rd world countries. It became known as “Minimally Invasive Education”

The final insight is in the name – minimally. It was discovered that if you complemented top-down growth with bottoms-up, children learned even better. 

They added a “grandmother” to stand behind the children to say encouraging and motivating things. (e.g. great job!, keep it up!, I wonder what that does?)

The addition of a grandmother improved children’s learning scores by 65%.

How does this relate to startups? 

We found that this same approach – a bottoms-up + top-down model – was common amongst high-growth startups. 

It wasn’t the only commonality. 

We identified four amplifiers of growth and one catalyst:

1. Treat product-led growth as an "industrial revolution"
2. Sales is PLG, too 
3. Start small, test & learn
4. Learn from the best

• Hire a Susana (aka hire a data engineer)

1. Treat product-led growth as an "industrial revolution"

(Inspired by Dave Boyce.)

When Sugata Mitra stared out the window of his air-conditioned office in New Delhi, he wondered, “How will poor children ever receive a good education?”

His worldview – teachers in rooms full of computers would not scale.

He needed a revolution…

This is reminiscent of a story long ago when we used to hire humans to do all the work. For example:

• Drain swamps, build factories, plow, plant, harvest
• Craft, assemble, cut, sew, bolt

Eventually, an efficient business owner thought …

It’s a similar story for modern startups.

Startups used to hire humans to do all the work. For example:

1. Make phone calls to generate awareness
2. Meet with prospective customers to ask them about their needs
3. Meet with prospective customers to demo our product’s ability to meet those needs
4. Meet with prospective customers to help them set up our product
5. Meet with prospective customers to help them experience value with our product

Eventually, an efficient business owner thought … 

• “What if we could hire machines (products) to do those same jobs?”

Thus, modern SaaS was invented and the birth of “product-led growth.” 

(This wouldn’t be a 2024 article if it didn’t have Gen AI)

The image above oversimplifies PLG, of course.

PLG is a strategy that requires company-wide focus and relentless execution.

The biggest challenge with executing PLG is – figuring out how we humans fit in.

2. Sales is PLG, too 

When salespeople used to get involved, their job was to hand-hold a prospective customer to experience value.

With PLG, salespeople’s jobs shift from hand-holding to enabling. 

PLG principle: Enable time to value, on the prospect’s terms

This principle emphasizes that people – children, adults, elders, and even prospective customers – learn best by a combination of intrinsic and extrinsic motivations.

To enable a prospect, a salesperson goes from asking, "How can I persuade this prospect to use our product?" (extrinsic motivator)

To, "How can I enable this prospect to discover how our product meets their needs?" (intrinsic + extrinsic motivators)

Referring back to our hole-in-the-wall analogy – the salesperson needs to emulate the grandmother, not the teacher in the computer lab.

The challenge with this transition – from hand-holding to enabling – is that the product has to take on a lot of the work the salespeople used to do. It has to:

1. Generate awareness
2. Ask prospective customers about their needs
3. Demo the product’s ability to meet their needs
4. Help them with set-up
5. Help them experience value

And, the more complex the software is (e.g. B2B), the harder this is to achieve. 

This is precisely the point where everyone gets confused about product-led growth (PLG).

They assume that PLG isn’t PLG until the product replaces salespeople in every step of the transition (#1-5 above). 

This is a myth. Salespeople never get replaced. If done right, they will just become more efficient. 

Therefore, it’s better to think of PLG as a continuum, not as a TRUE or FALSE. 

The big question then is – How do we speed up this transition from hand-holding to enabling so that the product picks up work the salespeople used to do?

There’s two primary ways:

1. Data
2. Self-serve UX

We’ll return to self-serve UX later but one of the best parts about modern SaaS is that it comes with large amounts of data.

This is also a double-edged sword because this data needs to be accessed by the entire business and it needs to be actionable. 

Accessing application data (user behavior data on your product), making it consumable, and getting it into an actionable format is a major challenge, especially for LEAN startups. 

You then have to join your application data with the rest of your data sources (e.g. CRM, analytics, billing/account data, etc) otherwise it has no context. 

According to 41% of companies recently surveyed, this is between 51-100 data sources. 

That’s a lot of work, especially for a startup who’s still trying to achieve product/market fit.

We need a catalyst…

Hire a Susana 

Right after we launched our PLG strategy in 2022, we hired Julian Wiffen Director of Data Science who sits within the product team.

Months later, Julian hired Susana Cardoso, our first data engineer. 

Susana became the catalyst for helping us get our data into the hands of the entire business and making it actionable. 

We set up a Snowflake data warehouse and began joining all our data – front-end Heap analytics, back-end SaaS application data, account, billing, Salesforce (CRM), etc. 

Now, the data was accessible but it still needed to become consumable and actionable. It also needed to be embedded into our salespeople’s systems and workflows. 

This required a different skillset, infrastructure, and data model. 

So, we leveraged GrowthScience, Matillion’s Data Productivity Cloud, and Pocus.

• GrowthScience is a combined team of Growth & Data Science. 
• Pocus is a Product-Led Sales (PLS) platform that helps business users get data into their salespeople’s systems and workflows.
• Data Productivity Cloud is the productivity platform for data teams. It helps you move data easily, no matter where it lives, with or without coding.

To talk about this, that’s the easy part. The hard part is execution.  

3. Start small, test & learn 

There are no templates for executing PLG. At Matillion, we started small way back in 2021 when CPO Ciaran Dynes wrote a 6-pager proposing PLG and a Growth team.

The first PLG A/B test we ran was in 2022 and it was a simple one – change the website’s primary CTA from ‘Get a demo’ to ‘Get started’ and see what happens. 

It was successful (also controversial), but after months of testing and iterating, we saw an increase in monthly sign-ups by 108%. But, our down-funnel metrics (e.g. activation rate & sales qualified leads) didn’t increase as much.

We learned quickly that a stronger self-serve UX and becoming more data-driven was needed to support PLG.

But this didn’t happen overnight. In fact, it’s still a work in progress.

Achieving a good self-serve UX is a long journey, especially for B2B companies and those with more technical software.

This is because your product now has to take on a lot of the work the salesperson used to do. For example, help:

1. Set up your product and
2. Experience value with your product

We accomplished #1 by launching the Data Productivity Cloud, which enabled new prospects to set up in less than 10 minutes. 

For #2, we are still on the path, as are most B2B software companies. 

Enough about Matillion. Let’s learn from others who’ve already been successful.

4. Learn from the best

In our research, we identified three trends among successful startups in the last decade: 

These were:

• A product-led + sales-assisted motion
• Great distribution
• And solving hard problems

“Most successful B2B SaaS companies eventually adopt a sales-assisted onboarding strategy” -Product-Led Onboarding, by Ramli John

Slack is the quintessential example of product-led growth.They had 3M daily active users before their first sales hire.

Their CEO famously said, “I think we can get away without having a sales team in any kind of traditional way probably forever.”

But, by 2019, 40% of Slack’s revenue came from sales.

Come to find out, Slack ended up hiring a lot of salespeople to amplify their growth. 

Especially with larger deals (e.g., Enterprise), Slack began to leverage a product-led, sales-assisted motion.

But, it was too little too late because something else happened in 2016. 

Enter Microsoft Teams. 

Within three years, Teams surpassed Slack. By 2022, they had over 270M daily users (almost 10X Slack).

Slack had a superior product. 

But, Microsoft had superior distribution. They had:

• A better product-led, sales-assisted motion:
  • E.g. free product with a large sales fleet and enterprise relationships
• A better focus on the buyer’s needs, as well as the user’s
  • For example, bundling value with teams and 365 suite
• Sharing SaaS revenue with partners

“First-time founders are obsessed with product. Second-time founders are obsessed with distribution.” -Justin Kan, Founder of Twitch

Great distribution was the story with Microsoft Teams but it’s a similar story for a lot of companies.

One of those companies is Snyk, a B2B developer-first cybersecurity company valued at $7B.

Their core innovation was making security software product-led for developers. They quickly grew, just like Slack, but stalled at the monetization phase.

They learned that they were meeting the needs of their users (developers) well but struggled with the needs of their buyers (security).

For example, a Chief Information Security Officer desired a wide array of coding languages while a developer only cared about the code they develop in (e.g. Java or React).

So, just like Microsoft Teams, they shifted their GTM strategy to a strong product-led, sales-assisted motion which meant PLG for developers while selling to security leaders.

By 2018, they had nailed this distribution strategy and begun to double down.

“Developers are constantly online looking for solutions to their problems. We use content and strong SEO to insert Snyk into the developer’s workflow as they search for answers.” -GTM leader at Snyk

Over time Snyk experimented with a myriad of distribution strategies:

• The Snyk Vulnerability Database
• Snyk Advisor (holistic health assessment of open source packages)
• Snyk Website Scanner
• Snyk Code Checker (just copy and paste your code in)

Finally, the last trend amongst successful startups is solving hard problems.

Recently, at MindTheProduct London, CPO of Wise Nilan Peiris showcased many hard problems which took 2-5 years to solve.

• “Building a high growth sustainable business is slow hard work solving really hard product problems. Solving these takes time but in solving them you create huge amount of value for customers, which translates back to growth.” -Nilan Peiris

One such problem was that banks had been charging 3-7% exchange rates for 30+ years and beneath those exchange rates were hidden fees. 

Wise wanted to charge much lower rates, but they needed to spread awareness about these hidden fees.

So, they tried to change the law to make it illegal for banks to hide fees in the exchange rate.

They leveraged their customers by running petitions and even had a piece of legislation make its way through Parliament. 

It didn’t pass in England but did in Europe.

It was the biggest marketing campaign ever. 

Wise has a 0.35% exchange rate that a lot more people now know about.

Nilan goes on to explain many multi-year problem-solving missions, for example:

• Transferring money used to take 3-5 business days; Wise made it instant.
• Sending money out of Singapore required Wise to meet physically with every customer, so they had their customers lobby local legislators.

CPO Oji Udezue recommends two questions when trying to uncover hard problems:

• What is still hard for the customer? What saps their time, productivity, and energy?
• What will make the customer feel 3-5x more productive?

Do you have a hard problem you are solving this year?

What about the next 3-5 years?

Whatever it is, we hope you make it the next hole in the wall.

But, if it’s not, and you just want to execute a PLG strategy, we hope you sign up here for the Data Productivity Cloud, harness your inner Susana, and begin amplifying growth.