Twitter acquired Threader! Learn more

Julian Shapiro
+ Your AuthorsArchive @Julian I deconstruct how things work—like storytelling and critical thinking—and share learnings along the way. Writing: Work: Oct. 05, 2021 5 min read

1/ An intro to startup growth.

Most fast-growing startups have something in common:

Product-led growth (PLG)

I define it as when existing users drive your growth.*

PLG only works for these startups:

Example of PLG: When you join Slack, you invite teammates and partner organizations.

Why is PLG critical? It scales, it often has little to no marginal cost, it's defensible, it has network effects, and it can build moats.

It's the holy grail of growth.

(*Others may define PLG as freemium replacing sales.)

I see many PLG subcategories:

Category #1: Invitations

This is where the use of your product naturally encourages users to invite friends/colleagues because they both get more value from the product. Hence, 1+1=3.


• When you send $100 via Paypal, the recipient signs up to receive the money.

• When you talk to others on Zoom, they have to install Zoom first.

• To chat with teammates, you onboard them into Slack. When you want to talk to outside teams, you use Slack Connect.

Notice two important elements:

A) The person inviting and the person being invited both significantly benefit from the other’s use of the product.

B) The benefit to both parties has to be getting an important job done, which often means getting paid or enabling communication.

Category #2: Billboarding

Billboarding is my term for when the use of your product creates a public advertisement.

For example:

• Wearing Airpods in public. People recognize they're by Apple.

• Sending people your Calendly/Dropbox link that shows the URL/page.

• Having “Sent from my iPhone” in your email signature.

• Replacing your Twitter avatar with an NFT photo.

If you can get PLG such as this to work, it's often healthier than ads:

Ads face volatility, heavy competition, audience saturation, and are often unaffordable.

Further, compare PLG to SEO:

With SEO, you’re at the mercy of Google’s algorithm updates, and often extreme competition, plus it's not guaranteed to scale. And the feedback loop can be long.

So, for your own PLG, ask yourself:

• BIllboarding: Is it worth modifying my product to make its use visible to non-users?

• Invitations: How can I modify my roadmap to prioritize PLG features? What can I change to encourage users to invite new ones in such a way that both get a lot of value out of the app?

Note that cash rewards are not real PLG. Those are referral programs, which are worse than PLG.

They're generally not one-to-many to the same extent as PLG, they can easily attract the wrong persona, and most users don't engage with them. Certainly businesses don't.

I believe you should focus on enduring product-level hooks instead of superficial reward-level hooks.

You might be wondering: Where does word of mouth fit in?

WOM is amazing. Perhaps we can categorize it as part of PLG, in fact! Although it often has slower virality.

It's hard to slap on WOM. It naturally emerges from:

• The product experience (entertainment or utilitarian) was delightful and the customer wanted to share that joy with others.

That's partly how Disney+/Netflix grow:

People watch shows, love them, and tell others they have to watch them too.

The second way I think about word of mouth is "tribal affinity."

That's where you create a movement of prestige, exclusivity, and/or ideology, e.g. a chat app, clothing brand, video game league.

Once a member is initiated, they want to look good by helping others join.

Triggering delight and tribal affinity both emerge from product-level and brand-level decisions.

PLG is not a band-aid you slap on.

See how growth is so deeply tied to product? Don't put growth in a silo.

Moving on, another channel I find under-discussed:

Programmatic SEO.

It's not a fit for every startup (most channels aren't!) and it may not be easy, but it's often overlooked:

Examples of programmatic: Quora, Stack Overflow, Airbnb, Yelp, Wikipedia, IMDb, and others that produce a huge long-tail of pages that help searchers with niche interests.

By having thousands of long-tail search pages that dominate the SEO margins, you can aggregate traffic into a meaningful volume.

For your startup, ask yourself:

Is there content relevant to my industry that I can write about or generate at scale that will attract visitors who are likely to convert into customers?

If so, perhaps work with writers and landing page experts to run a proper test.

Before I wrap up, I'll point out that I'm focusing on a couple channels, but what works for your startup may be entirely different!

Maybe it's sales, community-led growth—lots of options.

And that brings us to Part 3:

When do we test all these channels?

Once retention and NPS are good, I like to start working on every growth channel that I have high conviction will work.

Assuming, that is, I have the bandwidth to test each properly.

I'll expand my team with outside contractors as needed if I'm short-handed.

It’s important for you to quickly identify which channels will work (it can take a while to find out) and use that feedback loop to iterate aspects of your product.

You could modify your roadmap to lean into growth channels that work—if doing so is ALSO good for your product.

Maybe it's a new PLG feature, or a feature to encourage double the user-generated content.

As for the channels I'm uncertain of, I defer them until I've tested the ones I'm confident in.

To wrap up, when growing a startup, I prefer a sequence like this:

1. Fix every major drop-off point in your funnel. Plug holes.

2. Next, maximize retention and NPS within reason. Make sure people stay.

3. Next, scale up customer acquisition. Start with a few channels you have high conviction in.

Throughout this experience, your north star is to maintain a never-ending cadence of running new growth experiments—each with the potential to produce significant results.

You can follow @Julian.


Tip: mention @threader on a Twitter thread with the keyword “compile” to get a link to it.

Follow Threader