Waves, dams, and rivers

How demand works at scale

Fun experiment - I recorded a video going deep into this newsletter’s topic! Might keep doing this - let me know what you think, what to change, etc.

---

Think of demand as the “project” on someone’s to-do list. Demand only exists at the level of the individual. A company or market or persona doesn’t have demand; a human being does.

Most of the time, startups fail to find product-market fit because they have a grand theory of a “market problem” (like, “Sustainability is a 7 trillion dollar problem!”)… but don’t know why one real human gives a sh*t and will pay them money.

That said, when you first understand demand at the individual level, you can then zoom out and see how demand aggregates, and why some startups grow super fast and others don’t.

I’ve found three zoomed-out archetypes: Waves, dams, and rivers.

Knowing these archetypes can help you figure out how to grow fast: Who to target, what to sell, and what to build.

Waves

Generative AI is a new thing. People are trying to use generative AI in their work. As they use generative AI in their work, they run into weird new things like hallucinations, scaling issues, and MCP servers.

This is a wave breaking across the economy; the wave is made up of new projects on people’s to-do lists.

This wave might break unevenly, as some people adopt AI faster than others. Or it might be a tidal wave that breaks all at once - imagine COVID, or a new regulation that goes into effect at one time - and every business has a new project on their to-do list.

Here are the physics of demand waves:

• Demand waves usually break unevenly. To catch the wave, you often need to be tinkering around at the frontier with other people who are “early” to the wave.

• At the frontier, you’re specifically looking for projects that will be forced onto more and more people’s to-do lists as the wave crests.

The nature of waves is that it is very hard to tell how big a wave will be, or when it will get big and crash. There are a lot of ripples that we misinterpret as waves (think: the metaverse).

A few examples of startups that are chasing or rode waves:

• Composo - Building an evals tool for companies that are trying to scale AI products and features.

• Engines.dev - Building a sandbox tool for teams using AI software engineers.

• Vanta - Rode the wave of “SOC 2 becoming a required thing by enterprise buyers”

• Tuple - Like zoom for pair programming; Rode the wave of “Slack acquired the main competitor in this space and shut it down”; then rode the COVID wave

Dams

Generative AI is (still) a new thing. People have been making websites for decades. Generative AI enables you to make a website in seconds from a simple text prompt, rather than spending hours using a website editor, wishing you were never born.

When Lovable and Bolt came out, there was a whooshing sound, and they scaled very fast. This is because they broke a dam. Here are the physics of dams:

1. There was an existing project on a BUNCH of people’s to-do lists (build a website / web app / prototype)

2. They were COPING with existing options (hence: stuck behind a dam)

3. Something - usually, but not always, a new technology - enabled a serious improvement vs. the existing options, that made a HUGE number of people switch, basically overnight.

A few examples of startups that burst dams:

• Jump - AI notetaker for financial advisors (who had been taking notes manually prior)

• Lovable/Bolt - see above

• The Free Press - Media company that started out of a frustration with legacy media companies like the New York Times

• (gasp) WeWork

Rivers

Customer testimonials are not a new thing. There are a bunch of existing testimonial software tools out there. You try them all, think they’re meh, and decide to build a new, fresh, testimonial software.

This is the case where there isn’t a whooshing sound; you’re offering a better tool for testimonials for people who are mostly satisfied with their options. Some people choose yours sometimes.

The physics of demand rivers:

1. It is about existing projects on people’s to-do lists (like dams)

2. But people are mostly “fine” with their existing options, and we don’t have any radical unlock versus alternatives

3. So we tend to have to build a product up to feature-parity with alternatives, AND carve off some niche, to grow at any reasonable pace in a river.

That said, there are different kinds of rivers. Demand rivers can be faster- or slower-moving.

• Fast-moving rivers are where a high volume of people have a particular project on their to-do list at any one point in time: There are a TON of new tech startups, and they all need to run payroll, and most need to get SOC 2 compliant.

• Slow-moving rivers (AKA: lazy rivers) are where a low volume of people have a particular project on their to-do list at any one point in time: Think of how frequently, for example, the HR team at a university needs to do, well, anything.

Demand rivers can lead to fast-growing startups. Raising VC dollars to scale a company in a lazy river is usually a bad idea. That said…

The demand profile changes over time.

Let’s look at SOC 2 compliance: Vanta caught the wave of enterprise buyers requiring basically every software vendor to be SOC 2 compliant. Then came a bunch of competitors, who did not catch the original wave like Vanta did, but still were all able to grow fast in a fast-moving river.

Let’s look at Lovable and Bolt: They burst the dam for “building a website / front end” and grew insanely fast. Now, as everyone niches off to tackle “prompt-to-thing”, it’s unclear whether they are bursting other dams, catching waves, or trying to find hypergrowth in now-slow-moving-rives.

See how the demand profile changes over time? You catch waves, burst dams, or show up a little late or a little wrong and pout in a lazy river. (Though even the lazy river is better than what happens to most startups, who don’t understand demand at all: They show up with their floaties in a desert.)

So what do I do with this very informative information, rob?

First: Assess what kind of demand situation you’re currently attacking. Is there a dam, a wave, or a lazy river?

Then: If you’re not targeting a dam or wave, try to find the the dam or the wave… or, I suppose, slightly-faster-moving-water. This could mean:

• Targeting a different end-customer

• Building a very specific feature set that enables you to burst a dam or catch a wave (yes, probably AI)

• Hunkering down and keeping burn low until the water starts rising

Examples:

• A friend’s startup was trying to sell to enterprise HR buyers - a lazy river. He switched to target outsourced HR providers, and burst a dam there, basically 5x’ing revenue in eight weeks.

• Jump started trying to sell an AI notetaker to salespeople - a fast-moving river, but with very good existing options on the market. They switched to target financial advisors, and burst a dam there, growing insanely fast.

• I saw a new wave coming with my last startup: We were selling a hiring tool, and I kept hearing restaurant owners complain, “I don’t want to spend more money on hiring people, I want to spend money retaining the people I already have.” We launched a new employee retention product that was the main driver of our $1-$4M growth.

• Over the past few months, I have been trying to find the “dam” for Waffle. I couldn’t explain what I was looking for until now, but that’s what I’ve been trying to do, and perhaps am nearing it. More to come.

It is very, very hard to build a company in a lazy river, or even in a moderately paced river. You wind up needing to build more product, sell harder, and get more things right. It is, perhaps paradoxically, easier when you find a dam or wave. I’ve heard many stories from the companies that caught waves, where employees aren’t working as hard as people grinding away trying to generate traction in lazy rivers.

---

PS: I’m working with my next small group of founders in ~late July, early August. Here are the details. Get my help reviewing your sales calls and finding pull/PMF, and get a year of weekly problem-solving sessions with me!