The Un-SaaS Ch. 2 — The Toll Position

Chapter 2 of 26

This is a chapter from The Un-SaaS: A Toll Stack Engineer's Handbook. Each Friday, we're publishing a chapter as a bonus for our readers.

In 1757, a man named William Edwards tried to build a bridge across the Taff River in Pontypridd, Wales.

The first one collapsed. So did the second. The third attempt — a single-arch stone span, the longest in Britain at the time — stood. It still stands today, 267 years later. Edwards didn’t own the river. He didn’t own the land on either side. He didn’t generate the traffic. He just built the thing that let people cross.

For the rest of his life, and for generations of his family after him, anybody who wanted to cross the Taff at that point paid a toll.

Not because Edwards was clever about marketing. Not because he had a brand. Not because he posted daily about bridges on the 18th-century equivalent of LinkedIn. He just built infrastructure in the right place — where existing demand had no existing path — and collected a small fee every time someone used it.

The demand was already there. The river was already there. The people who needed to cross were already standing on the bank. Edwards didn’t create any of that. He built the bridge.

That’s the oldest business model in human history, and it’s the one this book is about.

An 18th-century stone bridge over a river with a toll booth at the center. An operator collects coins as travelers cross.

The Digital Version Is Sitting in Plain Sight

Let me show you something you can verify in the next sixty seconds.

Open YouTube. Find any creator with more than 100,000 subscribers who sells a course or digital product. It can be in any niche — fitness, investing, productivity, cooking, whatever. Now look at the video description.

You’ll find a link. Maybe it says “Get my course” or “Download my free guide” or “Check out my favorite tools.” Click it.

Where does it go?

In roughly 80% of cases, the link goes straight to a sales page or a checkout. No email capture. No landing page. No sequence. No follow-up. One shot: the person either buys on the spot or disappears forever.

Now do the math.

A creator with 200,000 subscribers puts out a video that gets 15,000 views. Maybe 1% of viewers — 150 people — click the link in the description. Of those 150, maybe 2% buy the $197 course. That’s 3 sales. $591.

What happened to the other 147 people who clicked? They looked at the checkout page, weren’t ready, and left. Gone. No email. No name. No way to follow up. No way to nurture. No second chance.

Those 147 people are dead clicks.

They expressed intent — they literally clicked a link about the product — and then evaporated because there was nothing between the click and the checkout. No bridge. Just a river with people standing on the bank, looking at the other side, and walking away.

Dead clicks are the single most under-monetized asset on the internet, and they’re hiding in plain sight in every creator’s YouTube description, every podcast’s show notes, every newsletter’s footer, and every Instagram bio.

What an Operator Installs

Here’s what happens when a Toll Stack Engineer shows up.

You build the bridge. Specifically, you insert a small piece of infrastructure between the creator’s traffic source and their checkout — a landing page that captures an email address before forwarding the visitor to the product.

That’s it. One page. One form. One redirect.

The creator changes nothing about their content, their brand, their posting schedule, or their product. They change one link. Instead of pointing at their checkout, their description link now points at your landing page. Everything upstream of the link stays the same. Everything downstream of the link now flows through your infrastructure.

The visitor arrives at your landing page. They see a headline that matches the creator’s promise (not your promise — their promise, just articulated better than a raw checkout page does). They enter their email. They get a short, well-crafted sequence — three to seven emails over five to fourteen days — that does what the creator’s raw checkout page never could. It educates, builds desire, addresses objections, and presents the offer at the moment the reader is most likely to say yes.

Some buy during the sequence. Many more buy over the next thirty to ninety days as the list compounds and periodic emails surface relevant offers. And every one of them — buyer or not — is now a known contact in a database you own and control.

The bridge captures what the river was drowning.

Three Things the Operator Owns

This is important, so I want to be explicit about what’s yours and what isn’t.

You do NOT own: the creator’s audience, the creator’s content, the creator’s product, the creator’s brand, or the creator’s traffic source. Those belong to the partner. You don’t want them. They are their competitive advantage, not yours.

You DO own three things:

1. The infrastructure. The landing page. The email sequences. The analytics. The redirect logic. The tag taxonomy. The A/B testing framework. You built it. It runs on your hosting. It’s wired to your accounts. If the partnership dissolves, you turn it off. The creator doesn’t have the keys to your infrastructure any more than you have the keys to their YouTube channel.

This is the control principle: your infrastructure, your kill switch. It’s the difference between being a vendor (who can be replaced by another vendor) and being an operator (who has installed something indispensable).

2. The data. Every subscriber who enters your landing page generates data — what they clicked, what they opened, what they bought, when they bought, what they didn’t buy. It also captures which sequence step converted them and which subject line they responded to. That data lives in your systems. Over twelve months of operation, the experiment log — the accumulated intelligence from hundreds of micro-tests — becomes the single most valuable thing you own.

The landing page can be rebuilt. The email sequence can be rewritten. But the experiment log that tells you which version of the landing page converts 31% better, and why, and for which audience segment — that’s irreplaceable. It took twelve months of continuous testing to produce. A competitor copying your current page is copying a snapshot. You own the trajectory.

3. The relationships. Not the creator-audience relationship — that’s theirs. But the operator-subscriber relationship you’ve built through email. The people on your list know you as the person who sends them useful, well-curated content. Some of them will follow you across partners. Some of them will become subscribers to a second toll position you’re running in an adjacent niche. Over time, a segment of your aggregate list becomes your audience — people who trust your curation regardless of which partner originated the traffic.

That cross-network intelligence — the ability to see patterns across multiple partners, multiple niches, and multiple audiences — is the operator’s version of proprietary data. No individual partner has it. No competitor who runs one position has it. It only exists in the operator who runs several, and it gets more valuable with every position added.

What an Operator Is and Is Not

A Toll Stack Engineer is not an affiliate marketer. An affiliate drops a link and hopes. An operator builds infrastructure, captures data, runs experiments, and compounds.

A Toll Stack Engineer is not a freelancer. A freelancer bills by the hour or the project and moves on. An operator installs something permanent and collects a share of the lift for as long as it operates.

A Toll Stack Engineer is not an agency. An agency hires people to do work for clients. An operator works alone — with AI agents — and owns the infrastructure rather than selling labor against it.

A Toll Stack Engineer is not a SaaS founder. A SaaS founder builds a product, finds an audience, and sells to it. An operator skips the first two steps entirely — the product exists, the audience exists — and installs the conversion infrastructure that connects them.

Here’s what a Toll Stack Engineer is.

An operator who identifies existing demand flowing through someone else’s business and builds a small piece of infrastructure that captures and converts that demand more effectively. They negotiate a revenue share on the lift, and then optimize that infrastructure continuously using data and AI agents.

The bridge builder. The toll collector. The person who doesn’t need to own the river or the land — just the structure that lets people cross.

Why the Partner Says Yes

This is usually the first objection from engineers: “Why would anyone let me do this?”

It’s a fair question if you’re thinking about it from a labor frame — why would someone hire me to do this thing? But the toll position isn’t a labor sale. It’s a low-risk bet.

Consider it from the partner’s perspective.

A creator with 200,000 YouTube subscribers is leaving dead clicks on the table every single day. They probably know it. They probably feel bad about it. But they don’t have the technical skills to build a proper landing page, wire an email sequence, set up an A/B testing framework, and run continuous optimization experiments. And even if they could, they don’t want to. That’s not why they got into content creation.

Now an operator shows up and says: “Let me run a two-week test. I’ll build the infrastructure on my own dime. You change one link. If my version loses, you revert with one click and you’ve lost nothing. If my version wins, I keep running it and we split the lift.”

What’s the downside? Bounded and reversible. The partner risks very little. They change nothing permanent. If it works, they make more money. If it doesn’t, they go back to what they had. No retainer. No invoice. Minimal risk.

The partner says yes because the deal structure makes “no” irrational.

And once the test wins — once Variant B is converting 25% or 31% or 40% better than the raw checkout link — the partner has little reason to revert. Not out of contractual obligation. Out of self-interest. Voluntarily giving up a 31% conversion lift would be economically insane, and the partner knows it.

You are not selling a service. You are proposing a bet that is overwhelmingly likely to benefit the partner and designed to limit downside. That’s why the conversion rate on partner outreach is so much higher than engineers expect.

The Mental Shift

Most engineers hear “build something inside someone else’s business” and think dependency. They think about platform risk — building on top of Twitter’s API, or Shopify’s ecosystem, or AWS’s pricing whims. They’ve been burned before. The instinct to own the whole stack runs deep.

But the toll position inverts the dependency. The partner is more dependent on you than you are on them — because you own the infrastructure, the data, and the experiment log. If the partnership ends, you lose one revenue stream. The partner loses their entire conversion lift and has to rebuild from scratch. The asymmetry favors the operator.

And unlike platform dependency, you’re not building on top of a single point of failure. You’re building a portfolio — twelve positions across seven partners across four niches. The loss of any single partner is a haircut. The loss of any single niche is uncomfortable but survivable. The portfolio is designed, from day one, so that no single point of failure can take down the whole thing.

The real mental shift isn’t about tactics or infrastructure. It’s about seeing yourself differently. You’re not an employee. You’re not a freelancer. You’re not a founder.

You’re an operator. You install toll positions. You collect the toll. You add the next one to the stack.

And the dead clicks you just identified in that YouTube description? They’re standing on the bank of the river right now, waiting for someone to build the bridge.

The Practitioner’s Dead Clicks

A 1960s doctor's waiting room. A physician pins business cards to a corkboard while patients leave through a side exit.

Everything I just told you about dead clicks in YouTube descriptions? It’s worse offline.

At least the YouTube creator has a link. It goes somewhere. It can be tracked, even if nobody’s tracking it. The link exists in the digital world where clicks are countable, funnels are buildable, and infrastructure is installable.

Now consider a real estate agent.

She closes twenty deals a quarter. After every closing, the new homeowner asks the same question: “Do you know a good contractor?” She recommends the same three renovation firms every time. The referral mechanism is a verbal mention and maybe a business card. “You should call Rivera Renovations. They’re excellent.”

That’s it. No tracking. No follow-up. No confirmation that the homeowner actually called. No data on whether the referral converted. No revenue flowing back. Just a name spoken into the air and a business card that has a 50/50 chance of making it out of the parking lot. It ends up in a cup holder, never to be seen again.

She made twenty contractor referrals this quarter. Each one represents a homeowner who will spend $15,000 to $80,000 on renovations. That’s $300,000 to $1.6 million in downstream revenue — flowing through her client relationships, past her desk, out the door, and into someone else’s business with zero capture, zero tracking, and zero compensation.

Those are dead clicks. They just don’t look like clicks because there’s no mouse involved.

The pattern is everywhere once you see it. A wedding planner recommends photographers, florists, and caterers to every couple — thirty vendor referrals per wedding, forty weddings a year. An interior designer suggests the same furniture stores and tile contractors to every client. A business consultant mentions the same CRM tool in every engagement — “You really should try HubSpot” — twenty times a month.

Twenty verbal referrals a month. No affiliate link. No landing page. No tracking pixel. No commission. Just professional goodwill evaporating into the atmosphere.

Let’s do the math on that consultant. She recommends HubSpot’s paid tier to roughly twenty clients per month. HubSpot’s Starter tier is $20/month. Their Professional tier is $890/month. Let’s be conservative and say half of her referrals convert, and the average plan is $200/month. That’s ten conversions at $200/month — $2,000 in monthly recurring revenue she’s generating for HubSpot. At HubSpot’s 30% recurring affiliate commission, that’s $600/month flowing into nobody’s pocket. Seven thousand dollars a year, compounding as each new referral adds to the base. She has no idea this money exists.

The operator opportunity here is not to become an affiliate marketer. It’s to build the bridge that the practitioner doesn’t know is missing.

You approach the consultant and say: “You already recommend HubSpot to your clients. What if I built you a recommendation page — branded to your practice, with your voice — that your clients land on before they sign up? You get a better experience for your clients. I handle the infrastructure. We split the revenue that’s currently going to nobody.”

She says yes because the deal is obviously better than the status quo, which is recommending a tool for free. The clients get a curated onboarding experience instead of a raw signup page. The consultant gets passive income from behavior she was already performing. You get a toll position built on referral traffic that was already flowing.

And here’s the part that should make your engineer brain light up: the consultant doesn’t have one tool she recommends. She has a stack. CRM, accounting software, project management, scheduling tools, legal templates, insurance providers. Each one is a referral stream. Each stream is currently unmonetized. Each one represents a toll position you can build in an afternoon.

The practitioner’s dead clicks are the offline version of the creator’s dead clicks — and in many cases, the downstream revenue per referral is ten to fifty times larger. A YouTube creator’s dead click is someone who didn’t buy a $47 ebook. A real estate agent’s dead click is a homeowner who spent $40,000 on a renovation she recommended but never tracked. Same structural problem. Bigger pipe.

A note on regulated industries: healthcare, legal, financial, and insurance referrals carry compliance constraints that make paid referral structures legally complex. Consult qualified counsel before building toll positions in regulated referral markets.

The Second Pattern: Owning the Demand

Every toll position described so far attaches to someone else’s traffic. A creator’s audience. A practitioner’s referral stream. You install infrastructure between existing demand and existing supply, and you collect a toll on the flow.

There’s a second pattern. Instead of attaching to someone else’s demand, you build your own.

A friend of mine wanted to invest in a boat tour operator on Kauai. His plan: fund their marketing — better website, ad account, social presence. I asked him what he’d own at the end of the investment. The answer was nothing. The website would be theirs. The ad account under their business manager. The social following on their profiles. If the relationship ended, he’d walk away with invoices and a lesson.

So he built a page instead. His domain. His content. “Best Nā Pali Coast Boat Tours” — ranked, optimized, owned. He embedded the operator’s booking widget on the page and collected a flat fee per booking. The operator got their cheapest distribution channel. My friend got a ranked asset on a domain he controls.

The structural difference from the standard toll position: the enforcement mechanism is asset control, not relationship pressure. If the arrangement ends, the operator can route the owned demand to another provider under the terms defined in the deal memo. The leverage is permanent because the asset is permanent. No late-night texts about attribution. No quarterly reviews. Both parties know the page exists, and both parties know what happens if the deal ends.

This is the demand-capture asset — a toll position where the operator originates the demand through search rankings instead of attaching to a partner’s traffic. The economics are the same (flat-fee toll on each booking), the plumbing is the same (booking platform as system of record), but the ownership structure inverts. You don’t need the partner’s permission to operate. You need their booking system to clear.

The standard toll position and the demand-capture asset aren’t competing models. They’re complementary substrates. An operator who runs three partner-attached positions and two demand-capture assets has substrate diversity — different risk profiles, different enforcement mechanisms, different traffic sources. When we get to portfolio construction in Chapter 17, this distinction matters.

Next Friday: What does the finished stack actually look like? A typical Tuesday in the life of an operator — including the honest numbers.