Build
The Un-SaaS Ch. 9 — The $15/Month Stack
The entire toll position infrastructure runs on fifteen dollars a month. Here is every tool, why it was chosen, and what to skip.
August 21, 2026 · 12 min read
Chapter 9 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.
Part III — Learn It: How the Machine Works
I once watched a consultant spend forty-five minutes explaining why his tech stack cost $387 a month. ConvertKit at $79. Leadpages at $49. Zapier at $49. Calendly at $16. A CRM at $99. Plus a handful of smaller tools, each solving one problem that a previous tool created.
He was proud of it. He called it his “revenue engine.” It processed about $2,200 a month in affiliate revenue.
That’s a 17.6% infrastructure tax. On a toll position, where the whole point is operating lean and compounding margins, a 17.6% cost floor before you’ve earned a dollar is an anchor around your neck.
Here’s a different approach: $15 a month. Total. For everything.
Not $15 per tool. $15 total, across your entire stack, running multiple toll positions simultaneously. The same infrastructure that the consultant built for $387 — landing pages, email, redirects, analytics, automation — on a budget that rounds to a large coffee twice a month.
That’s not a compromise. That’s the control principle in action.
The Control Principle
Before I show you the stack, I need to explain why it’s built this way.
The control principle is the single most important architectural decision in a toll position: your infrastructure, your kill switch.
Every piece of infrastructure the operator builds must run on systems the operator owns and controls. Not the partner’s systems. Not a vendor’s walled garden. Not a platform that can change terms, raise prices, or shut off access.
Here’s why this matters.
If your landing page lives on the partner’s WordPress site, the partner can remove it. If your email sequences run through the partner’s ConvertKit account, the partner can lock you out. If your analytics are embedded in the partner’s Google Analytics, you lose the data when the partnership ends.
Your infrastructure runs on your accounts, your hosting, your domain. The partner’s only touchpoint is a single URL — one link they paste into their YouTube description or podcast show notes. Everything behind that URL is yours.
If the partnership dissolves, you turn off the landing page. The URL goes dead. The partner loses their conversion lift and has to start from scratch. You keep your data, your experiment log, and your subscriber relationships.
That asymmetry — where ending the partnership hurts the partner more than it hurts you — is the structural feature that makes toll positions durable. And it only works if you control the infrastructure.
The Four Layers
A toll position requires exactly four layers of infrastructure. Not five. Not twelve. Four.
Layer 1: Landing Pages.
This is where dead clicks become captured subscribers. A single-page site — one headline, one persuasion flow, one email capture form, one redirect — sitting between the partner’s traffic source and their checkout.
You don’t need Leadpages or Unbounce. A static HTML page hosted on Cloudflare Pages, Netlify, or Vercel is free. If you prefer a visual builder, Carrd is $19/year (not per month — per year). Either way, the page loads in under a second, has zero dependencies, and costs effectively nothing.
For operators comfortable with code: a simple HTML template with Tailwind CSS and a form that posts to your email provider’s API. You build it once, clone it for each new position, and customize the copy. Total cost: $0.
Layer 2: Email.
This is where captured subscribers become buyers. A welcome sequence (five to seven emails over fourteen days), ongoing broadcasts (weekly or biweekly), and segment-specific campaigns as the list grows.
The email provider is the one place worth spending money, because deliverability matters. A cheap provider that lands in spam is worse than no provider at all.
Options in the $0-$15 range:
- Buttondown: free up to 100 subscribers, $9/month up to 1,000
- Mailerlite: free up to 1,000 subscribers
- Brevo (formerly Sendinblue): free tier with 300 sends/day
- Postmark: $15/month for transactional + broadcast, exceptional deliverability
For a new operator with one or two positions and a few hundred subscribers, the free tiers work fine. As the list grows past 1,000, the $9-$15/month tier covers multiple positions simultaneously.
Layer 3: Redirects.
This is the invisible layer most operators skip — and the reason most operators can’t optimize.
A redirect layer routes every outbound click through your own infrastructure before forwarding to the destination (the partner’s checkout, an affiliate link, a product page). This gives you three things raw affiliate links never provide. Click tracking: which subscribers clicked what, when, from which email. A/B destination testing: send 50% of clicks to the partner’s checkout and 50% to an alternative. Dynamic routing: change the destination without changing the link in your emails.
You can build a redirect layer with a $0 Cloudflare Worker, a simple database (a JSON file or a free Supabase instance), and about two hours of engineering time. Or you can use a tool like Switchy or PixelMe at $0-$7/month.
Layer 4: Analytics.
This is where data becomes intelligence. Page views, capture rates, email opens, click-through rates, conversion rates, and the experiment log.
Plausible Analytics: $9/month for up to 10K pageviews (privacy-friendly, lightweight). Or self-hosted Umami: $0. Or simple event tracking via Cloudflare Workers: $0.
For the experiment log itself: a structured spreadsheet (Notion, Google Sheets, or Airtable free tier) where each test gets a row with hypothesis, variant, sample size, result, and learning. This is the data moat. It doesn’t need fancy tooling — it needs discipline.
The $15 Configuration
Here’s what the stack looks like assembled, at the $15/month price point:
| Layer | Tool | Monthly Cost |
|---|---|---|
| Landing Pages | Carrd ($19/year) or static HTML on Cloudflare | ~$1.58 or $0 |
| Buttondown (1K subscriber tier) | $9 | |
| Redirects | Cloudflare Worker or Switchy free | $0 |
| Analytics | Plausible or self-hosted | $4 or $0 |
| Experiment Log | Notion/Sheets (free) | $0 |
Total: $10–$15 per month. Compare that to the consultant’s $387. Same capabilities. Different architectural philosophy.
The savings aren’t the point. The control is the point. Every component in this stack runs on your accounts, under your domain, controlled by your login. Nothing is entangled with a partner’s systems. Nothing is locked into a vendor’s walled garden. If any single tool doubles its price or shuts down, you swap it in an afternoon.
And the portability matters more than most operators realize on day one. Each tool in the $15 stack talks to the others through standard APIs and webhooks — not through proprietary integrations that lock you into a vendor’s ecosystem. The landing page posts a form to Buttondown’s API. The redirect Worker fires a tag back to the email system. The analytics layer reads events from both.
When you outgrow a component, you swap it. If Buttondown triples its pricing next year, you migrate your subscriber list to Mailerlite in an afternoon and update one API endpoint. Your landing pages don’t change. Your redirects don’t change. Your experiment log doesn’t change.
Try that with ClickFunnels. Your pages, your sequences, your automations, your analytics — all living inside one vendor’s walled garden. Migrating means rebuilding from scratch. Which means you never migrate, even when the tool stops serving you. That’s not infrastructure. That’s a hostage situation with a monthly subscription fee.
The $200-$400 Alternative (and Why You Don’t Need It)
To be fair, there’s a version of this stack that costs $200 to $400 a month and provides genuine additional features. ConvertKit or ActiveCampaign for advanced email automation. ClickFunnels or Leadpages for drag-and-drop page building. Hyros or RedTrack for advanced attribution.
Let’s itemize what that money actually buys. WordPress hosting on WP Engine or Kinsta: $30 to $60 a month. A premium page builder like Elementor Pro or Leadpages: $37 to $49. A mid-tier email provider like ConvertKit or ActiveCampaign: $49 to $79. An analytics suite like Hyros or RedTrack: $49 to $99. A form tool or quiz builder: $19 to $39. A link management tool: $12 to $20. A scheduling tool, a CRM add-on, a Zapier plan to wire them all together — another $30 to $65.
That’s seven to nine tools, each solving one problem, each with its own login, its own billing cycle, its own API quirks, and its own support queue. The consultant I mentioned at the top of this chapter? He spent more time managing the integrations between his tools than he spent optimizing what those tools produced.
Now itemize the $15 stack. Cloudflare Pages for landing pages: $0. Buttondown for email: $9. A Cloudflare Worker for redirects: $0. Plausible for analytics: $4 to $6. Notion or Google Sheets for the experiment log: $0. Four tools. Four logins. Zero integration middleware. The landing page posts directly to the email API. The redirect Worker writes directly to the analytics layer. No Zapier. No middleware tax. No “the integration broke at 2 AM and nobody noticed until Thursday.”
These tools are good. They’re also designed for agencies running campaigns for multiple clients at enterprise scale. A solo operator with one to twelve toll positions doesn’t need enterprise-grade email automation. You need reliable email delivery, simple landing pages, and clean analytics.
The $200-$400 stack makes sense when your portfolio is generating $15,000+ per month and the additional features produce measurable lifts. At that point, upgrading is an investment, not a cost. Until then, the $15 stack does everything you need without eating your margins.
Start lean. Upgrade from evidence, not aspiration.
The Kill Switch
Here’s the scenario nobody thinks about until it happens.
A partnership goes sideways. Maybe the creator pivots to a new niche. Maybe they bring in a manager who wants to renegotiate your deal from 20% to 8%. Maybe the creator says something publicly that makes you not want your infrastructure associated with their brand.
On the $15 stack, you shut down that position in under ten minutes. Change the landing page to a redirect to a generic resource. Pause the email sequences for that partner’s subscribers. Update the Cloudflare Worker to route clicks to an alternative. The partner’s YouTube description link now points to your redirect, and your redirect goes wherever you decide. Your subscribers, your data, your experiment log — all intact, all portable, all ready to attach to the next partner.
On someone else’s infrastructure — the partner’s WordPress site, the partner’s ConvertKit account, the partner’s analytics — you walk away with nothing. No subscriber list. No experiment data. No behavioral tags. Months of optimization intelligence locked inside someone else’s login.
The kill switch isn’t a theoretical concern. Every experienced operator has at least one story about a partnership that ended abruptly. The ones who controlled the infrastructure kept their assets. The ones who didn’t started over.
This is why the control principle comes before the cost principle. A $50/month stack you own beats a $15/month stack someone else controls. It happens that the $15 stack gives you both — cost efficiency and total ownership. But if you had to choose, choose ownership every time.
The kill switch works the other direction too. If a partner tries to renegotiate from a position of “we don’t really need you anymore” — and some will — the math changes the moment they look at what happens without you. The position reverts to the partner’s pre-existing setup — whatever they had before you arrived. You don’t threaten. You don’t need to. The partner’s own analytics tell the story: they know what their conversion rate was before your infrastructure and what it is now. The deal memo specifies notice periods and exit terms, so both sides have time to plan.
The partner can always remove or replace your URL on their end. The control principle protects the operator’s investment in infrastructure and data — it doesn’t hold anyone hostage. But a partner who has seen six months of lift data rarely wants to go back to what they had before.
That value gap — where the operator’s infrastructure produces measurably more revenue than the partner’s pre-existing setup — only compounds if you own the infrastructure. On someone else’s stack, the value you created is locked inside someone else’s login. And when it is, you’re not an operator. You’re a contractor with a revenue share pretending to be something more.
What Each Layer Actually Does for You
Chapter 10 walks through the wiring details. Chapter 14 turns that stack into the 10-day deployment sprint that gets it all running. But the philosophy matters before the implementation, so here’s why each layer earns its place.
The landing page layer is the capture mechanism. Its only job is converting anonymous traffic into a known subscriber. A static HTML page on Cloudflare Pages loads in under 400 milliseconds, has zero JavaScript dependencies, and gives you a Core Web Vitals score that would make a Google engineer weep. More importantly, you control every pixel. No platform branding. No “Powered by Leadpages” footer eroding the creator’s voice match. No template constraints forcing your pre-sell content into someone else’s layout decisions.
The email layer is the monetization engine. Everything that earns revenue runs through email — the welcome sequence, the broadcast campaigns, the reactivation sends. Buttondown at $9/month handles up to 1,000 subscribers with full automation and an API that lets your redirect layer fire tags on every click. At 1,000 subscribers across two or three positions, you’re generating $1,500 to $3,500 per month on a tool that costs less than a sandwich.
The redirect layer is the intelligence nerve center. Every outbound click — every product recommendation, every affiliate link, every CTA button — routes through a Cloudflare Worker you wrote in fifteen lines of code. That Worker logs who clicked what, when, from which email, on which device. It fires behavioral tags back to Buttondown. It sets affiliate cookies. And crucially, it lets you change any destination URL once and have every link ever sent update instantly. When a merchant drops their commission rate, you don’t rewrite fourteen emails. You change one line in a Worker.
The analytics layer is the feedback loop. Plausible or Umami tells you what’s happening on the landing page — capture rates, scroll depth, bounce rates, traffic sources. Combined with the email platform’s open and click data and the redirect layer’s click-to-conversion data, you have a complete picture from first visit to final purchase. Three data sources, aggregated in a fifteen-minute weekly review, producing the twelve numbers from your operator’s dashboard.
Four layers. Four jobs. Zero redundancy. The $387 consultant had nine tools doing the work of four — and three of them existed solely to fix problems the other six created.
What Your AI Agents Need
One more layer that doesn’t appear in the cost table: the AI agent workforce.
Your agents handle drafting (landing page copy, email sequence variants, experiment hypotheses), QA (checking links, previewing emails, validating analytics events), and routine optimization (promoting experiment winners, reverting losers, flagging anomalies).
The agents run on whatever AI provider you prefer. As of 2026, the cost per agent task is measured in cents, not dollars. A full day’s worth of agent operations — drafting three email variants, analyzing a week’s conversion data, proposing two new experiments — costs less than a dollar in API calls.
The agent workforce isn’t a separate infrastructure cost. It’s a per-task operating expense that scales with activity. When you’re building a new position, agent costs might hit $3-$5 for the week. When all positions are running in maintenance mode, it drops to cents per day.
The $15 stack is human infrastructure. The agent workforce is operational muscle on top of it. Together, they give a solo operator the capabilities of a small agency at 1/30th the cost.
Next Friday: Domain, hosting, email, analytics, payments — the complete technical architecture, wired end to end.
Build
More on building toll infrastructure?
This article covered one piece of the build dimension. The newsletter delivers a new install pattern every week.
Unsubscribe in one click. Useful operator material only. Promotions, sponsors, and partner links will be disclosed.