Replacing the SaaS Stack with $75/mo of Cloudflare

I run Obsidian Axis Group on $74 per month.

That's not a marketing line. It's the actual sum of every line item that keeps the business running, every account, every API, every piece of infrastructure that the rest of OAG depends on. CRM, automation, lead capture, scheduling, reporting, knowledge base, file storage, AI tooling, payments. $74 per month.

Most operators I work with, $2M to $50M revenue, run on $500 to $3,000 per month of SaaS doing the same things, and don't realize how much of it compounds with every new hire. This post is the line-by-line breakdown of what we replaced, what it cost, and how the replacement works.

The before number.

Here is a typical small-operator SaaS stack. I'm using actual numbers from a recent fractional COO engagement at a $22M building-products distributor, anonymized, but real:

That's $28,716 per year, $143,580 over five years. And that's at 10 seats, every new hire moves the per-seat tools up. The sales rep you onboarded last quarter probably bumped HubSpot, Mailchimp, Calendly, and Asana all simultaneously.

The after number.

Here is what we run on now:

The savings number: $2,319/month. That's $27,828 per year. Over five years, including the compounding seat costs as the team grows: well past $200,000 of avoided SaaS spend.

How CRM gets replaced.

HubSpot at $1,000/month is the biggest single line item, so it's the most useful one to walk through. What does HubSpot Sales Pro actually do? Strip away the marketing, and it's: a database of contacts and deals, a way to capture leads from a form, a way to send sequences of emails, a reporting view, and integrations with the rest of your stack.

The StackOS replacement is four pieces:

1. The contacts and deals database is a Supabase table. Two tables actually, `contacts` and `deals`, with foreign keys, indexes, and Row-Level Security policies that mirror what HubSpot calls "team permissions."
2. Lead capture is a Cloudflare Worker route. The form on your site POSTs to it, the worker validates the input, writes to Supabase, fires a transactional email via Resend, and redirects. About 80 lines of code.
3. Email sequences are a Cloudflare Worker on a cron trigger that wakes up nightly, queries Supabase for contacts ready for the next sequence step, and sends via Resend.
4. Reporting is a small dashboard that queries Supabase directly. We use a free Notion view connected via API for non-technical team members.

The first time you build this, it takes a half-day. The second time, two hours, and we now have a template. The cost to run all four pieces, at OAG's volume: under $2 a month.

How automation gets replaced.

Zapier at $280/month is the second-biggest. Zapier's value is connecting two systems that don't otherwise talk. The StackOS replacement is: a Cloudflare Worker that runs on a cron schedule, with code that reads from system A and writes to system B.

For example: "every night, pull yesterday's Stripe charges into a Supabase table, then summarize and email the owner." In Zapier, that's three Zaps and $50/month worth of tasks. As a Worker, it's 60 lines of code on a cron trigger. Cost: under $0.10 per month.

The argument against this is "writing code instead of clicking through a UI." That's only true if you're inventing the pattern from scratch. Once you have one cron-Worker template, every new automation is a copy-paste-edit, not a from-zero build.

What this isn't a fit for.

Three honest caveats.

Companies above 50 seats. StackOS scales fine technically, but the friction of changing UIs across a 50-person org outweighs the savings. Past 50 seats, the savings stop being meaningful as a percentage of revenue.

Teams with no technical operator. Someone has to know how to copy an API key into Cloudflare's dashboard. Not deep engineering, but not zero. If nobody on your team has ever run a script, hire someone for the StackOS Build session ahead of time.

Brand-new companies. If you're at $0 revenue, just use HubSpot's free tier. Don't optimize infrastructure for a business that doesn't exist yet.

How to actually do it.

Three paths.

If you want to do it yourself, the StackOS Framework PDF is $29. It walks through every account, every setup step, every piece of code with copy-paste examples. Most hands-on operators ship the basics in a Saturday.

If you want it built once and built right, book a StackOS Build session, three hours, live, custom to your business. We do the audit, the architecture, and the build with you. Walk in renting your stack, walk out owning it.

If you want it installed as part of a deeper operating-discipline engagement, that's the fractional COO track. Multi-month, embedded, the whole operating system gets the StackOS treatment.

The math on any of the three pays for itself in the first month. The five-year compounded savings is six figures.

Keep reading.

• What We Cut in the First 90 Days, A typical $10M–$50M operator carries 30–40% operational waste. The first 90 days of an Axi…
• How to Read an Operational Waste Audit, Three numbers come out of the Diagnose phase of every fractional COO engagement. Most oper…
• How to Actually Move a Team from Storming to Performing, Most LMM teams stop at Norming and the leadership thinks that is success. It is not. This …
• Lean Six Sigma for the Lower-Middle Market, Without the BS, DMAIC, SIPOC, Black Belts, control charts, kaizen events. At enterprise scale these earn t…

Talk through it.

If any of this is applicable to where you are, book a scoping call. No pitch deck.