The 4 Hooks I Use to Evaluate Whether to Self-Host or Buy SaaS

I run a one-person company that does the work of a small agency. Every month I sign up for a new tool, every month I think about cancelling three others, and every month at least one SaaS vendor pushes a price hike that makes me question whether I should just be hosting the thing myself.

After enough of those cycles, I stopped guessing. I built a four-hook test. Before I add anything to my stack, the tool gets pulled through these four questions. If it fails on the wrong hook, it goes on the self-host list. If it passes all four, I buy.

Here is the test, the reasoning behind each hook, and a worked example from last month.

Hook 1: Availability, what happens if this vendor disappears tomorrow

The first question I ask is brutal but cheap: if this company goes bust on Friday, what breaks on Monday?

For most operators, the honest answer is uncomfortable. The CRM holds your pipeline. The email platform holds your subscriber list. The form builder holds your active funnels. If any of those vanish, you are reconstructing weeks of work from memory.

I rate every tool on a 1-to-5 availability scale:

• 1: vendor disappears, I lose the data and the workflow stops
• 3: vendor disappears, I have the data exported but I am rebuilding for a week
• 5: vendor disappears, I have full backups, alternative providers ready, and a clean migration path

Anything that scores below 3 on the things I actually depend on goes onto the self-host shortlist. Not because self-hosting is free of risk, but because I can survive my own server failure with a backup snapshot. I cannot survive a vendor I never had control of.

Hook 2: Data lock-in, how easy is the exit

This is the hook everyone underweights. The free trial looks great. The features look great. Then you check the export options and discover you can pull a CSV of contacts but not the email history, or you can export the email history but not the segmentation logic.

When I evaluate any tool now, I ask three sub-questions:

1. Can I export every record I would need to rebuild this workflow elsewhere?
2. Is the export format usable, or is it deliberately mangled?
3. If the vendor changes their pricing model, can I be off the platform in under a week?

If the answer to any of those is no, the tool gets a hard mark against it. I will still use SaaS tools that lock me in, but only when the productivity gain in year one is high enough to justify the cost of leaving in year three.

The two categories where I never accept lock-in are: anything that holds my client relationships, and anything that holds my content history. Both compound over years. Losing access is a setback I cannot afford.

Hook 3: Pricing curve, what happens when I grow

Most SaaS pricing is fine at the start. The pain shows up between user 10 and user 100, or between 1,000 contacts and 10,000 contacts, or when you add a second business and discover the per-seat model now charges you twice.

I plot a rough five-year pricing curve before I buy anything that scales by usage. The questions:

• What does this cost me at current usage?
• What does it cost me at 5x usage?
• What does it cost me at 20x usage?
• At what point does the annual cost exceed the cost of running my own server with the same capability?

For most categories there is a crossover point. Below it, SaaS wins on convenience. Above it, self-hosting wins on economics. The mistake operators make is staying on the SaaS side of the crossover for years past the breakeven point because switching feels harder than paying.

When I do the maths properly, switching is usually a one-week project. The annual saving funds the next thing I want to build.

Hook 4: Team access, who else needs to touch this

The fourth hook is the one I forget the most often. I will choose a self-hosted tool that I love, and then realise three months later that I need a contractor, a virtual assistant, or my brother to use it, and the access model is not built for shared use.

The questions I ask now:

• How many people will need access in 12 months?
• Do those people need to be technically competent to use it?
• Can I onboard a new user in under 10 minutes?

This hook is the main reason I still pay for some SaaS tools that fail the first three hooks. If a tool needs to be used by a non-technical collaborator on a Tuesday afternoon and they cannot wait for me to fix it, the calculus changes. I will accept lock-in and a worse pricing curve in exchange for an interface someone else can use without me.

A worked example: replacing my email platform last month

Last month I migrated my newsletter off a SaaS provider that had quietly raised prices three times in 18 months.

Running the four hooks:

• Availability: subscriber list was exportable but campaign analytics were not. Score: 3.
• Data lock-in: CSV export available, segments not, templates not. Medium lock-in.
• Pricing curve: at my current list size, fine. At 5x, the price doubled. At 20x, I would be paying more per month than my server hosting costs annually.
• Team access: only I needed access. Self-hosting was viable.

The self-host equivalent took me a weekend to set up. It runs on the same server as three other tools. Total monthly cost: zero, on top of the server I was already paying for. The migration moved 1,800 subscribers in under an hour.

The new arrangement scores 5, 5, 5, 4 against the four hooks. The only weakness is team access, because the admin interface assumes the operator can read a config file.

The playbook I now run

When evaluating any tool, I do the following in order:

1. Run the four-hook scan before I sign up for anything.
2. For any tool scoring below 12 across the four hooks, look up the self-host equivalent on the same day.
3. If a self-host option exists and would land me at 16 or higher, allocate one weekend per month to migrate one tool off SaaS.
4. Keep a written log of what I self-host, what I pay for, and the reasoning. Review every six months.

The biggest payoff of running this discipline is not the money saved. It is the calm that comes from knowing what would happen if any single vendor in my stack went away tomorrow. The answer is now: not much. That is worth more than any feature comparison chart.

If you want the full stack I built around this approach, my Self-Hosted Infrastructure Stack for Solo Founders walks through every layer. If you want the lighter version, my notes on agent-readiness cover the AI side of the same problem. More on how I work and what I do is on the about page.

For the agency build side of this thinking, Weir Digital Media productises the same approach for clients who want it done for them.

Built by Weir Digital Media.