The SaaS Tax: Finding the $20,000 You're Wasting on Tools
A 6-month ATS build uncovered $20K+ in wasted SaaS spend. Fillout workarounds, n8n over Make.com, and a $1,200/month Notion bill nobody questioned.
We weren't trying to audit anyone's software spend.
We were building an applicant tracking system for Jemini, a brand consulting firm that had outgrown Monday. Six months in, the recruiting system was live. But somewhere around month three, we started noticing what else was going on with their stack.
Not because we went looking. Because when you redesign how work flows through an organization, the redundancies surface on their own.
Here's what turned up:
| What They Were Paying | What Replaced It | Annual Difference |
|---|---|---|
| ATS vendor quotes ($5K–$15K/year) | Custom build on Airtable + Zite. 3–5 paid Airtable editors. Everyone else (interviewers, hiring managers, leadership) on Zite interfaces, no per-user cost. | $5,000–$15,000 |
| Fillout Business tier ($200/month) for a custom domain | CNAME record in Cloudflare + tny.cc redirect ($5/month) | $2,340 |
| Notion ($1,200/month) | Flagged for replacement with a self-hosted wiki already running on a $16/month Elest.io server | $14,208 |
| Calendly per-seat licensing ($13/user for rotating interviewers) | Fillout Pro ($19/month, unlimited users) | Hundreds per year |
| Monday subscription | Cancelled | Full subscription cost |
North of $20,000 a year. Nobody had approved that number. It just accumulated.
How $200/month becomes $1,200/year for a URL
Jemini needed some features like autocomplete fields and custom domains. Fillout locks those custom features behind their Business plan at $90/month.
The actual technical requirement: redirect one URL to another. A CNAME record pointing company subdomain to a $5/month redirect service handled it. Same result in the browser. $1,020 less per year.
Eric, was blunt about it:
"Not looking to spend $1,200 a year for an address autocomplete. We're doing well. We're not doing that well."
SaaS vendors gate individual features behind tier jumps. You pay 3x the price for 1 additional capability. The other 39 features in that tier go unused.
The per-seat problem, and how architecture solves it
ATS platforms charge per user. Reasonable when five recruiters use the system daily. Less reasonable when 20 interviewers need to submit feedback once a quarter.
The Jemini system handles this at the architecture level. Airtable seats go to the people who build and maintain the backend: recruiters, the ops lead. Three to five paid users. Interviewers, hiring managers, and the CEO access the system through Zite interfaces. No seat fees. The 20th interviewer costs the same as the 3rd.
At $15/user/month for 20 occasional users on a traditional ATS, that′s $3,600/year for people who log in a handful of times over the year.
The subscription nobody questioned
Eric again:
"I actually kind of want to get off Notion. We don't use it, and it's literally $1,200 a month. It's just wild how expensive it is."
$1,200 a month. For an internal wiki. That the team wasn't reading.
The replacement: a self-hosted wiki running on a server Jemini already paid $16/month for (the same Elest.io instance hosting their n8n automations). The marginal cost of adding the wiki: close to zero.
This wasn't a hidden cost buried in an integration. It was a line item on a credit card statement. Charged monthly. For over a year. Nobody had stopped to ask whether it was still worth it.
When per-operation pricing works against you
Make.com bills by operation count. Each API call, filter step, and data transformation ticks the meter. For workflows that run once a day, the cost is trivial. For a 5-minute polling interval running around the clock, it adds up.
Jemini needed Float (their resource management tool) to sync approved PTO requests to Google Calendar. That sync checks for new approvals every five minutes. On Make.com, that volume of operations scales the bill unpredictably.
Moving that workflow to a self-hosted n8n instance on Elest.io collapsed the cost to a flat $16/month. Same logic. Same integrations. Fixed bill.
Eric described the cumulative effect of small SaaS charges:
"I spend 45 minutes doing my expense report every month. It's like death by a thousand cuts."
Running the audit yourself
You don't need outside help for this. You need your company credit card statement and about 90 minutes.
Pull every recurring software charge. Export the statement. Search for the obvious subscriptions and the ones you forgot about.
For each tool, answer three things:
- How many people used this in the past 30 days?
- What breaks if we cancel it tomorrow?
- Are we on this pricing tier because we need it, or because we needed one feature from it once?
Look for three patterns:
The tier tax (paying 3x for 1 feature). The seat multiplier (paying per user for people who log in twice a year). The forgotten subscription (nobody would notice if it disappeared).
Multiply each monthly charge by 12. $200/month is $2,400. $1,200/month is $14,400. Stack three or four of these and the annual number gets hard to ignore.
What this is actually about
The savings at Jemini were a side effect of building a system intentionally. When you map out how work actually flows, you choose tools based on what they need to do. When tools accumulate by accident (someone signed up, it worked, the card kept getting charged), they multiply without justification.
$20,000 a year in unjustified software spend isn't unusual. We see some version of it in most engagements. The number varies. The pattern doesn't.
If you want to go deeper on measuring what no-code work actually returns, we wrote about that here.
And if the audit turns up something bigger than a billing cleanup (a workflow held together by three tools that don't integrate, or a team fighting their software instead of using it), that's a systems problem.
We spend 4–8 hours mapping workflows and identifying where friction actually lives. It costs $299. Every one we've done has paid for itself in the first finding.