Here's what my agency's tool stack looked like for six years:
- Moneybird for invoicing
- ClickUp for projects (paid per seat, eventually expensive)
- Notion for docs
- Google Forms — not Typeform, because Typeform's branded tier was €90 a month
- Google Drive for files
- Spreadsheets for freelancer payouts
- WhatsApp for clients who refused to install anything else
- Slack for the team
- Zapier stitching Google Forms to Slack to ClickUp
- A notes app I never opened
- Word documents with pasted-in screenshots — for every round of design feedback, because there was no system for that
The monthly bill looked reasonable. My accountant didn't flag anything.
I was still losing entire days to coordination overhead. I just couldn't see where it went.
This is the part that doesn't show up in any P&L: the integration tax.
What "free" actually buys you
When a tool says free for up to 3 projects or free up to 5 users, the price isn't a number. The price is the work the tool refuses to do.
A free Google Form doesn't let you brand the page. So clients submit a form with Google's logo on it before they've signed your premium quote. A free ClickUp tier doesn't let you guest-invite clients without friction. So you copy their feedback into a Slack channel they aren't in. Free Notion can hold notes, but it doesn't know what a project is, so you build a database with five linked tables and forget which one is canonical six months later.
The free tier gives you a function. It doesn't give you the workflow. The workflow is what costs money — except in this case the cost is taken out of your hours instead of your card.
The integration tax, in concrete terms
One mundane workflow: a client submits an intake form, and you start a project from it. In a stitched stack, that's nine steps:
- Client fills out a Google Form.
- Zapier picks up the submission and posts it to a Slack channel.
- You read it in Slack. You decide to spin up a project.
- You go to ClickUp. You manually create the project, the tasks, and the assignments.
- You go to Drive. You manually create the folder structure:
Client / Project / Assets / Drafts / Final. - You go to Moneybird. You manually create the client and the quote.
- You go back to Slack. You DM the client a link to the quote, the Drive folder, and a Calendar invite for kickoff.
- The client replies on WhatsApp because that's where they prefer.
- Their kickoff feedback ends up in WhatsApp screenshots that nobody can find later.
Each step is small. None of them is the work. None of them is design. And every single one of them gets repeated for every new project, every week, forever, until you stop running an agency.
It's rarely a failure story — the tools all work. It's a maintenance story. A thousand five-minute tasks that compound into entire afternoons spent not designing.
"But the free stack saved me money"
That's a real argument. You avoided €90 a month for branded Typeform, around €30 a seat per year for ClickUp Business, around €600 a year for Slack's paid tier. Real subscription costs you kept out of your margin.
Now run the math the other direction. This is opinion, not a survey.
Five-person agency. Owner spends, conservatively, five hours a week on tool maintenance — wiring Zaps, fixing folder structures, copy-pasting form data into project boards, hunting for files, reformatting WhatsApp feedback into a brief your designer can actually act on. At a billable rate of €80 an hour (low for an agency owner), that's €400 a week. Roughly €1,600 a month. Around €19,000 a year.
Apply a softer version to each designer who has to context-switch between four tools to find what they need. Even at one hour a week per person, you lose four-figure annual revenue across the team to friction.
Against that, €90 plus €30 plus €600 is a rounding error. The "savings" from the free stack are negative the moment you account for the hours.
I don't have a survey. I have my own six years and the agency owners I've talked to since. Take the math directionally — even if the numbers are half what I've sketched, the conclusion holds.
The cost no one calculates: design feedback
There's one cost I want to single out, because it pushed me hardest.
Design feedback, in a stitched stack, has no native home. Clients send a Word doc with screenshots pasted in. Or a WhatsApp photo with a typed caption: "make this part bigger, the orange feels off." Or — if you're lucky — a Loom recording where they verbally point at a corner of the frame.
The cost is interpretive. Designers spend the first hour of every revision cycle decoding what the client meant before they touch the file. Multiply across every revision round on every project, and you have a structural drag on the deliverable side of the business. Not on tools. On the actual thing you sell.
This is the cost the free-stack math never includes, because it isn't a subscription. It's the slow erosion of your designers' best hours into figuring out what someone meant by "the orange feels off."
What this stack does to your AI
A note for anyone wiring agents into their workflow.
Whether the operator is a human or an AI agent, the coordination problem is the same: information lives in ten places. A human deals with it by tab-switching. An agent deals with it by needing ten integrations, ten authentications, ten different schemas. People are wiring up MCP servers right now — one per tool — hoping the AI will assemble context from email, Slack, WhatsApp, Drive, ClickUp, Moneybird.
It mostly works. It's also fragile. Every new tool is another connector to maintain.
If your data lives in one place, your AI doesn't have to assemble context. It already has it.
This isn't an argument against agents. It's an argument that the substrate matters. Agencies that consolidated before adding AI are getting better answers out of their agents than agencies that bolted AI on top of a stitched stack.
What changed for me
I built Oase with Lam because the math finally tipped. I'd spent six years stitching tools and translating between them. The day I sat down and added it up — the design-feedback time, the freelancer-payout spreadsheets, the marketing context I had to manually pull from five places before I could write a single blog post — the actual subscription savings from the free stack were nowhere near the cost of the workflow they didn't give me.
After switching, I cancelled Typeform (€90/mo), Slack (€600/yr), ClickUp (€30/seat/yr), Notion, ChatGPT, and a handful of smaller subscriptions. Total: more than $2,200 a year in direct subscription savings — and that's the small number. The big number is that the daily friction of running my agency is gone. I open one app, I ask Dune what's on my list today, and I get to work.
I'm not the right voice on whether Oase is the right tool for your agency. That's for you to decide after running a few projects on the free plan. What I want you to take from this piece is the calculation, not the conclusion.
Sit with your stack for a week. Track every task that isn't design. Every form-to-project copy-paste. Every WhatsApp screenshot you transcribe. Every five-minute folder creation. Multiply by your real hourly rate, not the one you bill clients at.
Whatever number you land on — that's what "free" actually costs you.
If it's small, your stitched stack is genuinely fine. Keep it.
If it's closer to mine, the next move is obvious.
Louie co-founded Oase with Lam after six years of running a design agency on a stitched-together tool stack. Oase is built for design agencies and creative freelancers up to 25 people. The free plan includes 3 client portals and 5 projects. Sign up at onoase.com.