Why I Tell 80% of My Clients Not to Build Custom Software

A founder messaged me last year. He wanted a custom platform to manage client onboarding for his marketing agency.

He'd already gotten a quote. €45,000. Six months of development. Custom dashboard, automated workflows, client portal, reporting — the whole thing.

He asked me to build it.

I told him not to.

Instead, we built his entire onboarding system using Airtable, Make.com, and a few well-placed automations. Total cost: about €6,000. Took two weeks. His team was using it on day 15.

A year later, it's still running. Zero downtime. He's onboarded over 200 clients through it. And he saved roughly €39,000 he would have spent on software he didn't need.

This isn't a one-off story. This is what happens in about 80% of the conversations I have with founders.

They come to me wanting custom software. I talk them out of it.

And somehow, that's become the best business decision I've ever made.

The instinct to build

I get it. The instinct to build custom software is strong.

You have a problem. You imagine the perfect solution. Something built exactly for your business, with your logic, your features, your interface. Nobody else has it. It's yours.

It feels like the right move. Especially if you're technical yourself, or if you've raised money and feel like you should be building something proprietary.

But there's a gap between the software you imagine and the software you actually need.

Most founders don't need custom software. They need their processes to work without manual effort. Those are two very different things.

Custom software is a product. It needs to be designed, built, tested, deployed, hosted, maintained, updated, and eventually rewritten when the framework it's built on becomes outdated.

An automation workflow is a system. It connects tools you already use, runs in the background, and solves the actual problem — which is almost always: "my team is wasting hours on repetitive manual work."

I've spent over 10 years building software. I co-founded a development agency with a team of 12 people. We build custom software for clients every day.

And I'm telling you: most of the time, you don't need it.

When custom software actually makes sense

Let me be clear — I'm not anti-custom software. I run a dev agency. It's literally how I make money.

But I only recommend it when it's genuinely the right answer. That's about 20% of the time.

Here's when custom software makes sense:

You're building a product, not a process. If software IS your business — if you're building a SaaS tool, a marketplace, a platform that users pay to access — then yes, you need custom development. You can't build a product on top of Airtable and Zapier. Well, you can for the MVP, but eventually you need real infrastructure.

You have 50+ users who need a polished interface. When your team or your customers are the ones using the tool daily, the experience matters. No-code tools are powerful but they have UI limitations. If the interface needs to feel like a product, you probably need to build one.

Your requirements are genuinely novel. If no existing tool or combination of tools handles your use case, custom development might be the only path. But be honest with yourself here. Most processes aren't as unique as founders think they are. "We do onboarding differently" usually means "we do onboarding with three extra steps." That's not a reason to build custom software. That's a reason to add three steps to your automation.

Compliance or security demands it. Regulated industries — healthcare, finance, legal — sometimes need custom systems to meet specific compliance requirements. Data sovereignty, audit trails, access controls. Sometimes off-the-shelf tools can't meet those standards. Sometimes.

That's the 20%.

The other 80% don't need custom software. They need someone to connect the tools they already have and automate the parts that waste time.

The 80%: what they actually need

Most founders come to me with some version of the same story:

"We use Slack, Google Sheets, HubSpot, Gmail, and Notion. Nothing talks to each other. My team spends half their time copying data from one place to another. I want a system that does it all."

And they think "a system that does it all" means custom software.

It doesn't. It means automation.

Here's what these projects usually look like:

Client onboarding. The founder I mentioned at the top? His whole problem was that onboarding a new client took 8 hours of manual work across 5 different tools. Contract signed → project management setup → welcome email → access granted → first meeting scheduled. All done by hand.

We automated the entire thing. Contract signed in PandaDoc triggers a Make.com scenario that creates the project in Asana, sets up the client folder in Google Drive, sends the welcome email sequence, grants platform access, and schedules the kickoff call. His team gets a Slack notification when it's done.

Eight hours of work became zero hours of work. No custom software needed.

Lead management. A SaaS startup was manually routing leads between three sales reps. Leads came in from the website, LinkedIn ads, and a landing page. Someone had to check each lead, figure out who should handle it, forward the email, update the CRM, and send a follow-up.

Response time was 24-48 hours. Leads were falling through the cracks.

We built a workflow: lead comes in → automatic qualification based on company size and source → routed to the right rep → personalized email sent within 5 minutes → CRM updated → Slack notification to the rep. Response time dropped to under 5 minutes. Zero leads lost.

Total build time: about a week. Cost: a fraction of what a custom lead management system would have been.

Invoice processing. An e-commerce agency was generating invoices manually after every order, tracking payments across three systems, and chasing late payments by hand. Their accountant was losing her mind.

We automated the whole chain. Order placed → invoice auto-generated → payment tracked → reminders sent for overdue invoices → everything synced to their accounting software → monthly report generated automatically.

They saved about €2,400 a month in labor costs. Error rate dropped by 95%.

These are not unusual examples. This is what most automation projects look like. Real problems, solved with tools that already exist, connected by workflows that run silently in the background.

The costs nobody tells you about

Here's what the custom software pitch usually looks like:

"We'll build exactly what you need. €30K-€80K. Three to six months. You'll own the code."

Sounds reasonable. But here's what happens after launch:

Maintenance. Custom software breaks. APIs change, dependencies need updating, security patches come out. Someone needs to maintain it. If you're lucky, that's the team who built it. If they've moved on to another project, you're hiring someone new to learn a codebase they didn't write. Budget at least 15-20% of the build cost per year for maintenance.

Hosting. Your software needs to live somewhere. Cloud hosting costs vary, but for a real application with a database, file storage, and decent performance, you're looking at €200-€1,000+ per month. That adds up.

The developer who leaves. This is the one nobody talks about. The developer or agency who built your system had context. They understood your business, your edge cases, your weird requirements. When they leave — and they will eventually — that context walks out the door. The next person has to reverse-engineer what the first person built. That takes time and money.

Scope creep and v2. After launch, you'll want changes. New features, new integrations, that one thing you forgot to mention in the initial spec. Each change costs money and time. Before you know it, you're building v2.

The rewrite. Frameworks age. What was modern 3 years ago is legacy today. At some point, someone will tell you the whole thing needs to be rewritten. And you're back to square one.

Now compare that to an automation workflow:

A Make.com scenario costs €9-€99/month depending on usage. Maintenance means checking it once a month and updating it when a tool changes. If it breaks, the error log tells you exactly where and why. If you want to change it, you drag and drop modules in a visual editor. No hosting. No dependencies. No rewrites.

The total cost of ownership is not even close.

Why being honest makes me more money

This might seem like a bad business strategy. I run a development agency. Why would I tell clients not to build software?

Because the clients I save from a €50K mistake don't forget it.

They come back when they DO need custom development — and now they trust my recommendation because I didn't try to sell them something they didn't need the first time.

They refer their friends. "Talk to Enis. He'll be honest about whether you actually need to build something."

They become long-term retainer clients. The automation I built works, they want more of it, and they want someone to keep optimizing it.

Being honest isn't just the right thing to do. It's the highest-ROI sales strategy I've found.

Over 10 years in software, the thing that's grown my business the most isn't a marketing funnel or a sales script. It's telling people the truth — even when the truth means a smaller invoice today.

How to decide for yourself

If you're a founder trying to figure out whether you need custom software or automation, ask yourself these questions:

Can your process be drawn as a flowchart? If you can map it step-by-step on a whiteboard — trigger, action, condition, action — it can almost certainly be automated without custom code.

Is your team under 50 people? Small teams rarely need custom internal tools. The overhead of building and maintaining them outweighs the benefits. Use existing tools and connect them.

Do your needs change every 6 months? If your processes are still evolving — and for most growing companies, they are — custom software becomes a liability. You'll pay to rebuild things you're still figuring out. Automation workflows can be updated in an afternoon.

Do you have a full-time developer to maintain it? If not, don't build it. Custom software without dedicated maintenance is a ticking time bomb.

If you answered yes to most of these, automation is probably your answer. Not custom software. Not a €50K project. Just the right tools, connected properly, running in the background.

And if you do turn out to be in the 20% that genuinely needs custom software — I'm happy to build it. That's what Outecho does every day.

But I'll only recommend it if it's truly the right call.

One more thing

I wrote a free book called The Invisible Employee that walks you through how to identify what to automate and how to build your first workflow. It uses the same frameworks I use with clients.

If you want to explore automation before investing anything, start there: /book

And if you want someone to look at your specific situation and tell you honestly what you need — custom software, automation, or nothing at all — book a free 15-minute assessment. No pitch, no pressure. Just an honest answer.

That's what I do.