Workflow automation sounds great until your manager or accountant asks: “what does it actually deliver?” Then you sit there with a Zapier price tag in your head, a hunch about time savings, and no model to back it up. This article gives you that model. You learn which costs to actually include, which benefits often get forgotten, and how to reach an ROI percentage in four steps that holds up afterwards. Plus three worked examples from Dutch SMBs, with real numbers. No marketer assumptions, just figures (for broader context also read our AI automation SMB guide) you can base an investment decision on.
What is automation ROI exactly?
ROI stands for Return on Investment. In workflow automation, it means how much value an automated process delivers compared to what you put into it. The standard formula is simple: (benefits minus costs) divided by investment, times 100. You get a percentage back. 100 percent means you break even. 500 percent means you get five times your money back.
The problem is not the formula, it is what you put in. Most owners count only the one-time build price as cost and only “time saved” as benefit. That gives nice numbers, but they do not hold up. An honest ROI calculation includes tooling, maintenance, training and opportunity cost. On the benefit side, error reduction, shorter lead time and scalability belong as well. Only then do you get a figure you can still look at three years later without blushing.
Which costs should you include?
This is where most calculations go wrong. Not one cost line, but six. We always separate one-time from recurring, and visible from hidden. An honest automation case looks at a twelve to thirty-six month window.
| Cost item | Type | SMB indication | Forgotten by |
|---|---|---|---|
| Setup / build | One-time | €495 to €3,950 | 0% (visible) |
| Tooling license (Zapier / Make / n8n) | Recurring | €10 to €100 per month | 20% |
| AI or API usage (OpenAI / Claude) | Recurring | €5 to €80 per month | 60% |
| Hosting (self-hosted only) | Recurring | €10 to €30 per month | 40% |
| Maintenance and updates | Recurring | 2 to 6 hours per year | 70% |
| Team training | One-time | 2 to 8 hours internal | 80% |
| Opportunity cost of internal time | One-time | 4 to 16 hours internal | 90% |
The biggest blind spot is opportunity cost. If your operations manager spends twelve hours on intake calls, requirements and testing, that is not free time. Those hours could have gone elsewhere. For internal time, use at least the gross hourly rate plus thirty percent overhead. For an operations role in the SMB, that puts you at around €60 per hour, not the €30 on the payslip.
Want to know which tool category fits your case? Read Zapier vs Make vs n8n first. You will see the same tooling numbers there, plus what to add per tool.
Which benefits should you include?
Benefits are more than time. Five categories belong in every serious case. The first two are always quantifiable, the other three sometimes soft but still countable.
- Time savings in euros. Hours per week saved, times workable weeks (use 46 instead of 52, accounting for holidays), times fully loaded hourly rate including overhead. This is the biggest line in 80 percent of SMB cases.
- Error reduction. How much did errors cost you last year? A wrongly processed invoice costs on average €40 in correction time. A missed lead that went to a competitor, that is a full order value. Research by McKinsey on AI in business operations shows error reduction in automated flows averages 30 to 50 percent.
- Shorter lead time. A quote out the door in two hours instead of two days scores higher. A customer who gets an answer within five minutes is more likely to stay. Translate lead time into conversion percentage where you can steer on it.
- Customer satisfaction (NPS, retention). Harder to quantify, but if your NPS moves from 30 to 45 thanks to faster handling, that maps to retention. Be conservative: number of customers times average annual value times expected retention lift.
- Scalability without extra people. If volume grows by thirty percent without an extra hire, that saves a full annual salary plus employer costs. For an operations role, easily €55,000 per year.
According to CBS figures on Dutch labour productivity, average productivity per hour worked in services hovers around €65 in added value. SMBs that outperform sit higher. For conservative calculations we use €45 per hour as a floor, €75 as a realistic middle.
The formula: ROI in 4 simple steps
With cost and benefit lines mapped, you can calculate ROI. We always use the same four steps, so clients can reproduce afterwards how we got to a number.
- Step 1: add up all one-time costs. That is your investment. Setup price plus training plus opportunity cost of internal hours. Example: €1,950 build + €240 training + €960 internal time = €3,150 total investment.
- Step 2: add up all annual costs. Tooling + API + hosting + maintenance. Example: (€30 + €40 + €15) × 12 + €300 maintenance = €1,320 per year.
- Step 3: add up all annual benefits. Time savings + error reduction + soft benefits (conservatively). Example: 8 hours × 46 weeks × €60 = €22,080 per year.
- Step 4: calculate year one ROI. Formula: ((benefits - annual costs) / total investment) × 100. Example: ((€22,080 - €1,320) / €3,150) × 100 = 659 percent in year one.
For a deeper look at when a no-code approach hits these numbers and when custom code makes a better business case, read no-code vs custom automation. The ROI formula stays the same, the inputs differ a lot.
Example 1: invoice processing
A Dutch installation company with around 1,200 incoming invoices per month. The accounts-payable staff member spent on average 10 hours per week booking, checking and pushing to the ERP. Lots of copy-and-paste, regular small errors in VAT codes or creditor numbers.
We built an automation with document processing where invoices in the mailbox are read automatically, validated against open orders and pushed to the ERP. Human step for exceptions. Setup price €3,500, tooling €70 per month, maintenance €200 per year.
| Item | Calculation | Amount |
|---|---|---|
| One-time investment | €3,500 build + €180 training + €480 internal time | €4,160 |
| Annual costs | €70 × 12 + €200 maintenance | €1,040 |
| Time savings | 10 hours × 46 weeks × €55 | €25,300 |
| Error reduction | 24 corrections per year × €40 | €960 |
| Total benefits year 1 | Sum of benefits | €26,260 |
| Year 1 ROI | ((€26,260 - €1,040) / €4,160) × 100 | 606 percent |
| Payback period | €4,160 / (€26,260 - €1,040) × 12 | 2.0 months |
In year two the number gets even better because the one-time investment falls away. Net benefits: €26,260 minus €1,040 = €25,220 per year, on a recurring cost of €1,040. Effective ROI from year two well above 2,000 percent. That is why we always model over 36 months, not just year one.
Example 2: lead routing and AI classification
A B2B services agency with around 80 incoming leads per week across channels: website form, LinkedIn, email, phone callbacks. Sales people spent about five hours per week on manual qualification, tagging, CRM entry and routing to the right account manager. Conversion to meeting: 12 percent.
We built sales automation where incoming leads are classified by AI on intent and company size, enriched automatically and routed to the right account manager. The Simply case study for recruitment follows the same pattern. Setup price €2,450, tooling plus AI usage €95 per month, maintenance €300 per year.
| Item | Calculation | Amount |
|---|---|---|
| One-time investment | €2,450 build + €240 training + €720 internal time | €3,410 |
| Annual costs | €95 × 12 + €300 maintenance | €1,440 |
| Time savings | 5 hours × 46 weeks × €70 | €16,100 |
| Faster response → conversion up to 15 percent | 3 extra leads/wk × 46 × €450 order value | €62,100 |
| Total benefits year 1 | Time savings + conversion lift | €78,200 |
| Year 1 ROI | ((€78,200 - €1,440) / €3,410) × 100 | 2,251 percent |
This is a case where the soft benefit (faster response time) outweighs the hard time savings. In two-thirds of the sales automations we build, we see this pattern. Research from Harvard Business Review on lead response time shows that leads picked up within five minutes have twenty times higher conversion than leads after an hour. Work that through and the value of routing becomes obvious.
Example 3: AI assistant over own documents
A British consultancy with 22 consultants lost on average 2.5 hours per consultant per week searching through their own reports, meeting notes and earlier deliverables. Classic knowledge overload: everything is somewhere, nobody knows exactly where.
We built them an EU-hosted AI assistant over their document library, including source attribution on every answer. See the full Planit case study for the detailed write-up. Setup price €6,500, tooling plus EU hosting plus API €280 per month, maintenance €600 per year.
| Item | Calculation | Amount |
|---|---|---|
| One-time investment | €6,500 build + €450 training + €1,440 internal time | €8,390 |
| Annual costs | €280 × 12 + €600 maintenance | €3,960 |
| Time savings | 22 consultants × 2.5 hours × 46 weeks × €85 | €215,050 |
| Faster client deliverables (soft) | Not counted | 0 |
| Total benefits year 1 | Time savings only | €215,050 |
| Year 1 ROI | ((€215,050 - €3,960) / €8,390) × 100 | 2,516 percent |
With bigger teams or higher hourly rates, ROI scales up fast. Two hours per person per week sounds modest, but times 22 people, times 46 weeks, times a consultancy rate, you easily land above €200,000 per year. An AI assistant solution like this often pays back within three to four weeks.
What are typical ROI pitfalls?
Five mistakes we see again and again. Avoid them and your model still holds three years later.
- Too optimistic time savings. “We save 20 hours per week” becomes 8 hours in practice. Halve your gut number and be pleasantly surprised if it comes out higher.
- Not counting error costs. A missed invoice or wrong lead costs more than you think. Use at least €40 per admin error and the full order value for a missed deal.
- Forgetting opportunity cost of internal time. Two weeks of your operations manager are not free. Count them at €60 to €90 per hour including overhead.
- Underestimating tooling and API costs. A free Zapier plan does not cover serious volume. Always calculate the subscription at your actual volume, not at the demo flow.
- No maintenance budget. Every automation needs 2 to 6 hours of maintenance per year. Updates, small tweaks, monitoring. Always include this, also for self-hosted.
According to Deloitte research on process automation, average predicted ROI on automation projects sits 30 percent above realised ROI after twelve months. Not because projects fail, but because business cases were too rosy. Stay realistic and you avoid disappointment.
How do you measure afterwards if it held up?
A good ROI calculation does not end at go-live. We schedule a review after three months by default to check if the assumptions held. Three KPIs are usually enough.
- Actual time savings per week. Ask the team before and after, or measure via volume processed times historic average time. We want to know whether the estimated ten hours per week landed at eight to twelve.
- Error rate and correction time. How many corrections did the team make on automation output? Above three percent is a signal to tune, below one percent is a healthy automation.
- Adoption inside the team. Who actually uses the flow? If half the team keeps working around it, your business case is off. Retrain or redesign.
Logical next step + ROI calculator (coming soon)
We are building a free ROI calculator on this site. Soon you will fill in your own volumes, hourly rates and current time spend, and get a calculation for your specific case. Until it ships, we can run the same numbers manually with you in a free intake call.
Plan a free intake call. In thirty minutes we look at your three most time-consuming processes, give an honest ROI estimate, and tell you whether automation pays off. No sales pressure, real numbers you can take back internally. Our way of working and guarantees sit in our terms and conditions, so you know what you are signing up for.
