Why most automation projects disappoint, and how to avoid being next

Every year, manufacturing companies across the world invest millions in automation. New robots, vision systems, "Industry 4.0" upgrades, and expensive software with impressive demos. And every year, a significant portion of these investments fail to deliver what was promised.

The failures rarely happen because the technology doesn't work. The technology usually works fine. The failures happen earlier, in the decisions that led to buying that particular solution in the first place. There's a pattern in the projects that succeed versus the ones that disappoint: the difference isn't budget or technology sophistication. It's whether someone asked the right questions before signing anything.

Here are five questions that can save you from a costly mistake.

1. What problem are we actually solving?

This sounds obvious, but it isn't. When a vendor shows you their solution, they're showing you what their technology does well, which is their job. But their solution was built for a general problem, not necessarily your problem.

I've seen companies invest €200,000 in automated quality inspection because "we have quality issues." Six months later, they still have quality issues because the real problem was upstream in their process, not in the inspection phase. The automation worked perfectly; it just inspected defects that shouldn't have existed in the first place.

Before any automation discussion, you should be able to write down what specific problem is costing you money today, how you know this is the actual root cause, and what "solved" would look like in measurable terms. If you can't answer these clearly, you're not ready to evaluate solutions.

2. What happens if this doesn't work?

Every technology vendor will tell you their solution works. They'll show you case studies, references, and impressive statistics, and they're probably not lying. It works somewhere, for someone. But will it work here, in your environment, with your products and your people? The honest answer is: maybe.

Technology projects carry real risk, and the question isn't whether risk exists but whether you've thought through what happens if things go wrong. Can you reverse this decision if it fails, and at what cost? What's your fallback plan during implementation? What happens to production if the system goes down? Who owns the problem when something breaks at 2 AM on a Saturday?

You should also ask the vendor directly what percentage of their implementations meet the promised ROI timeline, and whether they can connect you with a customer where things didn't go as planned so you can learn how they handled it. The vendors who get nervous at these questions are telling you something important.

3. Who else has done exactly this, and can we talk to them?

References matter, but generic references don't. "We've implemented this at 50 companies" means nothing if none of those companies are similar to yours. A vision system that works brilliantly for automotive parts might fail completely with your products, and a robot cell that runs 24/7 at a large factory might be completely wrong for your batch sizes.

What you actually need is a reference customer in a similar industry, with similar products, volumes, or challenges, who has been running the system for at least 12 months, and who will speak with you honestly without the vendor on the call. If the vendor can't provide this, that's a red flag. Either they don't have relevant experience, or their customers aren't happy enough to recommend them.

When you get the reference call, ask what surprised them after implementation, what they would do differently, what costs or challenges weren't in the original proposal, and whether they would buy from this vendor again. These conversations are often more valuable than any demo or sales presentation.

4. What's NOT included in this quote?

The proposal looks complete and the number seems reasonable, so you're ready to move forward. But before you do, read it again and look specifically for what's missing.

Common items that "aren't included" until they suddenly are: integration with your existing systems like ERP, MES, or quality systems; training beyond the initial handover; spare parts and maintenance contracts; software licenses after year one; modifications when your product changes; support in your language and timezone; the engineering time from your own staff during implementation; production downtime during installation; and the "phase 2" that's somehow always necessary.

A €100,000 project can easily become €150,000 when these items surface, and they always surface, usually after you've signed. Ask explicitly what the proposal assumes you will provide or do yourselves, what the total cost of ownership over 5 years looks like including maintenance and licenses, and what a typical "phase 2" would include and cost. Get the answers in writing.

5. What do we need to change beyond the technology?

This is the question companies most often skip, and it's also the question that most determines success or failure. Technology doesn't work in isolation. It works within your processes, operated by your people, maintained by your team, and managed by your organization. If you install a sophisticated automation system but don't change how you work around it, you'll get sophisticated disappointment.

Consider whether you have people who can operate this system, and if not, how you will get them. Think about whether this requires process changes upstream or downstream, who will maintain it after the vendor leaves, whether your production planning needs to change, how this will affect the people currently doing this work, and whether your organization is actually ready to work differently.

The companies that succeed at automation treat it as an organizational change project that includes technology, not a technology project that the organization will somehow adapt to.

The real question behind all these questions

You'll notice a pattern in these five questions: they're not about the technology. They're about clarity. Clarity about what you're trying to achieve, clarity about risks, clarity about what you're actually buying, and clarity about what needs to change.

The technology vendors can't give you this clarity. It's not their job, and they have obvious incentives to skip past these questions toward a signed contract. Your internal team might not be able to give you this clarity either, because they're busy running operations, may not have done this before, and might be too close to see the blind spots.

This is where an outside perspective helps. Not someone selling you technology, but someone helping you think clearly about whether, what, and how to buy.

A final thought

The best automation investment might be no automation investment, at least not right now, not this solution, not before you've fixed something else first. That's not a popular opinion in an industry that profits from selling technology, but it's often true.

Ask the hard questions. Get clear answers. Make decisions you can defend in two years, not just today.

Eemil Kiviahde is the Principal Consultant at Kitron Consulting, helping manufacturing and industrial companies make better technology decisions. He offers fixed-price assessments and independent evaluation of technology investments, with no vendor affiliations and no technology to sell.

→ Contact for a free 20-minute consultation

Blog