← Back to Field Notes

March 5, 2025 · The Bloomfield Team

7 Questions to Ask Before Adding Technology to Your Shop Floor

A 2024 survey by the Manufacturing Leadership Council found that 62% of technology investments in small and mid-size manufacturing operations fail to deliver their projected ROI within two years. The failure rate is not a technology problem. The decisions leading up to the purchase are where the damage happens.

Before signing a contract, committing engineering time, or redirecting budget, seven questions deserve honest answers. The order matters. Each one builds on the one before it.

1. What specific process breaks today, and what does that breakage cost?

Every technology purchase should trace back to a measurable process failure. Quoting takes too long and you lose bids. Setup times run 40% over estimate and erode margins. Operators repeat the same mistakes on a part family because tribal knowledge never made it into a document.

If you cannot name the broken process in one sentence and attach a dollar figure within an order of magnitude, the project is not ready. The number does not need to be precise. It needs to be real. "We lose approximately $200,000 per year in scrap on titanium jobs because setup parameters are inconsistent across shifts" is a project foundation. "We need to modernize our operations" is a budget request without a destination.

2. Who owns this process today, and do they agree it is broken?

Technology that solves a problem the operator does not recognize will sit unused. The estimator who has built quotes the same way for 15 years may see nothing wrong with a three-day turnaround. The production scheduler who manages the board with magnets and dry-erase markers may consider that system perfectly functional.

Both may be right. And both may be wrong. The point is not to override their judgment. The point is to verify that the people who will actually use the tool agree that the current state needs to change. Without that buy-in, the most sophisticated software in the world becomes the most expensive screen saver in the building.

3. What data does the solution need, and does that data exist in your operation?

Every manufacturing technology runs on data. An AI quoting tool needs historical job records with consistent part descriptions, material costs, cycle times, and margins. A scheduling system needs machine availability, job priorities, and realistic setup time estimates. A quality system needs inspection data tied to specific operations, operators, and machines.

The question is whether your operation generates that data in a form the technology can use. If your ERP contains five years of job records with consistent naming conventions and complete cost breakdowns, you are in good shape. If your ERP is a data entry afterthought where descriptions vary by whoever typed them and half the fields are blank, the data preparation work comes first.

Skipping this step is how manufacturers end up six months into a project with nothing to show for it. The technology works. The data does not.

4. How does this connect to the systems you already run?

A standalone tool that requires manual data entry to function is a tool that will stop being used within 90 days. The integration question is not a technical afterthought. It is a requirement.

Can the solution read from your ERP? Can it pull from your quoting spreadsheets, your supplier emails, your CAD files? Does it push results back into the systems where your team already works, or does it create a new destination they have to remember to check?

For more on this challenge, see our guide on connecting systems that were never meant to talk to each other.

The best manufacturing technology sits inside the existing workflow. The operator does not change what they do. They get better information while doing it.

5. What does your team need to learn, and how long will that take?

Every new tool carries a training burden. The question is how heavy that burden is and who bears it. A tool that requires a two-hour orientation and produces value on day one is a different commitment than a system that requires 40 hours of training across three departments and a dedicated internal champion for six months.

Neither is inherently wrong. Both need to be understood before the commitment. A shop running at full capacity with zero slack in the schedule cannot absorb a six-month implementation without production impact. That cost needs to appear in the ROI calculation alongside the software license.

6. How will you measure whether it works?

Define the success metric before the project starts. Quote turnaround drops from four days to one. Scrap rate on aluminum jobs falls below 2%. Setup time variance between operators decreases by 30%. First-pass yield on complex geometry improves from 82% to 94%.

The metric should be something you already measure or can start measuring immediately. If you cannot measure the outcome, you cannot tell whether the investment worked. You will be left with feelings and opinions instead of evidence, and feelings do not survive the next budget cycle.

More on measuring technology outcomes in manufacturing: How to Measure Whether Your AI Tool Is Actually Working.

7. What happens if this fails?

Every technology project carries risk. The vendor goes under. The integration proves more complex than estimated. The tool works perfectly and the team refuses to use it. Understanding the failure modes before you start lets you build contingencies.

The shops that handle technology adoption well tend to start with a pilot. One process. One team. One measurable outcome. If the pilot works, expand. If it does not, you have lost a small amount of time and budget instead of a large one. That approach requires patience, which is precisely why it works.

The Question Behind All Seven Questions

Every one of these questions circles the same underlying issue. Do you understand the problem well enough to know what a solution looks like? Technology vendors will happily sell you their answer. The work of defining the right question belongs to you.

For a deeper look at evaluating technology for manufacturing operations, see our complete guide to AI in manufacturing.

The shops that get this right build operational advantages that compound over years. The shops that skip the questions and buy the demo end up with expensive tools that nobody opens after the first quarter.