ROI of Machine Vision, How to Build a Business Case Your Board Will Approve

A machine vision project is rarely turned down because the technology does not work. It is turned down because the business case was too thin. A board does not approve better inspection or higher quality in the abstract. It approves a return on an investment. If you want a yes, you have to speak that language. That means putting a credible number on what the system saves and how fast it pays for itself.

The good news is that the numbers are usually there. Machine vision pays back in four main ways. Most strong cases are built from some combination of them. Here is the framework we use, followed by how to turn it into a one-page case a board can approve, with a calculator to do the arithmetic for you.

The four sources of return

Work through each of these for your own line. Put a conservative number on the ones that apply and leave out the ones that do not.

Defect and scrap reduction

This is usually the biggest return and the easiest to prove. Every defective part costs you money, in scrapped material, rework, returns and warranty claims. A vision system that catches defects earlier and more consistently than a tired human inspector cuts that cost directly. The maths is simple. Take the parts you make in a year, the share defective now, the share you would expect with automated inspection and the cost of each defective part. The difference is your annual saving. As an illustration, 1,000,000 parts a year at a 2 percent defect rate and £3 a defect is £60,000 a year. Halve the rate and you save £30,000 a year, every year.

Throughput improvement

Inspection often sets the pace of a line, especially where people are doing the checking. Automated vision inspects every part at line speed without tiring, which lets you run faster or run more consistently with fewer stoppages. The extra good parts you ship have a value. Multiply the additional units a year by the margin you make on each one and you have the throughput benefit. On a busy line this is often larger than people expect, because even a small percentage uplift compounds across every shift of the year.

Labour savings and redeployment

A vision system can take over repetitive manual inspection. The saving is real. How you frame it matters. The strongest cases rarely lead with cutting jobs. They lead with two things. Freeing skilled people from dull, error-prone checking and moving them onto work that needs judgement. And the fact that a camera does not get tired, distracted or inconsistent on a Friday afternoon. Put a number on the inspection hours the system replaces and cost them at the fully loaded rate. That is your labour benefit. Count it whether those hours become redeployed people or simply capacity you did not have to hire.

Quality and compliance

The last source is the one that does not show up every month but can dwarf the others when it does. A single recall, a rejected shipment, a lost customer or a failed audit can cost more than the system many times over. If your sector carries traceability or verification requirements, vision is often what lets you meet them and keep the contract at all. This return is harder to put an exact figure on. Treat it as risk avoided. Estimate the cost of one failure and how much less likely the system makes it. Carry that as a conservative line in the case rather than the headline.

Turning the four into a business case

Add up the annual benefit from the categories that apply to you. Set it against the cost of the system, both the one-off capital cost and the running cost of support and maintenance. Two numbers fall out. They are the two a board wants. Payback period is the system cost divided by the net monthly benefit, usually expressed in months. Return on investment is the net annual benefit measured against the cost. For most sound machine vision projects the payback lands somewhere between a few months and about two years, which is comfortably inside what most boards will approve.

Making it board-ready

A few things separate a case that gets approved from one that gets sent back. Be conservative with every assumption, because a board will trust a modest number it can defend far more than an optimistic one it cannot. Keep it to a page. Lead with the payback and the return. Put the detail behind them. State the cost of doing nothing, the defects, the lost throughput and the risk you carry by leaving things as they are, because that is the real alternative to approving the spend. And show that you have taken the risk out of the project, ideally with a feasibility study or a proof of concept on your own parts. Then the board is approving something that has already been shown to work.

Run your own numbers

The framework is only useful with your figures in it. That is what our online ROI calculator is for. Put in your own defect rates, volumes, margins and costs and it works out the annual benefit, the payback and the return for you in a couple of minutes. Start there to see whether the numbers stack up, then build the detail into your own one-page case for the board.

Try the machine vision ROI calculator

The fastest way to see whether vision stacks up for your line is to put your own numbers in. Our new ROI calculator does it in a couple of minutes and gives you the payback and the return to take into the room. If you would like a hand pressure-testing the figures, or a feasibility study to firm them up, talk to us.

Try the machine vision ROI calculator and see whether the numbers work for your line.

Get in touch at info@clearview-imaging.com or call +44 (0)1844 217270

Related reading    ROI calculator   |   Engineering services   |   Contact