Imagine for a moment you are loading a machine. Loading the machine requires you to take a piece of paper off the tray holding the parts and write the information on this piece of paper on another piece of paper, which is then retained as a record of the current process cycle. The paper is then placed in a temporary storage location so it can be mated with the proper tray once the process is complete. Performing this job requires just over fourteen minutes. Seems simple enough, what is the harm in doing it this way? You'd be correct, it is simple and while 99 times out of 100 your team does it correctly that 1 time out of 100 causes a lot of lost time and confusion at the end of the process.

How do we improve this? First we identify the pain points.

• It takes approximately one quarter of an hour to perform this function. (Productivity)
• Data being transfered could be copied incorrectly (Quality)
• Papers aren't properly mated after the process ends (Quality)
• Errors occur due to rushed or poor handwriting (Quality)

It goes without saying that errors in quality are also productivity hits. We can make this process better. Safety isn't really a factor here. So next on the priority list is quality. We will focus on that.

The two major problems that can arise in quality are due to separation of the paper from the tray, and human error in copying or reading the data that is recorded. Let's try to eliminate that, but remember we need this intermediate paper as a record of cycle for the machine.

First Automation Steps:

1. Gather papers from trays to be loaded.
2. Utilize barcode scanner to obtain key data from papers on the tray.
3. Collect any data not barcoded such as tray number, number of parts in tray, etc.
4. Store this information to be printed after all trays have been scanned.
5. Display information to operator for verification.
6. Print paper run form for process machine.
7. Place tray papers at end of process machine.
8. Unload machine and match up papers.

This removes human writing from the equation and reduces the number of quality incidents due to illegible handwriting. Additionally, it reduces the chance that a number is copied incorrectly, because we are pulling data directly from the sheet using barcodes. Is it perfect? Nope, but it's a start.

This step obviously requires some custom programming to be performed. A simple program would be created that presents a form to the user to collect data about the tray. It would then store this data an display it later for verification purposes. An added benefit not already discussed is this data can now be transfered to a central database to provide an electronic record of a batch being ran. Sometimes this eliminates the need for paper alltogether.

Sounds good, but how does this help your bottom line? By reducing quality errors, we increase first time yield. Therefore, your team spends less time reworking parts; instead they have time to run more parts. We also found that by scanning barcodes, and only typing minimal data into the program we could cut the time it took to perform the paperwork task, in half. What did take approximatly fourteen minutes per run, now only took seven minutes.

Isn't there still room for errors, and what about all that wasted time moving papers from trays to the end of the machine? Yes, there is always a possiblity for errors. However, this method provides a printed map that allows the tray's to be matched back up with their papers with far greater accuracy than was previously possible. And, yes, we are still wasting time moving paperwork. You will, almost always, find that problems are solved by an iterative process. This means that the large problem is broken down into smaller more managable problems and these are the problems that are solved. We then re-analyze the problem, and make more improvements.

Lastly, let's take a quick look at how much potential money this project saved you throughout the course of 1 year. For simplicity sake, we will only count the seven minutes that were saved by the new process. Other assumptions are that all quality issues related to the process fixed, the operators performing this takes are making $10/hour, and that we are running 24 hours a day for 261 days a year.

Originally, it cost 0.233 * 10 * 24 * 261 = ~$15,000/year.
Utilizing the improvements, it cost 0.117 * 10 * 24 * 261 = ~$7,500/year.
This means that within one year you have probably paid for the program and put a little more to the bottom line. Subsequent years go straight to the bottom line!