Completely agree with Madam A. D-tor.
In metal detection there is always a product signal. That signal will vary on the water activity of your product, level of ionic compounds in there (so dissolved salt normally pushes up the signal). It is important to test a metal detector with the correct product and the correct metal test sticks. Why?
Imagine a situation where the dummy product you're using has a lower product signal than your real product. You set up your metal detector to be able to reject your spiked packs but the combination of the metal test stick with the dummy pack is actually a lower signal than metal in a real pack. That is likely to then lead to false rejects on your normal product. You may think that's not a problem but not all retailers then accept product being retested and also a large number of false rejects then is likely to lead to staff not checking those packs as carefully as they should increasing the risk of real metal being sent out.
Imagine the situation where the dummy product has a higher product signal than your real packs. Well then you're really screwed. Why? As the combined signal from the dummy pack and the metal stick will be higher than a real life situation where the metal stick is in product. That means a metal contaminant of the same size as your metal test sticks is likely to pass through your metal detector resulting in an increased risk of real metal being sent out.
Now you might be thinking "hey, I've checked a few packs, it's fine, it's good enough, I don't get false rejects and it still works if I put it on a real pack." Well if that is the case your metal detection test piece sizes are almost certainly too large and if you used real spiked packs, you would be able to reduce your test piece size or at least have a real awareness of what your test piece size should be.
I am astounded no retailer nor third party auditor has picked this up and as presumably your product has a long shelf life there really isn't any excuse as you can keep the product spiked packs for some time too so it would cost you very little if anything more than what you're doing.
If you've not had the challenge yet, I'd anticipate it and it may even be a major (once someone with their head screwed on visits). I would raise it at that level, possibly even critical if there is a significant product signal difference between the product and the dummy pack without the test stick.