As people have explained above, there are a number of possible causes.
- The powder is contaminated, but not homogeneously.
- The sample was contaminated in the lab
- You have a rogue result
Is it your own lab or external? How well do they work? Are they ISO17025? Was the second result their retest or from a resubmitted sample? My experience of working in a micro lab is that errors can occur. However, the usual approach is that the powder would be reconstituted using 10g powder and 90g of diluent (could be BPW or Sodium Citrate depending on your requirements). About 100g of the rest of the powder is then put aside for up to 1 week to allow retesting if unusual results are found. From the 100g reconstituted material, 1g is plated and incubated. Unfortunately, this does not always give representative results as the powder is unlikely to be fully mixed.
Plating from the retained sample will not necessarily give the same result as the sample is taken from a different set of powder, hence the different results.
Plating 2 samples will limit the effects of contamination in the lab and/or rogue results.
However, you can never be sure that there was no contamination as your original sample has been use up.
Isn't life fun!