From my experience, contractors generally struggle with utility trench backfill. Therefore, I recommend a maximum loose lift thickness of 8 inches. As far as testing frequency, I would say every lift and at least every 200 feet or every time the backfill material changes. Also, sand cone calibration is a good idea, however, if the same gauge is used everyday, one sand cone per 10 tests is fine even if it takes several days to record 10 tests. Number of Proctor tests, one for every material used (United Soil classification System, CL, ML, SC, SM, GC, or GM). A gradation test in conjuction with Atterberg Limits can be used to classify the soil. Also, color change does not mean the soil classification has changed. Hope this helps.