I used to make colloidal gold for research purposes and it was definitely a tricky business. I don't know about silver, but there were a lot of poor methods out there for gold. The biggest issue with poor preparations was a wide dispersion of particle sizes--which usually caused problems with the downstream application. A good colloid is the size you want it to be with a fairly narrow spread in particle size. A centrifuge is necessary for concentrating and washing the colloid and a scanning spectrophotometer is used to assess the particle size and spread.