Respectfully: you did a trial where you were looking for one thing, failed to find it, but found something else. You need to do a trial where you define in advance what "success" looks like: what range of objects, within what range of distance, you would regard as evidence that your dowsing had found something. Then you need a control: you need to dig some random holes in random sandpits and find out how often one of the objects from your pre-defined list cropped up within the pre-determined distance. Finding an object isn't evidence that dowsing works; finding an object more often than random might be. You then need to repeat your trial enough times to generate statistical significance at a pre-declared level. And you need to declare in advance the trials you are going to do, to guard against cherry picking the ones that give positive results.
I'd love for there to be something in dowsing, for multiple reasons. Likewise homoeopathy and reflexology and fortune telling. Show me good evidence and I'd believe you like a shot.