For each species there is a species-specific minimum effective concentration of colchicine for polyploid induction. In plants that minimum concentration is usually somewhere below 0.01%.
In seed treatment it is typically used somewhere in the range of 0.01-0.1%.
If a concentration too high above the minimum effective threshold is used, the treated plants can be stunted and lack vigor. For instance in a clover species that was made polyploid 0.01% worked better than 0.025%, both having been treated for 8 hours. However, some researchers find that going overboard on concentration- while it kills many and stunts the survivors- can be more effective in making the survivors polyploid.
The fewer cells, the better. If you treat one cell, the resultant effect will be homogenous throughout the plant. If you treat a clump of cells, particularly ones in various stages of division, you'll get a 'mixoploid'- a sort of chimera. Because of this, many people treat seeds. However, many researchers had quite good success by treating seeds just as they germinated.
Colchicine is rapidly destroyed by sunlight.
Some survivors will be diploid.
In early work, staining and counting chromosomes was a bitch. As such, and because some treated plants will be diploid, researchers hoped that cell size was altered. By doubling or quadrupling the nuclear DNA compliment, the cell nucleus is enlarged. This often (but not always) leads to enlarged whole cells. If researchers identified variance in size on the treated plants versus controls they would test all treated plants first by staining and measuring cells in pollen or a leaf of a given size, then they'd do the chromosome count tests on subjects with enlarged cells. Comparing pollen cell size at home would be significantly easier than counting chromosomes at home.
If your target species is self-incompatible, grow out as many apparent successes as you can to increase chances of identifying successful breeding pairs.
@XDX Polyploidy is often used to make bigger, prettier, or more vigorous plants. Most ornamental roses and many morning glories you see are polyploids. Its also how things like seedless watermelons and grapes are made (they are triploids fertilized by diploids).
There are other uses too, including potential in improving low-genetic diversity things like S. divinorum or saffron.