It will be best if you have drop data from your own distribution chain. But another approach can be using data from others (see below). If you are using data from other sources please also introduce a control on the quality of these data. If you experience too many damages then you have to increase drop heights or shocks slowly until the problem stops. If you see less or none damages from the market your test can result in over packaging. Then you have to reduce drop heights or shocks.
In case of no practical data you can use: