You might want to play around with different pixel functions than just max (mean, min, first/last image, etc). Depending what type of data your raster represents and what you are trying to do with it, one of these might be more appropriate than the max.
For example if you were working with a DEM, taking the max is going to exaggerate the height, so maybe taking the mean would be a better option.
In contrast, if working with categorical data (e.g. land use) taking the mean would not be appropriate, and taking the max just biases the results towards whichever category happens to have the highest category number. Instead you might want to choose the most common value if you have 3+ rasters overlapping.
If your data is imagery based, if the images are on different days perhaps taking the most recent pixel would be best (assuming you can order the rasters by date).
When combining rasters it is also important to check whether your input rasters have values all in the same range. For example with imagery taken on different days sometimes one image might have particularly high or low values, or a far smaller/larger range of values than the other images. One single extreme value might not be visible to the naked eye in your merged image, but it could definitely throw off the coloring making everything appear to be all roughly the same value. As an example, if a color ramp is stretched from 0 to 100 you'll be able to see the difference between 40 and 60 much better than if that same color ramp is stretched from 0 to 10,000.