Chinese has no alphabet, and a Sanskrit word entering a Chinese Buddhist text either had to be translated or had to be written with characters chosen for their sound. Translators did both, in enormous quantity, and the transcription system that resulted is one of the largest bodies of phonetic writing ever produced. It is also the reason the Siddham letters had to travel alongside it.
Transcription loses information in a specific way. Chinese had no equivalent for several Sanskrit distinctions, notably the retroflex series and the voiced aspirates, so different Sanskrit sounds were written with the same character and the difference is simply gone. A reader working back from the characters cannot recover it. Worse, the sound values of the characters themselves shifted over the centuries and differed by region, so a transcription made in one place and period is read differently in another.
That is the practical argument for keeping the script. A dharani transcribed into characters can be chanted, and it was, but it cannot be reconstructed; the Siddham original can be read directly. The Chinese Buddhist canon accordingly carries both, and the transcriptions are now a major source for the historical pronunciation of Chinese itself, since a Sanskrit word of known sound written in characters tells a linguist what those characters sounded like at the time.