Research = determining a candidate compound, e.g. a new antibiotic, (e.g. by taking and tweaking an existing compound, by basing off of the actual viral/bacterial genetic code such as with the Biontech/Moderna CoV vaccines or by isolating a compound found in a plant, fungus or other natural agent which has been determined to have the effect you want), and initial checking against cell cultures and/or lab animals for effectivity.

Development = taking a lot of candidate compounds and evaluating them in the classic three-stage procedure - phase 1: determine how the compound is processed in actual human bodies to check if it is actually safe to ingest and what side effects can already be observed (n ~ 20-80), phase 2: determine effective dosage (n ~ 100-500), phase 3: check effectivity in a double-blind trial (n ~ 1k-10k), as well as the fourth phase (after-license monitoring) [1][2].

The more a compound progresses, the more expensive the trials get to conduct, not just because the participants usually get some compensation for their risk, but also because all the data has to be tracked and processed. And a lot of candidate compounds fail somewhere along the path (either because they are ineffective or because the side effects are too severe), which makes the money invested effectively wasted (from a capitalist point of view, not from a scientific!).

[1]: https://en.wikipedia.org/wiki/Clinical_trial

[2]: https://studienteilnehmergesucht.de/ratgeber/die-vier-phasen...