There are some great observations about data in Why the Pandemic Experts Failed that apply in any context.

I won’t spoil the list, other than to say that I strongly agree with the first one: All data are created; data never simply exist. We rarely put enough thought or effort into planning how data will be generated, and then have to make up for this in the modelling phase.

Our models are always dependent on the quality of data we put into them. And, yet, we often spend much more time refining and testing our models than we do with our input data and the production process that generates them.