Monitoring LUT Accuracy on Set, A Mini vs. C70 Comparison

I recently tried a new approach to ensure our on-set LUTs translated accurately to post, specifically with differing camera gamma spaces. We were shooting a narrative short with an ALEXA Mini as A-cam and a Canon C70 as B-cam, both delivering ProRes internally. My usual method is to load the CDL/LUT onto the SmallHD monitors and trust it, but this time, the DIT provided a more aggressive look-LUT that felt a bit crushed on the ALEXA Mini.

What I tried was using an Atomos Ninja V on the C70, which allowed us to apply the same CDL/LUT file directly to its monitoring output, rather than relying on the camera's internal LUT processing which can sometimes interpret things differently. For the ALEXA Mini, we stuck with loading the LUT directly into the camera's SDI output.

What worked well was the Atomos's ability to preview the exact same Resolve-generated LUT on the C70's output without much fuss. We could visually confirm that the look was consistent, even if the underlying gamma curve of the C70 was distinct from the ALEXA Mini's LogC3. What didn't work as perfectly was the visual matching itself; despite using the same LUT, the monitors still showed a subtle difference in the deep shadows between the two cameras, likely due to their native sensor characteristics and not the LUT application. It prompted a lot of discussion with the DIT about baked-in looks for certain shots where we knew the C70 would match a specific exposure range.

My question is: beyond matching the specific monitoring LUT, what are your best practices for ensuring consistency in the final look when blending cameras with fundamentally different color science and gamma spaces, especially when a more aggressive look-LUT is desired from the outset?