Find influential (dependent) variable: Analyse which action of a cook was more important

#1
I have the following experimental design and two research questions. I am searching for help how to answer question 2

Experimental design
Person A prepared during the baseline every day 3 different meals that vary in CO2-content. he recorded how often he sold each meal. the baseline lasted 40 days

During "Phase2" person A changed the recipe of the meals to lower the CO2 per meal and tried at the same time to convince the customers to choose more often meal 1. Again the same type of data was recorded

Data:
Phase - day - meal - CO2 - sold
baseline - Day 1 - meal 1 - 200g - 50
baseline - Day 1 - meal 2 - 400g - 50
baseline - Day 1 - meal 3 - 500g - 30
baseline - Day 2 - meal 1 - 100g - 45
baseline - Day 2 - meal 2 - 200g - 55
baseline - Day 2 - meal 3 - 400g - 28
...
Phase2- Day 1 - meal 1 - 100g - 60
Phase2 - Day 1 - meal 2 - 300g - 20
Phase2 - Day 1 - meal 3 - 300g - 20
....

Research question
I am interested in the following questions:
1) did the average CO2 per meal drop in phase 2? (did the cook manage to lower the average co2 emissions per meal?)
2) if it dropped: which action had the higher impact: that the the cook reduced the co2 of the meal, or that more people chose the co2-friendly meal?

Approach for question 1)
I calculated the average co2 per meal, weighted with how often they were sold. i first calculated the daily average, and then the average over the whole phase by averaging over all days. And then i compared the average co2 per meal during the baseline and during phase 2.

For example,
for basline day 1 this would be (200*50+400*50+500*30)/130=346g/meal
for baseline day 2 this would be (100*40+200*55+400*28)/128=204g/meal
...
average baseline = 250g /meal (based on data for 40days)

for phase2 day 1 this would be (100*60+300*20+300*20)=180
...
average phase 2 = 170g/meal (based on data for 40days)

How to address question 2?
The cook reduced the co2 emissions from 250g/meal (baseline) to 170g/meal (phase2). But which action was more effective: that he reduced the CO2 of his meals, or that i made the consumers choose the meal with the lower CO2?

i started answering this question with scenario analysis, but it would be nice if there is a clean statistical approach to solve this question.