Assuming spherical customers in a vacuum...
McDonalds serves around 68 million customers per day (https://findanyanswer.com/how-many-customers-does-mcdonalds-...)
in around 39 000 restaurants (https://www.statista.com/statistics/219454/mcdonalds-restaur...)
Of those restaurants, around 9% have broken ice cream machines (https://as.nyu.edu/content/dam/nyu-as/psychology/documents/c...), meaning there are around 61.88 million customers who had the opportunity to order a McFlurry.
I wasn't able to find good data on what percentage of customers order a McFlurry, but the linked article says they make 60% of dessert sales. I'm going to make a wild guess at this point that 50% of customers order a dessert of some description.
That means about 34 million people ordered a dessert, about 20 million McFlurries and 10 million other desserts.
Of those people, 3 million were not able to order a McFlurry because the machine was broken, so let's remove those 3 million from the total.
31 million people, 20 million McFlurries.
Assumption then would be that on average, around 1/3 of customers will order a McFlurry in a random restaurant (50% x 66%).
68 million customers across 39 000 restaurants is around 1 700 customers per restaurant per day, of which around 570 will order a McFlurry.
A regular McFlurry costs $2.39 (https://www.fastfoodmenuprices.com/mcdonalds-prices/), so a restaurant takes around $1 350 from selling McFlurries.
Taylor have a helpful calculator on their website (https://taylornewengland.com/sales/soft-serve-ice-cream-prof...), with likely optimistic estimates and don't show any wastage, but we're ballparking anyway. They say that food cost + napkin + cone comes to around $0.29, meaning daily profit from ice cream would be around $1 200.
Conclusion, it's pretty unlikely the difference is on the 'orders of magnitude' scale.
As a side note, given that the cost of a new machine is around $18 000 (https://www.wired.com/story/they-hacked-mcdonalds-ice-cream-...), after 15 days of the machine being out of order, the cost in lost sales is as much as buying a new machine (lower by whatever the Taylor callout fee is).
Given that 9% of 365 is around 33 (more than double the cost of a new machine), it seems that the more cost-effective solution would probably be to have a second, standby machine, from which McFlurries could be served while waiting for the callout.