There's a claim that 15 cinemas were seeing good turnouts – I overlooked this in my average calculation because I don't know the size or location of these screens or how many tickets they need to sell for a film to be 6th of 15 screens. It's possible that the, say, top 3 films will account for a disproportionate amount of the sales. From a personal experience – this would be like when I tried to see Star Wars VII and it was sold out and I saw a different film on the 2nd screen which had only ~20 people in it, the number 1 film was outperforming the 2nd performing film by a ratio of about 1:100 (Screen 1 is big!).
However, if we say that these screens were getting 100 people a day (a little over triple my average figure calculated above), that actually makes little difference to the average visitors per screen at all cinemas.
$1,235,000 USD USD at $9 USD/ticket ~= 137200 tickets.
Assuming 15 cinemas had 100 ticket sales per day, that accounts for 15 x 7 x 100 = 10500
Remaining cinemas = 668-15 = 653
Therefore average daily ticket sales at the other cinemas = (137200 – 10500) / (653 x 7)
= 27.7 visitors per day.
Now if you say "no, no, those multiplexes were reporting 6th of 15 screens and 9th of 30 screens, they must have had a lot more than 100 people a day" then, ok, let's try a ten fold increase.
15 screens have 1000 people per day. This leaves only 32200 visitors for the other 653 screens.
Which will be 32200/(653*7) = 7 ticket sales per cinema per day.
Now – that also can't be right, or all the other 653 cinemas would have dropped the film.
So although I don't know exactly what the visitor rate at the well performing multiplexes was I can postulate that it was of the order of 100-200 ticket sales per day at best, as larger numbers destroy the averages at the other cinemas to the point that it is no longer credible that they would continue to show the film.
I then postulate that 700-1400 people could still be reasonably judged to represent a high percentage of the likely audience for the locale (note – we none of us actually know where these high performing multiplexes were) and a business decision could justifiably be made to replace this film with a new film.