I would guess there is a difference between the effect of various payout percentages in online games vs physical machine games. I would assume online players tend to have more time to spend playing on average which gives them more time to hit their spending limit.

I would love to give a source for the 95% figure but I only have that info second-hand from an ex-employee.

Regardless of whether or not you believe that statement, you should see that the logic that lower payout ratio equals a lower payout amount simply doesn't hold. Imagine if the payout turned out to be 50% rather than 90%. Would you expect the company make more or less money? Even if players believe they were simply unlucky, that will still be enough for some of them to stop playing. The optimal payout in terms of revenue for the company is something that can only be determined by real world experiments and there is no evidence in the article about this case that the lottery company determined that an advertised payout of 90% and an actual payout of 87% by experimental analysis. Hence my tendency to believe this was not an act of deliberate deception.