It is a rubbish article with rubbish conclusions.
Over the time of the study, there were no changes to paper tickets. If we take the statements made in the article at face value, then if a user decides to use Opal over an existing ticket in the limited number of situations where Opal might actually be more expensive for exactly the same travel, then presumably the user has done that because the card offers some additional benefit to them - perhaps it is more convenience.
Voluntarily getting more money out of people in exchange for additional convenience or whatever for a public service that needs all the revenue it can lay its hands on would be a brilliant success story. That's not how it is portrayed.
However, there are some weasel words in the article - the baseline period mentions spending on rail tickets only, but the April/May figures talk about spending on "transport". They are not the same things - so why are they being compared? Transport is more general than rail tickets - you would expect
spending on transport to be higher than spending on rail alone. The multi-mode applicability of Opal comes into play here as well - someone spending money on Opal may be catching more than the train. I doubt the authors of the study have access to be able to tell the difference.
Finally, as matthewg mentions, there is no way of knowing that the users have maintained the same travelling habits. Again, if their use of public transport actually increased in association with the adoption of the card, then that is another brilliant success story.
I suspect the study is just complete noise. Beyond the multi-mode issue and beyond the change in travel habits issue, the authors of the study sell an app that analyses your bank account transactions. Prior to opal a purchase of a rail ticket or other fare is more than likely to have been done by feeding cash into a machine. Chances are that the app would be oblivious to that sort of spending, unless the user diligently tracked their cash spending. With Opal, the top-up is more than likely done by a direct bank transaction (one that is automatically triggered by card balance) that the app can see and can easily track. That aspect alone could cause a massive difference in reported spending habits.
The fact that the authors of the study and/or the journalist don't mention this massive potential for statistical bias (or any of the other factors mentioned above) prior to drawing their conclusions in the headlines and by-lines says volumes about their lack of competence.