I support open access and would like to see everyone getting access to scientific articles. However, we as scholars have also become irresponsible and mostly thinking about ourselves in this case. Here the question is not: Are commercial publishers making a lot of profit, but why are we as scholars producing content which is read so few number of times that a commercial publisher has to dictate the scholarly community? If we were writing good content which is read by millions of people, we would not have required the commercial publishers (e.g. Today a musician does not require a music label. If their music is good, it can be released on Youtube).
Profit margins are necessary to run journals effectively. We do not want a situation that a journal shut downs as it becomes financially unviable to run it. Could you yourself run a business at 5 % profit margin when the product you are selling has limited commercial value?
A scientific article has limited value for the general public except for people interested in the topic and the academics working in the specific field. A scientific article is sometimes a tribute to the old scholars working in the field i.e. citations to their work and a way to popularise the concepts developed by a niche academic community. There are so many citations to academic friends and self citations that one must question: How would the general public benefit from this content?
Why would be millions of people be interested in reading scientific articles and why would a distributor of the product (i.e. journal) take the risk to promote it when they know that the articles will not be read by more than 2000 people even in 10 years?
The scholarly publishing industry has an average profit margin of 30-40 %. Profit margins differ from industry to industry. Some industries even have negative profit margins but it does not create an alarming situation. Operating at 10 % profit margin for a scholarly publisher creates financial risks in the long run and limited funds available for spending on upgradatian of digital infrastructure and investing in new technologies for journals. Infact, the costs are quite high now than the old model when journals were published in paper. Digital does not mean free. Software engineers who run and maintain the journal digital infrastructure (for 1000's of submissions every year as the number of scholars are increasing) require very high salaries due to their rare skills. There are massive costs for running a journal and profits are required to continuously improve the customer experience.