Law Enforcement is a necessary part of our society and our economy. Good law enforcement improves the community and poor law enforcement creates problems. It is the same thing in any business that has a customer service (i.e. public engagement) aspect when deal with customers (i.e. society in general) We do know that the vast majority of law enforcement officials are good people who do the right thing in any given situation but there are "bad apples" that will cause these types of issues.
Thinking of this from a purely disciplinary labor relations issue it is important to have not only feedback loops but a review of employees who engage in misconduct. That is missing here and that is part of the problem. When there are enough complaints and other performance issues there should be an investigation and review that leads to removal. Without it people just continue on and on. It doesn't matter if its a company or a public service institution because it will almost always cause angst.
However, the stick approach doesn't create positive culture within organizations (i.e. there were officers standing around not saying anything). One must also think about rewarding good officers for their pro social and positive behavior. I have met them and these officers are an asset to their departments. Much is based in their personality and their desire to provide "justice" on a higher level. They go above and beyond to listen before acting and make wise decisions that protect the public in the short and long term. Such officers should be commended and rewarded for their behavior.
The metrics are also part of the problem. The ultimate goal of law enforcement is "justice" and therefore that requires an open mind. The amount of tickets, amount of arrests, and amount of people put behind bars really isn't a great metric. It may be necessary at times but it isn't the ultimate goal. Such metrics influence behaviors and in turn influence outcomes and how people interact with the community. Hence the metrics and what is rewarded influences the cultural outcomes (i.e. culture as a determinant of how people are expected to act in a group).
Furthermore, we may want to consider a few other things such as recruitment practices, how public perception impacts available candidates in recruitment, and the type of personalities we want to recruit into such departments and how we select law enforcement leadership. On the back end we could also consider better support services such as counseling for law enforcement officials who struggle with their own personal problems and how that makes its way into workplace choices (i.e. kneeling until death).
We know there is work to do here and how that is left to the "experts". The problem is the "experts" haven't really solved the problem even though it has been brewing for some time and one could expect this very public feedback loop (i.e. rioting). This means the experts were trained in a certain methodologies and continue to use the same methodologies when understanding and solving similar problems. Supporting law enforcement is about supporting the ability to do their jobs effectively and that means rethinking how things are done so as to be the most effective system possible. Doing the same thing over and over is going to look more insane in the future.
Finally, I feel that sometimes helping people is a more important metric than actual enforcement. Educating people is helpful, referring people to mental health and other services can save tax payers a lot of money down the road through not having to incarcerate (i.e. 40k to 50K per year), and when necessary arresting those who continue to menace society without consideration to their superficial differences. In essence, we should ask people outside current law enforcement training and experts on how to better protect officers and protect the public at the same time.