The organization does not have to be huge to have a need for scaled agile decision making.
A while back (before SAFe existed) I was working in a small organization (30 people, 3 Scrum teams, 2 products). We were doing Scrum and XP and one of the founders was CEO and Product Owner for one of the products and the other founder was Product Owner for the other product. The Product Owners had the full mandate to decide over their product (of course after lot’s of discussions with customers, support staff, management, developers, the other Product Owner, sales etc). We did Scrum of Scrums a couple of times but realized we didn’t really need it, the teams were collaborating as needed without it. All was peaceful in agile land.
Then it was decided to really push and gain market shares and increase the pace. Investors and a new CEO was brought in to help make this happen. After a while it turned out that neither the CEO nor the new investors were completely happy with “I am PO, I am making all product decisions. I will listen to you and consider your opinions, but in the end it is my call”. With the pure Scrum approach both CEO and investors became stakeholders rather than decision makers. We couldn’t really blame the PO, because this was how we had designed the role. After a while we worked out some new structures. We introduced a Program Backlog with Epics that was continuously prioritized in the management team. The PO could still decide what to do in each sprint as long as the order in which Epics were delivered was aligned with the Program Backlog (normally there were no dates in the Program Backlog- we avoided dates unless they were absolutely necessary). The board of directors (representing the owners) made decisions on which markets to enter and in what order (the markets had different needs so this had an impact on the Program Backlog) and other strategic decisions. For example if we were aiming for UK before Germany, how much we should invest to enter UK and so on. To me it makes sense that the people putting their money at risk and that are majority shareholders should have an opportunity to make decisions based on risk levels, size of opportunity and so on. Some markets were obviously far smaller risk and investment – usually together with less opportunity and less visibility in the world market.
Basically what we did was to introduce a decision structure were strategical product decisions were taken by the board of directors, tactical decisions in the management team (that included the Product Owners) and operational decisions by the Product Owners. This means we had centralized decision making on two of the levels, much like the SAFe framework. The board of directors worked in a similar way to Program Portfolio Management using Investment Themes. The management team worked in a similar way as the Product Management team with a Program Backlog. The roles were different, but this is one of my experiences that makes me comfortable with some centralized control rather than fully autonomous teams – when necessary. As you saw in the introduction we didn’t use it in the beginning – and we didn’t need it. The need arose through a new context, rather than through size increase.
This meant that we moved away from the Scrum Product Owner should have complete and solitary ROI responsibility over the product. And I am completely fine with that – it made sense then and it still makes sense.