Economic evolutions are ugly, chaotic but often still result in a fruitful end. Almost no one completely understands how these economic evolutions occur. Theories proposed by Schumpeter and Marshall have their own ideas but don't describe the phenomenon in its entirety. There is plenty of more room to discuss this issue.
A journal article in Jahrbucher fur Nationalokonomie und Statistik discusses evolutionary economics according to Schumpeter and Marshall (Rahmeyer, 2013). Schumpeter believed that newly established firms were agents of change while Marshall believed innovation is the side effect of manufacturing and division of labor. Both have some legitimacy in their arguments.
Along with Schumpeter it isn't hard to see that new firms, which start by new ideas, can become catalysts for market change when their ideas work and are rewarded on the market. Unsuccessful firms die and successful one's thrive and eventually are copied by other firms seeking similar advantages.
Considering that evolutions changes our economic system as we know as the continuous development of new ideas leads to a continuously adjusting society. If we consider increasing specialized knowledge through research and application of best practices we can consider the division of labor an important point.
What both theorists don't research in specific detail is the underlining system that allows firms to develop and people to become more knowledgeable. This is a result of infrastructure and social interactions that share knowledge about ideas and then use those ideas for innovative development. The capacity of the local infrastructure, costs, government, culture, and environment will determine if it is possible. Therefore, the entrepreneur is one important component but they live within an environment that will determine the likelihood of those entrepreneurs to be successful.
Additional research needs to be conducted on the environment of research.
Ranmeyer, F. (2013). Schumpeter, Marshall, and Neo-Schumpeterian Evolutionary Economics: A Critical Stocktaking. Jahrbucher fur Nationalokonomie und Statistik, 233 (1).