In much of private sector development, the selection of a particular economic sector or sub sector happens quite early in the development process. Typically a sub sector is selected either for its potential in terms of job creation, exports or some other criteria. These indicators are sometimes informed by some preliminary research, other times by careful statistical analysis.
Approaches such as market development (a.k.a Making markets work for the poor) often assume that if a market that is important to a particular sub sector can be improved, then the society (and of course the sub-sector) will benefit. This logic ignores that a market is deeply embedded in a social context, what Mark Granovetter called social embeddedness. Furthermore, market failures tend to be interconnected as markets are interdependent. Thus the chances that a market is failing for a particular good or services in isolation of many other markets is slim. Market development is an evolutionary process that requires the co evolution of product and service markets as well as market supporting organizations and systems (markets are about far more than supply and demand, unless you are a traditional economist). This evolution is enabled or hampered by the presence of social institutions like trust, habits, value as well as formal and informal organizations that provides market guidance, regulations that reduce transaction cost (for instance by providing quality assurance).
The chances that a particular sub sector and its supporting markets, say for instance dairy farming with its small farmer network that is connected to a retail market, will evolve as an exemplary island in a sea of chaos (as in disorder, randomness, inconsistent behavior, low trust and fragmented efforts to improve the system) is highly unlikely.
In these instances, a sectoral or sub-sectoral approach is futile. Of course, it could be argued that working with a particular sub sector reveals something about the society, an argument that we cannot disagree with. However, most development programmes are measured at the micro level (in other words at the level of the actors in a particular sub sector) rather than on the improvement of the overall system conditions which includes adequate or responsive domestic market supporting organizations and supporting social institutions.
If you really want to support faster change in a developing country, perhaps a more promising place to start is with the actors that are behaving differently from the rest. This is a very qualitative or subjective assessment, but with this group of outliers it may just be possible to prove that a different kind of behavior is profitable (or effective), and that the boundaries created by culture, habits and routines may need to be adjusted.