Optimal Market Intelligence Strategy When Management Attention Is Scarce

This paper extends the theoretical literature on firms' optimal information strategies to the situation when a firm's management attention capacity to process available data is scarce. In this case, a firm's optimal market intelligence strategy must trade off learning a little about a broad range of markets (a broad strategy) with gaining a very deep understanding of one or a few markets (a focused strategy). This trade-off is not present when data are scarce, an assumption made in most of the existing literature on optimal information search strategies. However, in data-rich environments, which are of increasing relevance given technology changes, we show a focused market intelligence strategy is always best when managers need to process a substantial amount of data before beginning to gain insights; i.e., there are increasing returns to attention. Interestingly, this focused strategy can also be best with decreasing returns to attention when (a) managers are sufficiently efficient in processing the available data and (b) managers have sufficiently strong initial priors on the unknown market parameters. We show a broad market intelligence strategy is only optimal when new data points are sufficiently redundant, i.e., when the learning rate is sufficiently decreasing with the allocation of more attention. Our results also indicate that advances in information technology can account for the pressure on firms to become more focused and that competition increases the likelihood of a focused strategy. Competition can lead to asymmetric outcomes where firms focus on different markets. Finally, we note that a focused market intelligence strategy, and thus an asymmetric allocation of attention, does not require a priori differences between firms, markets, or market-specific core capabilities. Consequently, a focused market intelligence strategy can result in market-specific core competencies and produce firm differences from equivalent starting conditions.