The returns to sales effort in B2B selling : prospecting for opportunities and the value of behavioral information
Abstract/Contents
- Abstract
- This dissertation studies the effect of sales effort on sales productivity in the B2B context. I study the returns to sales effort from a firm's salespeople, measured in terms of sales opportunities, and the value that behavioral information generates through its influence on salespeople's sales efforts. Firms commonly seek to improve the productivity of their salespeople (i.e., the number of sales or the number of sales opportunities generated by a salesperson) through policies (e.g., how salespeople are compensated, or what information is shared with their salespeople) that affect the amount and quality of sales effort supplied by their salespeople. These policies are costly for the firm to adopt, and their returns are unclear. For example, adopting more generous compensation may incentivize salespeople to expend more effort but also directly increases the cost of each sale. Similarly, sharing more behavioral information may encourage more sales effort or the targeting of sales efforts toward better prospects, but acquiring this behavioral information requires costly investments in marketing. Since these policies affect productivity through their effect on effort, the returns to these policies are difficult to evaluate without an estimate of the causal effect of sales effort on sales productivity. However, the causal effect of sales effort is difficult to estimate from observational data. Unobserved factors which make sales more or less likely may simultaneously affect a salesperson's decision to contact a prospect, which causes sales effort to be an endogenous regressor whose effect cannot be estimated without bias. Estimates of this causal effect are rare in the prior literature and, to my knowledge, non-existent in the context of B2B sales. My dissertation has two main contributions: first, I use a novel dataset of B2B selling to estimate the causal effect of sales effort through an instrumental variables approach and, secondly, I use this causal estimate to empirically evaluate the value of three information disclosure policies which vary the amount of information shared by the firm with their salespeople. To accomplish this, I obtain granular archival data from a US technology firm that records one year of selling activity by their salespeople towards prospective customers ('prospects'). I construct a novel panel containing 1,481,899 prospect-week observations (corresponding to 30,841 unique prospects, 12 months of selling and 45 salespeople). I can observe salespeople's efforts (306,019 emails and calls) and productivity (1,272 opportunities created) in this panel. I also observe each salesperson's territory and when they are informed about three types of behavioral information (i.e., website visit notifications, marketing leads and responses from prospects). This additional data allows me to construct a set of instruments to estimate the causal effect of sales effort. As salespeople are effort-constrained and behavioral information signals a prospect's interest in the product, the amount of behavioral information received by a salesperson from prospects in their territory in a given time period can shift the probability of an opportunity with a prospect by changing how the salesperson allocates effort. This forms the foundation of a set of instruments that are relevant and plausibly satisfy the exclusion restriction. I find that, on average, if a salesperson chooses to initiate contact with a prospect in a given week, it increases the probability of finding an opportunity with that prospect by 1.39%. I find that there is heterogeneity in the returns to sales effort. The return to sales effort is largest for prospects considered the least likely to become opportunities (3.41%), smaller for prospects that are ex-ante considered better prospects (1.58-2.34%), and the return is not statistically distinguishable from zero (at conventional levels) for the ex-ante best prospects. Using these estimates, I evaluate how the firm's choice to disclose (or censor) behavioral information from their salespeople affects sales productivity. Firms make costly investments in marketing efforts and technologies to generate and share behavioral information, and it has been a long-standing empirical question whether sharing behavioral information like marketing leads is valuable to salespeople. It is also unclear how much new types of information, like notifications triggered by prospects visiting the seller's website, improve sales productivity. In my data, the firm shares both leads and notifications with their salespeople. One way of quantifying the value of this behavioral information is to consider counterfactual worlds where salespeople are prevented from seeing certain types of behavioral information. By comparing the observed information disclosure policy (full disclosure) with a counterfactual world where some types of information are censored (e.g., no notifications), I can evaluate the change in expected effort allocations and the value of an information disclosure policy in terms of lost opportunities. I find that, compared to the status quo where the firm discloses all behavioral information to their salespeople, censoring notifications would have caused +6.5% opportunities created over a year (censoring marketing leads results in -0.8% opportunities, and censoring both results in +5.5% opportunities). Surprisingly, this indicates that the firm would be better off censoring notifications rather than disclosing them. This is particularly counter-intuitive because my estimates indicate that notifications about a prospect, on average, increase effort towards a prospect. Increased effort should increase the probability of finding an opportunity with the prospect and lead to, in expectation, more total opportunities created. A more detailed investigation, however, shows that while censoring notifications decreased the probability of finding opportunities with prospects that generated notifications, it also increased the total amount of effort (especially among prospects that did not generate notifications). This suggests that large volumes of behavioral information may trigger information processing costs among salespeople that, in aggregate, lead to worse outcomes than not providing any behavioral information at all. One way to balance the beneficial and detrimental effects of disclosing notifications would be to disclose behavioral information selectively. To facilitate the future evaluation of more complex counterfactuals, I introduce a novel theoretical model of prospect sales response, prospect information-gathering and salesperson effort allocation that could be used to simulate and evaluate targeted information disclosure policies.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2022; ©2022 |
| Publication date | 2022; 2022 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Zhang, Charles |
|---|---|
| Degree supervisor | Sahni, Navdeep |
| Thesis advisor | Sahni, Navdeep |
| Thesis advisor | Lattin, James M |
| Thesis advisor | Nair, Harikesh S. (Harikesh Sasikumar), 1976- |
| Thesis advisor | Reiss, Peter C. (Peter Clemens) |
| Degree committee member | Lattin, James M |
| Degree committee member | Nair, Harikesh S. (Harikesh Sasikumar), 1976- |
| Degree committee member | Reiss, Peter C. (Peter Clemens) |
| Associated with | Stanford University, Graduate School of Business |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Charles Yu Zhang. |
|---|---|
| Note | Submitted to the Graduate School of Business. |
| Thesis | Thesis Ph.D. Stanford University 2022. |
| Location | https://purl.stanford.edu/qj769kf6397 |
Access conditions
- Copyright
- © 2022 by Charles Zhang
Also listed in
Loading usage metrics...