Are conservative approaches to new product selling a blessing in disguise?

Central to our study is the concept of conservatism, which finds its roots in the Greek word conservare, meaning “to keep,” “to preserve,” or “to retain.” Psychology, sociology, economics, and political science research presents conservatism in various ways, including as an individual behavior, a personality trait, an attitude, a business strategy, or a social/cultural norm (Jost et al. 2003; Wilson 2013). Individuals’ conservatism associates with avoidance of cognitive complexity, a lower willingness to deviate from social convention, and a desire for stable beliefs as opposed to uncertainty (Jost et al. 2003).

We focus on conservative behavior, which entails conducting known courses of action before engaging in new and unknown activities when making decisions under risk. A useful lens for studying such behavior is provided by literature on perceived risk processing, which describes how individuals perceive risk and consequently make behavioral decisions (Conchar et al. 2004). Risk reflects the extent to which there is uncertainty about realizing potentially significant and/or disappointing outcomes of decisions.

An important assumption in perceived risk processing literature is that human decision makers are risk averse. The more outcome uncertainty surrounding a behavioral choice, the more likely individuals prefer less uncertain options. Although most sales studies also assume salespersons to be risk averse, risk aversion is seldom operationalized. We provide a more detailed and practical perspective.

CSB represents a sales domain–specific behavioral representation of risk aversion and is defined as the extent to which a salesperson maximizes selling efforts for existing products before engaging in efforts to sell the new product. CSB describes the order in which salespeople sell products from their portfolio. Rather than making a one-time choice in selling an existing or a new product, salespeople may change their product preference (and accompanying pitch) along the sales process. CSB may even act as a deliberate strategy to reduce customer objections.

CSB shares conceptual territory with related concepts; Table 1 provides a comprehensive overview. Extant research on salesperson behavior during the sale of new products mainly focuses on effort or its variations. For instance, selling (or working) hard reflects the amount of time spent in trying to achieve sales goals (Rapp et al. 2006). New product adoption is an interaction of selling effort and commitment to the new product (Kuester et al. 2016). New product selling intentions precede effortful behavior and capture a salesperson’s willingness to try hard in selling new products (Fu et al. 2010). CSB differs from these concepts in that it specifically accounts for the temporal ordering of selling new and existing products. Salespeople may invest much effort in selling the new product, but only after they have explored the options to sell the existing product. CSB also accounts for the fact that salespeople do not need to be positive or negative about the new product. This is an implicit premise of studies on effort-related concepts though.

Adaptive (or smart) selling is another related concept. This reflects a salesperson’s capacity to plan and execute a wide range of selling behaviors and activities based on situational considerations (Sujan et al. 1994). Unfortunately, most studies on adaptive selling consider the skill of adaptation but do not focus on specific alterations in selling behaviors or activities. CSB specifically suggests that the order of new and existing product selling may be such an adaptation.

Finally, two concepts in the new product selling literature specifically account for the trade-offs or complementarities between new and existing product selling. First, product selling ambidexterity holds that selling new and existing products can be balanced over time through alteration of activities (Van der Borgh et al. 2015). Studies on ambidexterity do not discuss the order of selling that leads to this balance. Because salespeople have to decide in each encounter which product to present to customers first, CSB provides a more informative lens on salespeople’s trade-offs than product selling ambidexterity. Second, cross- and up-selling behaviors reflect selling additional items to customers who have previously purchased one or more item(s) (Kamakura 2008). Such sales behaviors are usually successful because salespeople have a foot-in-the-door with these customers. CSB may influence sales outcomes through similar principles but does not require a history of purchase to take effect.

Later, we continue to distinguish CSB from these concepts in our empirical analysis. Next, we build our conceptual model.

We build on perceived risk processing theory to derive CSB’s antecedents. The theory posits that individuals go through three phases when they have to make product choices that involve risk: risk assessment, risk framing, and risk evaluation. In the first phase individuals perceive five dimensions of risk that ultimately influence their product choice: performance risk (i.e., chance that product does not produce desired outcomes), social risk (i.e., chance that product affects the way others think of individual), financial risk (i.e., chance that product involves losing money), psychological risk (i.e., chance that product does not fit well with self-concept), and physical risk (i.e., chance that product causes health injury) (Jacoby and Kaplan 1972; Mitchell 1999). In contrast to the first three risk dimensions, psychological risk and physical risk may be salient for some products (e.g., popular brands, luxury goods, food or health products), but are absent for most other products (Labrecque et al. 2016).

In the risk framing phase, individuals “sort and filter informational cues that will enable them to handle or reduce perceived risk” (Conchar et al. 2004, p. 427). Such information processing determines the relative importance of each risk dimension in an individual’s choice process. Mass media, managers, friends, or researchers may (de)emphasize one or more risk dimensions in individuals’ decisions through activating particular connections or associations in their cognitions―a process called priming. For instance, Mandel (2003) shows that instructing individuals to think about family and friends makes social risk more salient than financial risk in behavioral decisions.

In the last phase, risk evaluation, individuals decide on whether to make a risky choice or not. Individuals “manage the consequences of perceived risk through a process of mental accounting […] that constitutes perceived-risk evaluation” (Conchar et al. 2004, p. 431). In general, the larger the perceived risk, the larger the threat to extant wealth and the less likely individuals make the risky choice. However, individuals also weigh the risk dimensions as potential losses against evaluation standards, specifically their initial asset levels (e.g., past investments) and trait-based personality characteristics such as self-confidence (Mitchell 1999; Wiseman and Gomez-Mejia 1998).

Although perceived risk processing theory originates in consumer research, several works extend the theory to an organizational context and outline managerially-relevant factors that make up individual’s perceived risk dimensions in organizational settings (e.g., Sitkin and Pablo 1992; Wiseman and Gomez-Mejia 1998). We build on these studies to define the elements in the perceived risk processing framework and employ a qualitative study to ground our concepts and hypothesized relationships. Specifically, we interviewed 32 employees from 15 high-tech companies.¹ All employees had a role in new product development and launch within their respective companies. Functions included salespeople and their managers, R&D managers, product engineers, and marketers.

We first consider the perceived risk dimensions that salespeople rely on during risk assessment: performance, social, and financial risks. In our study context, psychological and physical risk dimensions are less of a concern because the newly introduced products do not pose risk to a salesperson’s health or self-identity (cf. Labrecque et al. 2016).

A salesperson’s uncertainty to what extent effort spent in the sales process will result in closing deals is largely a function of the (un)familiarity of the product to the decision maker and other stakeholders (Sitkin and Pablo 1992). One of the fundamental challenges in new product selling is new product radicalness: the extent to which the new product is perceived to be inconsistent with the systems, needs, and norms already adopted by the business customer (Micheal et al. 2003). As an R&D manager from an automotive company said: “Our salespeople are really good in selling simple vehicles […] But when they have to explain an innovative vehicle with four independent axes and all configurational options, the pitch becomes more complex because the customer does not understand the product.” New product radicalness directly affects the chance that a salesperson can attain the desired performance outcomes and thus embodies the performance risk a salesperson perceives.

In addition, managers model risk behavior and lend their personal legitimacy to the taking or avoiding of risks of their subordinates (Sitkin and Pablo 1992) through their selling orientations. A sales manager from an original equipment manufacturer indicated: “In contrast to more transactional selling situations where sales priorities are communicated company-wide, our complex B2B selling environment requires that I assume an important role in guiding salespeople on how to proceed with the selling task.” We thus focus on managerial new product orientation, which reflects to what extent salespeople perceive managerial practices, actions, and directives that guide employees’ attention, time, and effort toward the sale of new products in the firm’s product portfolio.

Employees align their behavior with leader orientations to minimize potential risks while maximizing benefits with respect to pay, promotions, and job assignments (e.g., Detert and Burris 2007). Not surprisingly, a salesperson who acts in line with managerial preferences generally is evaluated favorably by his/her manager (Podsakoff and Mackenzie 1994). In contrast, a salesperson who does not follow the dominant selling orientation generally is evaluated critically by his/her manager. The stronger a manager’s new product selling orientation, the more clearly employees perceive a strategic prioritization of selling goals, and the more convinced they become that deviating from working toward these goals changes the way their manager thinks about and evaluates them. We thus see managerial new product orientation as the key indicator of social risk.

Organizations also channel employees’ risk assessments through monitoring and rewards (Sitkin and Pablo 1992). In B2B settings managers typically find themselves unable to set specific rewards for new products because they cannot make an accurate estimation of the true customer demand for new offerings (Schöttner 2016). Moreover, installing new product-specific incentives increases the chance that salespeople push new products that customers do not need or want. Rather than linking salary and bonuses to new product sales volume, firms typically install long-term reward systems, as put by an R&D manager from a logistics company: “We motivate our salespeople to take a long-term perspective that aligns with our strategic objectives.” Long-term rewards lower the perceived financial risks associated with selling new products because they provide more leeway for salespeople to obtain their targets (Wei and Atuahene-Gima 2009). We thus regard long-term rewards as the key indicator of perceived financial risk and define them as incentives that aim to facilitate the achievement of various long-term objectives and specified strategic goals in a time frame of longer than one year.

We also examine the important role of external information during salespeople’s risks processing. New product information typically comes to salespeople in aggregated form and is centrally coordinated, as described by a manufacturing company product manager: “Our salespeople really need to be convinced about the added value of new products, how they address the problems and needs of their customers. We invest a lot of resources and time in away days, workshops, training, information meetings, drinks, exclusive trips, et cetera.” Managers expect that salespeople become less conservative because information lowers perceived risk through clarifying how the new product benefits customers and “signaling” the company’s commitment to the new product (Erdem and Swait 1998). However, in accordance with perceived risk processing theory, we do not posit a direct but a moderating effect of new product information as it influences how employees act on their perceived risk dimensions (Conchar et al. 2004).

Finally, we account for three categories of evaluation standards that previous perceived risk processing studies have outlined. First, individual preferences toward risk are captured in self-confidence (Conchar et al. 2004; Mitchell 1999), company tenure (Wang 2015), and past performance (Sitkin and Pablo 1992). These factors provide an individual with evidence from past or enduring abilities to overcome obstacles and therefore drive individuals to accept risks that others would avoid (Wiseman and Gomez-Mejia 1998). Second, people are more or less likely to take risk as a function of their past investments and resultant current asset levels; individuals with more favorable current assets are more likely to avoid risky choices (Conchar et al. 2004). We thus consider a salesperson’s pay scale and customer relationship quality (i.e., the salesperson’s perceptions of his/her customers’ trust in, satisfaction with, and commitment to him/her) to represent past investments and achievements within the company and its customer base, respectively. Third, individual risk taking depends on whether problems are presented as gains or losses (Kahneman and Tversky 1979). In our context, a radically new product may also offer much value to customers. We define new product advantage as the salesperson’s perception of product superiority relative to existing products with respect to quality, cost-benefit ratio, or technological innovativeness, and account for the possibility that such perceptions may affect risk behavior.

In sum, Fig. 1 presents our conceptual model. We describe our hypotheses next.

An individual’s risk perceptions, information processing, and risky choice are inextricably linked (e.g., Conchar et al. 2004; Dowling and Staelin 1994). In this nomological network the information available to decision makers determines the relative weight of the perceived risk dimensions through a process of cognitive priming (Mandel 2003). Priming an individual with a specific information cue creates cognitive activation tags. When individuals assess risky choice situations, the perceived risk dimensions have to “make contact with one of the tags left earlier and find an intersection” (Collins and Loftus 1975, p. 409). These intersections are easily available and retrievable at the time a risky decision has to be made and thereby affect the weight of perceived risk dimensions in an individual’s choice (Mandel 2003; Scheufele and Tewksbury 2007).

The dimension that most closely corresponds to the primed information becomes more important relative to other dimensions. New product information specifically addresses how the new product and its features satisfy customer needs. Rather than stressing social or financial elements in risky situations, managers provide new product information in an effort to reassure salespeople that the product will perform well in the market (Atuahene-Gima 1997). It therefore appeals to the salesperson’s perception of performance risk. Providing new product information to salespeople will strengthen the relationship between perceived performance risk and CSB and weaken the relationships of perceived social risk and perceived financial risk with CSB. Hence, we expect that new product radicalness will become a more important antecedent and managerial new product orientation and long-term rewards become less important antecedents of CSB. Formally:

Previous research has convincingly demonstrated that higher levels of salesperson’s effort to sell new products leads to positive performance outcomes (Johnson and Sohi 2016). Reasons include that a high level of effort conveys to customers the value of the product and the salesperson’s confidence in the product (Ahearne et al. 2010) and being more persistent helps overcome occasional setbacks and thus closing deals (Fu et al. 2009). Effort to sell new products also positively relates to a manager’s overall performance evaluation, as it signifies that salespeople are willing to go the extra mile and do not refrain from engaging in difficult selling tasks. Harris et al. (2014) even report that sales managers may prefer hard work and productivity over a salesperson’s intentions to satisfy customer needs. Although disconcerting from a marketing point of view, it shows the importance of salesperson effort in managerial evaluations of their subordinates’ performance.

When salespeople display CSB and thus delay their efforts to sell the new product during and across customer encounters, the resources remaining for intensively selling the new product in the end are more limited because of the shorter time period available. In addition, people revising their initial choice from a set of options need some time to recalibrate because of cognitive processes such as dissonance, rationalization, anticipated regret, or inertia (Hoch 2002). Salespeople who (have to) switch their selling efforts from existing to new products may therefore only grudgingly accept their new strategy. Their overall effort to sell new products will therefore be lower. In sum, we expect that effort to sell new products will mediate between CSB and performance outcomes. Formally:

Salespeople are rational actors who allocate their cognitive and physical resources across a portfolio of products in a way that maximizes overall performance (Ahearne et al. 2010). This implies that they may plan and organize their selling activities in a way that makes sales efforts more effective (Allcott and Sweeney 2016; Rapp et al. 2006). One approach for salespeople to structure their selling activities is to change the order in which different products from the product portfolio are presented to the customer. We thus posit that conservative selling can be a strategy that combines with effort to sell new products to affect a salesperson’s performance outcomes.²

First, we expect that CSB dampens the positive relationship between effort to sell new products and overall managerial performance evaluation. Managers assess their subordinates by judging the degree to which a salesperson matches their ideal of a “good salesperson.” This is typically reflected in high effort and productivity (Harris et al. 2014). When launching new products, managers expect this effort to be enduring because salespeople need to open up a new market by informing and educating customers about how the new product may address customer needs and problems (Fu et al. 2010). When salespeople first focus on selling existing products and delay their effort to sell the new product till the late stages of sales cycles, managers will feel that their employees had the chance to put in more effort to sell the new product but did not take this opportunity. Additional effort that employees put into selling new products after a period of conservatism thus translates less strongly into managerial evaluations. In contrast, when salespeople expend effort in selling new products throughout the sales cycle, managers may feel that employees constantly took initiative and were competitive in selling new products (Pettijohn et al. 2001). Because the sales activities of such employees lack a period where no effort was expended on selling the new product, managers are less likely to think that more effort could have been put in. Each additional unit of effort is then appreciated more because managers feel that employees may have reached their cognitive and physical limits and now go the extra mile.

Second, we expect that CSB strengthens the positive relationship between effort to sell new products and new product selling performance. Because CSB indicates the extent to which a salesperson maximizes selling efforts for existing products before engaging in efforts to sell the new product, customers likely experience a sequential presentation of products in a sales cycle with a salesperson who acts conservatively. Literature supports the notion that new products become more attractive when presented following existing products. For instance, sales literature in consumer settings suggests that foot-in-the-door techniques can help lower initial resistance to adopt because in their strive for consistent responses, customers agreeing to a small initial request are more likely to comply with a larger or riskier request (Cialdini and Guadagno 2004). Sequential presentation also makes a new product look more attractive to a customer; compared to an existing product, each additional feature of a new product may add desired capabilities and thus provide the customer with another reason to purchase (Thompson et al. 2005). This makes the effort spent on selling the new product more effective. In sum, we posit:

Following our qualitative grounding discussed earlier, for Sample 1 we gathered data from a global ICT company that operates in 90 countries and is representative of B2B selling contexts as (1) new products are introduced annually, (2) new products are complex and break from existing offerings, (3) the salesforce organization is unit-based, and (4) selling is a relational rather than a transactional activity. The company’s product portfolio consists of workspace management systems, connectivity solutions, and datacenters, among others. Such products have a relatively short life cycle.

The company’s sales force focuses on a set of approximately 500 business customers in industries such as finance, government, education, transport, and retail. At the time of study, the company had just introduced several new solutions that required significant changes in customers’ work processes. The radicalness of these solutions differed across sectors; not every sector faced equally substantial changes to their work processes. The new products immediately entered the salespeople’s product portfolios and accounted for a substantial portion of the company’s total annual revenue (28%). Sales units received collective briefings and training about the new product’s features, value proposition, and link with customer needs.

We collected data from three sources at different points in time. We asked all 244 salespeople and their 31 managers, organized in 31 sales units, to complete a questionnaire. After two reminders, sent over a three-week period, we received 172 responses from salespeople (70.5% response rate) and 31 responses from managers (100% response rate). All units sampled featured at least 3 responding salespeople. Six months after collecting the questionnaire data, we obtained performance data from company records.

With minor wording adjustments to enhance the applicability of some items, most of our constructs could be operationalized with scales validated in previous work. However, because CSB is a new concept, we carefully considered its operationalization. Following our review of relevant literature and general qualitative grounding, we interviewed four sales managers of our focal company and asked them to reflect on their experiences with new product launches in the salesforce and what actions they typically associate with salesperson conservatism. The managers consistently mentioned elements such as being cautious, sticking to existing sales routines, and preferring to maximize the potential of proven sellers first.

Based on the managers’ input and studies on political (e.g., Jost et al. 2003) and accounting (e.g., Watts 2003) conservatism, we developed an item pool. We conducted industry-specific investigations to define the average product life cycle and sales process duration. As a result, items referred to “new products” when those were introduced in the 12 months preceding the questionnaire. The initial pool of items was then refined based on further in-depth interviews with the sales managers, their salespeople, and their sales support staff. Next, we constructed a draft questionnaire and pretested it with six company employees and two industry experts. Following the pretests, we made minor wording adjustments to enhance the applicability of the items. The resulting scale consists of three items.

Table 2 contains the scale items for our measures. All responses were recorded on five-point Likert scales with 1 (“strongly disagree”) and 5 (“strongly agree”) as anchors. To assess new product information, sales managers completed four items adapted from Low and Mohr (2001) to indicate the extent to which the salespeople in their unit received timely, relevant, and accurate information on how new products address customer needs. It is therefore a unit-level measure.

The salesperson questionnaire included managerial new product orientation, measured with five items from Van der Borgh and Schepers (2014). Long-term rewards was measured with three items adopted from Wei and Atuahene-Gima (2009) such that a low score on these items indicate a focus on short-term rewards. New product radicalness was measured using a four-item scale developed by Langerak et al. (2008), and effort to sell new products was based on work by Sujan et al. (1994) and Hultink and Atuahene-Gima (2000).

New product selling performance was taken from company records and reflects for each salesperson the sales volume generated from the sale of new products as a percentage of his/her overall sales volume. Managerial overall performance evaluation was also collected from company records as we were given access to an aggregate measure of managers’ formal evaluations of each salesperson’s overall functioning. Sub dimensions tapped into individual performance in terms of output and behavior and included questions as “This salesperson obtained revenue targets for his or her customers,” “This salesperson sticks to the company’s formal rules and regulations,” and “This salesperson contributes to the company’s success.” For each employee, the aggregated score indicated a “poor” (1) to “excellent” (5) evaluation.

We controlled for the evaluation standards that salespeople may use in their risk evaluation phase. Specifically, we measured self-confidence with one item from Riggs and Knight (1994). In addition, company tenure (i.e., years with the firm), past performance (i.e., order intake target obtained on all products in previous year), and pay scale (i.e., a market-based salary structure dividing sales people in different levels of salary relative to the market) were obtained from company databases. Customer relationship quality was measured with four items from Palmatier (2008), and new product advantage was measured using a four-item scale developed by Langerak et al. (2008).

Because salespeople were nested within sales units, responses from salespeople of the same unit may be interdependent. To determine whether we should explicitly account for multiple levels in our analyses, we examined the intraclass correlation coefficients (ICC) for the variables in our model; ICCs ranged from .025 to .176. Even small ICC values (e.g., .05) indicate that researchers should control for dependence of observations to prevent considerable bias in the results (Cohen et al. 2003, p. 538). We thus accounted for the multilevel structure of our data and estimated a multilevel structural equation model (MSEM) with Mplus 7.11 software (Muthén and Muthén 2012). Compared to regression-based multilevel approaches, MSEM has the advantage not to conflate within- and between-group effects. MSEM separates the effects using latent variables at both levels and thereby accounts for measurement error (Preacher et al. 2010).

Finally, we obtained a relatively small sample and set out to test moderated and mediated effects that are non-normally distributed. Given these conditions we employed Bayesian methods because these provide more reliable estimations on small samples (Muthén and Asparouhov 2012) and do not assume or require normal distributions for the model parameters (Zhang et al. 2009).

To test whether the data fit the hypothesized measurement model, we conducted a confirmatory factor analysis (CFA) that accounted for the non-identification problem that may occur with small sample sizes (i.e., the CFA is Bayesian) and that considered the nested nature of our data (i.e., the CFA is multi-level). Table 2 reports the results.³

To determine the Bayesian CFA model fit, we examined the posterior predictive p (ppp) value. Our BCFA showed a ppp-value of .862, which indicated a good fit between the model and the data. One cross-loading (λ_{NPA, npr_1} = .198) and four residual covariances (σ_{npo_3,npo_4} = −.129; σ_{ltr_1,ltr_3} = −.149; σ_{eff_2,eff_3} = .055; σ_{eff_4,eff_5} = .075) were found significant, which warrants the use of Bayesian estimation techniques (i.e., subscript capitals indicate latent constructs, lower case indicates items. Please refer to Tables 2 and 3 for details on the acronyms).

All items loaded on their respective factors with substantial values, and no serious cross-loadings (i.e., > .3) were observed. Table 3 shows that the scales also achieved sufficient reliability, with Cronbach’s alphas varying between .73 and .93 and composite reliabilities varying between .82 and .94. Average variances extracted (AVE) exceeded .50 for each construct, in support of convergent validity. The constructs also displayed discriminant validity because the AVE of each construct exceeded the average variance shared with any other construct.

We evaluated the validity of our CSB concept and scale in Sample 1 using a second dataset (Sample 2) obtained from a commercially available panel of 191 B2B salespeople. The CSB scale again displayed desirable psychometric properties and satisfied the criteria for discriminant validity versus related concepts. Appendix A provides more detail.

We tested our hypothesized model using Sample 1 and took a stepwise approach. In the first step, we included the control variables and our hypothesized direct effects (Model 1). In the second step, we specified and added the interaction effects of unit-level new product information (Model 2). We standardized all independent variables before creating the product terms to enable model convergence and facilitate the interpretation of the coefficients without altering the underlying data. Thus, we obtained the following multilevel equations:

SPO_ij indicates salesperson performance outcome for salesperson i of unit j, other acronyms are explained in Table 3. Furthermore, γ₀₀, α ₀₀, and θ₀₀ are intercepts; γ₁₀... θ₁₁₀ are regression coefficients; e_ij, ε_ij, and ϵ_ij, are individual-level error terms; and u_oj, π_oj, and μ_oj are unique variations of unit j from the intercept (i.e., γ₀₀, α ₀₀, and θ₀₀), after partialling out the effects of all unit-level regression coefficients (i.e., γ₀₁, α ₀₁). We then proceeded with our Bayesian estimation procedure.⁴

Tables 4a and 4b displays the results of the estimations of our two models. Model 2 explained significantly more variation than Model 1. In support of H1, we found a direct positive effect of new product radicalness on CSB (CI_95% = [.02; .31]). The results also confirm H2, as the effect of managerial new product orientation on CSB is significant and negative (CI_95% = [−.33; −.01]). The main effect of long-term rewards on CSB is significant and positive (CI_95% = [.03; .29]), while the hypothesized effect was negative. This is an interesting finding; we return to this result and discuss potential explanations below. Hence, despite the significant relationship have to reject H3.

We now turn to the moderating effects of new product information. The results show a positive interaction effect of new product information and new product radicalness (CI_95% = [.04; .34]) toward CSB. In addition, the positive interaction effect of new product information and managerial new product orientation toward CSB was supported by the data (CI_95% = [.01; .32]). This confirms H4a and H4b. The interaction term of new product information and long-term rewards was significant and negative (CI_95% = [−.31; −.01]). Because the main effect of long-term rewards on CSB turned out positive, this negative interaction effect aligns with our overall argument that long-term rewards would become a less important driver of CSB following the provision of new product information to the salesforce. Therefore, although we formally reject H4c, we do find empirical support for the dampening effect of new product information expected in the hypothesis. We interpret this effect below.

Panels A, B, and C in Fig. 2 plot the results of the interaction effects to facilitate interpretation. The plots provide the results for values 1 standard deviation below and above the mean of the independent variable and the moderator. Panel A reveals that new product radicalness leads to more CSB when new product information is high. Remarkably, when little new product information is provided to salespeople, there seems to be no effect of new product radicalness on CSB. Panel B shows that under conditions of low new product information the salesperson follows managerial guidance and lowers his/her CSB, presumably because being conservative entails a (too) large perceived social risk. However, when a salesperson receives new product information, the effect of managerial orientation on CSB is neutralized because the perceived performance risk becomes more salient in lieu of perceived social risk.

Finally, Panel C shows that a similar logic applies to financial risk: when new product information is available, long-term rewards have little effect on CSB because perceived financial risk becomes less salient in individual’s risk assessment. Remarkably though, under low information conditions, the effect of long-term rewards is positive. Long-term rewards motivate employees to care about realizing the firm’s long-term revenue growth (Wei and Atuahene-Gima 2009). This goal can be accomplished through the sales of any product in a salesperson’s portfolio. Because salespeople more easily see how to realize revenue growth through proven sellers than through difficult-to-sell new products (Wieseke et al. 2008), long-term rewards could perhaps de-emphasize salespeople’s concerns about a new product’s success in the market. An alternative explanation is that long-term rewards and new product information substitute each other: when salespeople have sufficient information, there is no need for long-term rewards since salespeople are able to make well-informed decisions.

Overall, Fig. 2 shows three distinct moderating effects (cf. Cohen et al. 2003). Paired with new product radicalness, new product information has a synergetic (or enhancing) effect on CSB, paired with managerial new product orientation it has a buffering effect, and paired with long-term rewards it has a substitution (or antagonistic) effect.

With regard to CSB’s consequences, the results show a negative effect of CSB on effort to sell new products (CI_95% = [−.33; −.01]). This supports H5a. Furthermore, new product selling effort was positively related to both performance measures, i.e., managerial overall performance evaluation (CI_95% = [.05; .25]) and new product selling performance (CI_95% = [1.87; 9.77]). This provides support for H5b and H5c, respectively. Finally, CSB weakened the positive relationship between effort and manager’s evaluations (CI_95% = [−.15; −.02]), providing support for H6a. In contrast, CSB strengthened the positive relationship between effort and new product selling performance (CI_95% = [.27; 6.58]) in support of H6b. This provides evidence for the thought that although managers see CSB as an undesirable behavior, it may in fact benefit sales performance. Figure 3’s panels A and B present a more detailed account of the interaction effects of effort and CSB, again for values of 1 standard deviation below and above the mean of the independent variable and the moderator.

To test whether effort, in its relationship with performance, both mediates and interacts with CSB, we calculated the index of moderated mediation for CSB (Hayes 2015). In our “independent variable as moderator” model, the indirect effect of CSB on SPO through EFF is the product of the effect of CSB from Eq. 2 and the conditional effect of EFF on SPO from Eq. 3 (see Preacher et al. 2007 for a detailed discussion). Formally:which equals:with intercept α₄₀θ₂₀ and slope α₄₀θ₄₀CSB _ij . In this model α₄₀θ₄₀CSB _ij is the index of moderated mediation. Using Bayesian estimation, we obtained 95% credible intervals. The moderated mediation indices for CSB on a manager’s overall performance evaluation and new product selling performance are both significantly different from zero (CI_95% = [.01; .04] and [−1.92; −.05], respectively). This supports the moderated mediation effect of CSB and demonstrates that the negative indirect effect of CSB on managerial overall performance evaluation increases with higher values of CSB, as the slope of the line (or the index of moderated mediation) is positive, while the negative indirect effect of CSB on new product selling performance decreases with higher values of CSB. In sum, the findings further corroborate our expectations.

Of our control variables, company tenure positively associated with CSB (CI_95% = [.02; .32]). Probably, more tenured employees have gained more status and are reluctant to incur any damage by making risky choices. Alternatively, more tenured employees may be older and therefore more likely to resist change and be more conservative in their decision making. As expected, self-confidence associated negatively with CSB (CI_95% = [−.49; −.08]) and positively with effort to sell new products (CI_95% = [.17; .59]). Next, customer relationship quality positively associated with effort to sell new products (CI_95% = [.07; .38]). This suggests that trusted relations make salespeople more willing to increase their effort allocated to that account. An alternative explanation could be that these accounts demand more from the salesperson (Mullins et al. 2014). Finally, in line with previous studies we find that past performance positively associates with managerial overall performance evaluations (CI_95% = [.04; .23]). All of these results are intuitive and thus offer face validity to our data and increase the confidence in our outcomes.

Our work has several important theoretical implications. First, we add to existing work in new product selling by conceptualizing, operationalizing, and validating the concept of CSB. We conceptualize CSB as a sales domain–specific behavioral representation of risk aversion. The latter has been a central assumption in the sales domain and in new product selling in particular (Atuahene-Gima 1997). We show that CSB differs from related constructs such as new product selling effort.

Remarkably, we find that when salespeople first try to sell existing products before they engage in new product selling, their effort to sell new products more strongly translates into percentage of sales volume obtained with new products. Explanations for this effect can be found in how customers strive for consistent responses to sales requests (Cialdini and Guadagno 2004), creeping commitment in B2B sales cycles, and salespeople’s potential to better communicate the new product’s attractiveness through highlighting new features vis-à-vis old (or missing) features (Thompson et al. 2005). Previous literature has shown that adapting sales presentations to the nature of the sales situation benefits overall sales performance (Franke and Park 2006) but has remained silent on the specific alterations needed in new product selling. We show that the order of presenting the customer with new and existing products is a key factor in such adaptive strategies.

Second, we add to recent work on internal marketing of new products (e.g., Wieseke et al. 2008). In contrast to the traditional view that internal marketing efforts motivate salespeople to sell new products (Atuahene-Gima 1997; Johnson and Sohi 2016), we show that this relationship is not straightforward. Salespeople perceive different risk dimensions surrounding the sale of new products. Providing new product information changes the relative weight of the dimensions in salespeople’s decisions to act conservatively. More specifically, product radicalness (performance risk) becomes a more important driver of CSB, while managerial orientations (social risk) and long-term rewards (financial risk) become less important. Especially the latter finding is interesting because without information, long-term rewards tend to increase rather than decrease CSB. Long-term rewards may focus salespeople on selling existing products, while short-term rewards provoke risk taking.

Finally, we add to work on (sales) personnel evaluation. Previous studies have showed that sales managers are sensitive to subordinates exerting high levels of effort (Harris et al. 2014). Early work also indicates that sales managers often fail to account for the task difficulty facing their employees (Mowen et al. 1985) and that more tactical elements such as planning ability, judgment, and time management are of lesser importance in performance appraisals (Hawes et al. 1995; Pettijohn et al. 2001). Our findings align with and add more detail to these insights. We show that managers are less appreciative of effort to sell new products when this effort follows a period of salesperson conservatism. Although this practice does not honor the associated sales performance, it does suggest that sales managers take into account a longer time-frame in their appraisals and do not suffer from “recency bias” (Brown and Heywood 2005) such that recent events have more influence on appraisals than less recent events. Our findings suggest that sales managers can oversee the entire length of the sales cycle and that they factor in the timing of their subordinates’ effort.

Our work offers important insights for managers. First, we urge sales managers to get a better understanding of the conservative behavior of their salespeople. Many managers worry that salespeople maximize their selling efforts for existing products before engaging in efforts to sell the new product and that this practice makes them less effective in selling new products. However, a conservative selling approach may help salespeople to provide a better pitch to customers who are then more likely to agree to risky value propositions. This is especially likely in B2B settings, where sales cycles extend over time and adopting a new product means that the customer faces significant changes to its workflows.

Second, managers’ better understanding of conservatism should go hand-in-hand with rethinking salespeople’s performance appraisals. Many managers use criteria such as product knowledge, attitude, effort, initiative, and aggressiveness in their performance evaluations of salespeople. Specifically, they correlate observed lack of effort to sell new products to a salesperson’s attitude and work motivation, consequently attribute negative traits to him or her, and adjust their evaluation of this person’s work performance downward. This may be unfair in markets where conservative selling behavior is a smart selling strategy.

Third, managers have routinely tried to control the levels of conservatism in their salesforce by using incentive systems, being an advocate of selling new products, and ramping up internal marketing on new products toward their employees. Rather than doing all at once, we urge managers to consider the interrelationships between these control mechanisms (see Fig. 2). Our results point to three possible control scenarios: (1) Managers responsible for launching radically new products can enhance CSB by setting up intensive internal marketing efforts via the provision of new product information. In this scenario, managers are advised to save their scarce resources of personally motivating sales staff and installing long-term rewards systems because providing new product information makes these mechanisms less suitable to regulate CSB. (2) Managers responsible for launching less radically new products can enhance CSB by rewarding the obtainment of objectives and goals in a time frame for longer than one year. Importantly, this strategy only works when providing no or very limited new product information to salespeople. (3) Managers who are looking to decrease CSB―perhaps because of a more transactional selling context or because CSB does not fit the organization’s image―are most successful when they do not provide new product information and set short-term (rather than long-term) goals. This runs counter to present-day management practices and explains why companies are often unsuccessful in managing CSB. Personally motivating subordinates to sell new products without providing new product information is likely to further decrease CSB.

Finally, there are many trade-offs that sales employees and their managers have to make because new products are added to a portfolio of existing products. B2B managers could think about allocating a specific part of their salesforce to sell new products and the remaining part to sell existing products. An alternative to this solution is to hire new salespeople to sell new products. A problem here, though, is that a lower quality of customer relationships limits the knowledge exchange between salesperson and customer. This may hamper the transfer of customer feedback into the company’s development department and thus block bug fixing and product quality improvement.

Although we believe our study has clear merits, there are some limitations too. Some of them provide fruitful avenues for future research. First, we relied on salesperson perceptions of new product radicalness, managerial new product orientation, and long-term rewards. Although this fits with our perceived risk processing framework, recent studies indicate that employee perceptions may be different from customer-side perceptions or manager perceptions and relate differently to performance outcomes (e.g., Mullins et al. 2014). Including customers as a data source would provide further evidence for the validity of our findings, but collecting a sample of representative respondents is hard in B2B settings.

In addition, we demonstrated the robustness of our findings through triangulation with previous results and controlling for market context and conditions. We hold that our sample profile is typical for firms with field salesforces and frequent introductions of new products in many industries, such as machinery, chemicals, plastic materials, equipment and supplies, and pharmaceuticals. Assuming that with increasing frequency of new product introductions salespeople get more used to selling new and unfamiliar products, we expect that any CSB-related effects also hold for (or may even be stronger in) industries with a lower frequency of new product introductions. However, in industries where competition offers similar products, salespeople have limited time to convince customers, or where fad products are concerned, CSB may have different effects. Future research should hence substantiate the effects across industries to see in which situations CSB is most beneficial.