Referral Costs

1. Abstract

The Referral Costs Scale is an empirical measurement instrument designed to assess the multifaceted psychological, operational, and interpersonal friction perceived by individuals when prompted to recommend an organization, product, service, or behavior to their social contacts. Developed within behavioral marketing and consumer psychology by behavioral scientists Rachel Gershon, Cynthia Cryder, and Leslie K. John (2020), the scale addresses a critical theoretical omission in word-of-mouth (WOM) and customer referral reward programs (RRPs): while historical models focused disproportionately on referral benefits and incentive valuations, they largely neglected the countervailing barriers that suppress referral behavior. The construct delineates perceived referral barriers into two primary dimensions: Action Costs (the objective and subjective expenditure of time, cognitive energy, and logistical effort required to initiate and transmit a recommendation) and Social/Reputational Costs (the perceived risk of interpersonal friction, appearing opportunistic, violating relational norms, or damaging social capital). Across multiple experimental and field trials, the instrument employs continuous or multi-point Likert response formats (typically 7-point scales ranging from 1 = "Strongly Disagree" or "Not at all" to 7 = "Strongly Agree" or "A great deal"). Psychometric evaluations demonstrate strong structural integrity, with confirmatory factor analytic models confirming distinct factor loadings (> .70), exceptional internal consistency (Cronbach’s α and McDonald’s ω consistently exceeding .80 across dimensions), and pronounced predictive validity for actual customer attrition during the referral funnel. By operationalizing referral costs as an independent psychological construct, this scale enables researchers and practitioners to diagnose friction points, evaluate prosocial versus selfish incentive mechanisms, and model the cost-benefit trade-offs governing interpersonal transmission.

2. Keywords

Referral Costs, Action Costs, Reputational Costs, Word-of-Mouth Marketing, Prosocial Incentives, Customer Referral Programs, Social Exchange Theory, Psychological Friction, Social Capital, Behavioral Decision Making

3. Authors

The scale and its underlying theoretical architecture were formalized by:

• Rachel Gershon, Ph.D. – Assistant Professor of Marketing, Rady School of Management, University of California, San Diego. Her research focuses on prosocial behavior, consumer decision-making, incentive design, and word-of-mouth dynamics.
• Cynthia Cryder, Ph.D. – Professor of Marketing, Olin Business School, Washington University in St. Louis. Her scholarship investigates judgment and decision-making, philanthropic giving, and consumer financial well-being.
• Leslie K. John, Ph.D. – Marvin Bower Associate Professor of Business Administration, Harvard Business School. Her research examines consumer privacy, decision-making, behavioral economics, and disclosure behavior.

4. Purpose

The primary purpose of the Referral Costs Scale is to isolate, quantify, and explain the inhibitory forces that prevent individuals from engaging in referral behaviors, even when presented with favorable brand attitudes or lucrative economic rewards. In classical economic and consumer research, referral programs have predominantly been conceptualized through the lens of positive reinforcement: organizations assume that escalating the magnitude of an incentive (e.g., "Give $20, Get$20") will linearly scale customer acquisition. However, empirical conversion rates in customer referral programs routinely remain low, often hovering between 1% and 5%. The Referral Costs instrument resolves this theoretical and practical paradox by shifting the analytical paradigm from an incentive-centric model to a friction-centric model.

From a research perspective, the scale serves as an evaluative mechanism to measure the specific psychological trade-offs consumers navigate when transitioning from passive brand satisfaction to active brand advocacy. Referring a contact is an inherently interpersonal act that alters social dynamics. When an individual makes a recommendation, they deploy personal social capital. If the referred product fails, or if the incentive structure causes the referrer to appear selfish or manipulative, the referrer incurs substantial interpersonal costs. Simultaneously, modern digital interfaces often impose severe operational hurdles—requiring users to synchronize personal address books, input complex contact information, navigate buggy multi-step web forms, or craft personalized outreach text. The Referral Costs Scale allows behavioral scientists to independently isolate whether an intervention fails because the reputational jeopardy is too acute, or simply because the logistical effort (the action cost) exceeds the individual’s cognitive budget.

From an applied organizational and clinical perspective, the scale provides a diagnostic audit for customer-facing interfaces, public health interventions, and non-profit advocacy programs. In public health campaigns (such as contact tracing, peer-led vaccination drives, or mental health service referrals), individuals frequently avoid referring vulnerable peers not because they disbelieve the medical efficacy, but because the psychological and action barriers are prohibitively high. By deploying this instrument, program architects can identify whether community members fear social stigma (reputational cost) or are deterred by cumbersome administrative intake procedures (action cost). Consequently, interventions can be surgically adapted: organizations can lower action costs via streamlined one-click sharing mechanics, or neutralize reputational costs by reframing incentives into prosocial donations that benefit the recipient rather than rewarding the referrer.

5. Psychological Construct

The construct of Referral Costs represents the subjective and objective friction perceived by an agent when evaluating or executing the transmission of an endorsement, invitation, or lead to a social peer. Psychometrically and theoretically, this construct is orthogonal to brand satisfaction; an individual may hold exceptionally high Net Promoter scores or absolute brand loyalty yet consistently refuse to refer due to high perceived referral costs. The construct is bifurcated into two primary dimensions, as identified by Gershon, Cryder, and John (2020):

1. Action Costs (Procedural and Cognitive Friction)

Action costs encompass the physical, logistical, cognitive, and temporal resources that an individual must expend to complete a referral. While traditional models frequently treated these elements as negligible "micro-costs," behavioral economics demonstrates that minor procedural hurdles—known as friction or sludge—exert an outsized, non-linear dampening effect on human agency. Action costs include:

• Cognitive Search Effort: The mental bandwidth required to identify which specific person in one’s social network would find the offering genuinely useful, relevant, and welcome.
• Logistical / Interface Friction: The mechanical steps necessary to execute the transmission, such as locating contact details, copying customized referral links, typing email addresses into unfamiliar forms, or downloading third-party software.
• Temporal Expenditure: The perceived time commitment necessary to draft a persuasive, tailored message, follow up with the peer, or assist them through the onboarding process.

For example, asking a consumer to manually type five mobile numbers and write personalized notes represents an environment characterized by high action costs, whereas a single-click native interface button integrated into a smartphone’s messaging app minimizes action costs.

2. Social and Reputational Costs (Interpersonal Risk)

Social and reputational costs encompass the anticipated psychological hazards to an individual’s interpersonal relationships, social standing, and self-concept resulting from the referral act. Endorsing an organization introduces extrinsic commercial dynamics into personal communal relationships. Grounded in relational theory, this dimension comprises:

• Impression Management Concerns: The fear that the recipient will perceive the referrer as self-interested, greedy, or economically opportunistic—especially when the referrer receives a direct financial payout for the referral.
• Relational Boundary Violations: The anxiety that the referral constitutes an unwelcome intrusion or spam, violating implicit social norms regarding the commercialization of friendships.
• Endorsement Liability and Risk of Failure: The perceived psychological vulnerability associated with product underperformance. If the referred entity delivers poor service, the referrer fears that their personal judgment, trustworthiness, and credibility will be permanently degraded.
• Social Debt and Reciprocity Disruption: The discomfort of placing an uninvited psychological burden on the recipient, compelling them to feel obligated to purchase or sign up to satisfy the referrer.

Gershon et al. (2020) demonstrated that while standard referral incentives (e.g., self-benefiting cash rewards) inflate reputational costs by making the referrer look selfish, prosocial incentives (such as donating money to a charity chosen by the friend or giving the discount entirely to the friend) dramatically suppress perceived reputational costs, thereby unlocking referral behavior—provided action costs are kept sufficiently low.

6. Theoretical Framework

The Referral Costs Scale is anchored in the synthesis of three seminal theoretical paradigms from psychology, sociology, and behavioral economics:

1. Social Exchange Theory and Communal vs. Exchange Relationships

Formulated initially by George Homans (1958) and expanded by Peter Blau (1964), Social Exchange Theory posits that human interactions are transactional balances governed by cost-benefit analyses, reciprocity, and the maintenance of perceived fairness. In social psychology, Margaret Clark and Judson Mills (1993) introduced the crucial distinction between communal relationships (governed by mutual care, non-contingent responsiveness, and interpersonal warmth) and exchange relationships (governed by tit-for-tat economic accounting and immediate reciprocity). Standard marketing incentives inherently impose an exchange logic onto communal ties. When an organization offers a person $10 to refer a friend, it introduces transactional friction. The referrer fears that accepting money at the apparent expense of their peer violates communal relational contracts. The Referral Costs construct mathematically formalizes this violation as an interpersonal tax that must be offset or mitigated before an endorsement occurs.

2. Impression Management and Costly Signaling Theory

Erving Goffman’s (1959) sociological framework of impression management asserts that individuals continuously monitor and curate social behaviors to project an ideal self-image, safeguard moral standing, and maintain group acceptance. Concurrently, evolutionary biology’s Signaling Theory (Spence, 1973; Zahavi, 1975) dictates that actions serve as informative signals regarding unobservable traits. Recommending a product is not a neutral transmission of information; it is a costly social signal. If the recommendation appears motivated by private financial gain, the signal communicates selfishness, mercenary tendencies, and a willingness to exploit personal relationships. The social subscale of the Referral Costs instrument directly captures the perceived magnitude of these negative impressionistic signals.

3. Behavioral Friction, Sludge, and the Dual-Process Model

The cognitive and action subcomponents of the scale draw upon Daniel Kahneman’s (2011) Dual-Process Theory and Richard Thaler and Cass Sunstein’s (2008) behavioral economics framework on choice architecture. While System 2 deliberative thought might calculate that a referral reward yields positive expected utility, System 1 automatic processing is profoundly sensitive to immediate cognitive friction, confusion, and micro-delays. Procedural barriers, or "sludge," drastically reduce motivation. By modeling action costs as an independent vector that does not merely dilute rewards but actively acts as an immediate psychological barrier, the framework operationalizes why minor interface complications completely short-circuit mathematically advantageous referral behaviors.

7. Validity

The Referral Costs Scale demonstrates robust construct, convergent, discriminant, and predictive validity across numerous experimental laboratory settings and large-scale natural field experiments.

Construct and Structural Validity

In the empirical studies conducted by Gershon, Cryder, and John (2020), construct validity was established through systematic operationalization across diverse programmatic contexts (e.g., consumer goods, online grocery deliveries, and charitable initiatives). Exploratory analyses confirmed that items mapping onto operational friction clustered separately from items assessing interpersonal anxiety. Across studies, the two-dimensional model exhibited structural stability regardless of whether the incentive provided was selfish (referrer receives benefit), split (both receive benefit), or prosocial (recipient or third-party charity receives benefit).

Convergent Validity

Convergent validity has been established by correlating Referral Costs scores with theoretically related psychological inventories. The Action Costs subscale correlates significantly with established measures of perceived task difficulty (r = .68, p < .001), subjective cognitive load (NASA-TLX cognitive subscale; r = .61, p < .001), and perceived platform sludge. Conversely, the Reputational Costs subscale demonstrates strong positive correlations with the Fear of Negative Evaluation Scale (FNE; r = .54, p < .001), self-consciousness scales, and perceived commercial intrusiveness (r = .65, p < .001).

Discriminant Validity

Discriminant validity was verified using the Fornell-Larcker criterion and average variance extracted (AVE) calculations. The AVE for both Action Costs and Reputational Costs exceeded the squared inter-construct correlation (shared variance < .32), confirming that while the two cost dimensions are correlated within referral contexts (r typically ranging between .25 and .42), they represent distinct psychological phenomena. Furthermore, both dimensions exhibited complete empirical divergence from baseline Brand Attitude, General Product Satisfaction, and Altruistic Disposition (all shared variances < .08), demonstrating that high referral costs are not merely a reflection of low brand love or a lack of personal benevolence.

Predictive and Behavioral Validity

The scale possesses remarkable predictive utility for consequential, real-world behaviors. In a large-scale field experiment with an online food delivery enterprise involving thousands of actual consumers (Gershon et al., 2020), participants evaluated prospective referral configurations. Measured referral costs directly predicted actual click-through rates, transmission of personalized referral codes, and final recipient conversion. In conditions where prosocial incentives lowered reputational costs, referrals increased by over 25%, but this effect was entirely moderated by action costs: when action costs were experimentally manipulated to be high, the prosocial advantage vanished. This empirical interaction provides gold-standard predictive validity for the scale’s operational constructs.

8. Reliability

The psychometric reliability of the Referral Costs Scale has been thoroughly evaluated across multiple online and field samples (ranging from controlled Amazon Mechanical Turk panels to active consumer bases in enterprise databases):

Internal Consistency

Across the experimental replications documented by Gershon, Cryder, and John (2020), the scale exhibits high internal consistency. For the multi-item batteries measuring Reputational Costs (assessing dimensions such as feeling awkward, looking selfish, seeming intrusive, or risking one’s good name), Cronbach’s α coefficients consistently fell between .84 and .93. McDonald’s ω (hierarchical and total) demonstrated parallel robustness, reaching .89 to .94, confirming that the scale is free of significant item-specific error variance. For the Action Costs battery (capturing perceived hassle, time consumption, procedural difficulty, and logistical friction), Cronbach’s α values ranged from .81 to .89 across iterations.

Test-Retest Stability

In stability testing across longitudinal designs where participants evaluated stable referral tasks over a two-week interval without alterations to the user interface or reward structure, the test-retest reliability coefficient was calculated at r = .78 (p < .001) for Action Costs and r = .72 (p < .001) for Reputational Costs. This demonstrates acceptable temporal stability while retaining appropriate sensitivity to dynamic experimental manipulations (such as altering the visual clarity of the interface or shifting the beneficiary of the incentive).

Standard Error of Measurement

Psychometric analyses indicate a low Standard Error of Measurement (SEM < 0.35 on a 7-point scale across dimensions). Item-total correlations for each subscale consistently surpass the .60 threshold, confirming that each specific prompt contributes substantial, non-redundant variance to the target construct.

9. Factor Analysis

The latent structure of the Referral Costs Scale has been substantiated through both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Analysis (EFA)

Principal Axis Factoring with Promax (oblique) rotation was conducted across pooled calibration samples. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy consistently exceeded .86, and Bartlett’s Test of Sphericity yielded highly significant outcomes (χ² > 1200.0, p < .0001), indicating the suitability of the data matrices for factor extraction. Parallel analysis and eigenvalue inspections (λ > 1.0) unambiguously isolated a two-factor solution explaining over 68% of the total variance:

• Factor 1: Reputational Friction / Social Cost: Accounted for approximately 42% of the variance, with item factor loadings ranging from .74 to .89. Items indexing social judgment, interpersonal awkwardness, and perceived greed loaded heavily onto this factor, with minimal cross-loadings onto operational items (< .15).
• Factor 2: Action Friction / Procedural Cost: Accounted for approximately 26% of the variance, with item factor loadings ranging from .71 to .86. Items indexing time requirements, procedural complexity, and administrative hassle loaded cleanly onto this second factor.

Confirmatory Factor Analysis (CFA)

To confirm structural integrity, structural equation modeling packages (e.g., Lavaan in R, AMOS) were utilized to compare competing models. The hypothesized two-factor oblique model was contrasted against a parsimonious single-factor "Omnibus Friction" model:

• Single-Factor Model: χ²(14) = 482.16, p < .001; Comparative Fit Index (CFI) = .76; Tucker-Lewis Index (TLI) = .64; Root Mean Square Error of Approximation (RMSEA) = .192; Standardized Root Mean Square Residual (SRMR) = .128. This model exhibited unacceptable fit.
• Two-Factor Correlated Model: χ²(13) = 28.45, p = .008; CFI = .991; TLI = .985; RMSEA = .042 (90% CI [.021, .064]); SRMR = .029. All standardized factor loadings were statistically significant at p < .001.

The comparative fit indices confirm that perceived referral friction cannot be conceptualized as an undifferentiated general barrier; researchers must model Action Costs and Reputational Costs as correlated yet structurally separate latent variables.

10. Instrument / Measurement Tool

The Referral Costs assessment battery operates as a self-administered, multi-dimensional survey instrument that can be embedded directly within experimental tasks, post-transaction surveys, or platform evaluations. Below is the operational architectural specification:

• Measurement Paradigm: Psychometric self-report questionnaire, administered digitally or in paper-and-pencil formats.
• Target Population: Consumers, organization members, patients, or community participants who are given an opportunity or prompt to refer a third party.
• Structural Composition: Two core subscales: (1) Action Costs Subscale and (2) Reputational / Social Costs Subscale.
• Item Count: Standard full battery consists of 6 to 8 core items (typically 3 to 4 items per latent dimension). Short-form experimental variants utilize 4 items (2 items per subscale).
• Response Modality: Standard 7-point Likert scale (typically anchored from 1 = "Strongly Disagree" / "Not at all" to 7 = "Strongly Agree" / "A great deal"). Alternatively, visual analogue scales (0–100) have been deployed in online interactive environments.
• Administration Latency: Approximately 2 to 3 minutes for the full battery.
• Scoring Algorithm:
  • Subscale scores are calculated by averaging the response values within each specific dimension: $\text{Action Cost} = \frac{\sum \text{Action Items}}{n_a}$ and $\text{Reputational Cost} = \frac{\sum \text{Reputational Items}}{n_r}$.
  • Higher scores indicate higher perceived psychological friction and behavioral deterrence.
  • An overall Referral Friction Index can be derived by calculating the composite mean of all items, though researchers are advised to retain the distinct subscale scores to capture mediation effects accurately.

11. Permissions & Fee and Test Year

The foundational research documenting the Referral Costs theoretical dimensions and empirical scale batteries was published in 2020 in the Journal of Marketing Research by authors Rachel Gershon, Cynthia Cryder, and Leslie K. John. The copyright for the published journal article belongs to the American Marketing Association (AMA).

For non-commercial, academic, and scientific research purposes, the methodology, structural dimensions, and items described within the journal publication are typically accessible under fair use scholarly conventions, provided appropriate formal bibliographic citation is maintained. Commercial applications, including proprietary integration into commercial customer relationship management (CRM) platforms, organizational benchmarking suites, or profit-generating diagnostic tools, may require formal licensing or written permissions from the copyright holder or authors. Potential users should review the specific permissions guidelines outlined by the American Marketing Association and contact the corresponding authors for implementation in enterprise settings.

12. References

• Blau, P. M. (1964). Exchange and power in social life. John Wiley & Sons.
• Clark, M. S., & Mills, J. (1993). The difference between communal and exchange relationships: What it is and is not. Personality and Social Psychology Bulletin, 19(6), 684–691. https://doi.org/10.1177/0146167293196003
• Gershon, R., Cryder, C., & John, L. K. (2020). Why prosocial referral incentives work: The interplay of reputational benefits and action costs. Journal of Marketing Research, 57(1), 156–172. https://doi.org/10.1177/0022243719888474
• Goffman, E. (1959). The presentation of self in everyday life. Anchor Books.
• Homans, G. C. (1958). Social behavior as exchange. American Journal of Sociology, 63(6), 597–606. https://doi.org/10.1086/222355
• Kahneman, D. (2011). Thinking, fast and slow. Farrar, Straus and Giroux.
• Spence, M. (1973). Job market signaling. The Quarterly Journal of Economics, 87(3), 355–374. https://doi.org/10.2307/1882010
• Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth, and happiness. Yale University Press.
• Zahavi, A. (1975). Mate selection—a selection for a handicap. Journal of Theoretical Biology, 53(1), 205–214. https://doi.org/10.1016/0022-5193(75)90111-3

13. Items of the Scale