Fundamentals

Imagine a local bakery, aromas wafting onto the street, drawing in passersby. They offer personalized recommendations ● “The sourdough today is exceptional,” the baker might say to a regular, remembering their preference. This interaction, in its analog simplicity, touches on personalization.

Now, translate that to the digital world where algorithms, not bakers, make suggestions. The crucial question arises ● how do we ensure these digital interactions are as ethically sound as that bakery exchange?

Defining Ethical Personalization

Ethical personalization, at its core, is about respecting the individual while enhancing their experience. It’s not about simply increasing click-through rates or maximizing sales; those are byproducts, not the primary goal. The aim is to create a beneficial exchange where the customer feels understood and valued, not manipulated or exploited. For a small business owner, this might seem like abstract corporate speak, yet it’s the bedrock of sustainable customer relationships, the kind that keep doors open and lights on.

Ethical personalization success Meaning ● Personalization Success, within the domain of Small and Medium-sized Businesses, signifies achieving quantifiable improvements in business metrics, such as customer lifetime value or conversion rates, directly attributable to tailored experiences. isn’t just about better sales; it’s about building stronger, more trusting customer relationships Meaning ● Customer Relationships, within the framework of SMB expansion, automation processes, and strategic execution, defines the methodologies and technologies SMBs use to manage and analyze customer interactions throughout the customer lifecycle. that fuel long-term SMB growth.

Transparency ● The Bedrock of Trust

Consider walking into a store and realizing you are being followed by security without any explanation. Unease sets in quickly. The digital realm operates similarly. Customers need to know why they are seeing specific recommendations or offers.

Transparency means clearly stating how data is collected and used for personalization. For an SMB, this could be as straightforward as a website privacy policy written in plain language, not dense legal jargon. It involves informing customers that their browsing history informs product suggestions, or that their location data helps suggest nearby store promotions. Openness builds trust; secrecy erodes it. Think of it as the baker openly sharing their recipe ● a sign of confidence and respect for the customer’s curiosity.

Control ● Empowering the Customer

Imagine being bombarded with bakery ads after just one visit, even if you only bought a single croissant. Annoying, right? Ethical personalization Meaning ● Ethical Personalization for SMBs: Tailoring customer experiences responsibly to build trust and sustainable growth. respects customer autonomy. Individuals should have control over their data and personalization preferences.

This means providing easy-to-use opt-in and opt-out options for personalization features. SMBs can implement simple preference centers where customers can manage their communication settings, data usage, and personalization levels. Offering granular control, like choosing specific types of recommendations or opting out of certain data collection practices, empowers customers. It shifts the dynamic from a business dictating the experience to a collaborative partnership where the customer is in the driver’s seat. This control is akin to allowing customers to choose their own fillings for a pastry ● catering to individual tastes and preferences.

Value Exchange ● Mutual Benefit

Think about loyalty programs at coffee shops. Buy ten coffees, get one free. It’s a clear value exchange. Ethical personalization operates on a similar principle.

Customers share their data expecting something valuable in return ● a more relevant, efficient, or enjoyable experience. If personalization only benefits the business, extracting data without providing commensurate value to the customer, it becomes exploitative. SMBs must ensure that personalization genuinely enhances the customer journey. This could be through time-saving recommendations, exclusive offers tailored to their needs, or a more seamless and intuitive online experience.

The value exchange needs to be balanced and perceived as fair by the customer. Like the coffee shop loyalty program, the benefit should be tangible and appreciated by the customer, fostering a sense of reciprocity.

Fairness and Bias Mitigation

Picture a bakery always recommending expensive pastries to certain customers based on assumptions about their income. It feels discriminatory. Algorithms, if not carefully designed, can perpetuate and amplify biases. Ethical personalization demands fairness.

Metrics must account for potential biases in data and algorithms, ensuring that personalization efforts do not unfairly disadvantage or discriminate against any customer segment. For SMBs, this means regularly reviewing personalization systems for unintended biases, testing recommendations across diverse customer groups, and being mindful of data sources that might reflect societal inequalities. Fairness is about treating all customers equitably, regardless of their background or demographics, just as a good baker would offer the same quality ingredients to everyone.

Measuring Ethical Personalization Success for SMBs

Moving beyond abstract principles, how can an SMB actually measure ethical personalization success? Traditional metrics like click-through rates and conversion rates are insufficient. They capture engagement but not necessarily ethical alignment.

SMBs need a new set of metrics that reflect customer trust, perceived value, and long-term relationship health. These metrics might be less immediate and easily quantifiable than sales figures, but they are far more indicative of sustainable, ethical growth.

Customer Lifetime Value (CLTV) and Retention Rate

Consider a customer who feels valued and respected. They are more likely to return and become a loyal patron. Customer Lifetime Value Meaning ● Customer Lifetime Value (CLTV) for SMBs is the projected net profit from a customer relationship, guiding strategic decisions for sustainable growth. (CLTV) and retention rate Meaning ● Retention Rate, in the context of Small and Medium-sized Businesses, represents the percentage of customers a business retains over a specific period. become crucial indicators. Ethical personalization should contribute to increased CLTV by fostering stronger customer loyalty.

A high retention rate, particularly among customers who have actively engaged with personalization features, suggests that these efforts are resonating positively. SMBs should track these metrics over time, comparing customer segments who experience personalized interactions with those who do not. An increase in CLTV and retention among personalized segments indicates that ethical personalization is not just effective but also contributing to long-term business health. This is like the bakery seeing regulars return week after week, a sign of sustained customer satisfaction Meaning ● Customer Satisfaction: Ensuring customer delight by consistently meeting and exceeding expectations, fostering loyalty and advocacy. and loyalty.

Customer Satisfaction (CSAT) and Net Promoter Score (NPS)

Imagine asking bakery customers, “How satisfied are you with your experience?” and “Would you recommend us to a friend?” Customer Satisfaction (CSAT) and Net Promoter Score Meaning ● Net Promoter Score (NPS) quantifies customer loyalty, directly influencing SMB revenue and growth. (NPS) provide direct feedback on customer sentiment. Ethical personalization should lead to higher CSAT and NPS scores. Surveys and feedback forms can be used to gauge customer perceptions of personalization efforts. Asking questions about transparency, control, and perceived value can provide valuable insights.

A high NPS, indicating that customers are willing to recommend the business, is a strong signal of ethical personalization success. It suggests that customers not only appreciate the personalized experience but also trust the business’s approach. This is akin to customers enthusiastically recommending the bakery to their friends, a testament to positive word-of-mouth built on trust and satisfaction.

Opt-In and Preference Management Rates

Think of customers actively choosing to sign up for the bakery’s newsletter or customize their pastry preferences. Opt-in rates for personalization features and active use of preference management Meaning ● Preference Management: Strategic SMB discipline orchestrating anticipatory, ethical, hyper-personalized experiences for loyalty and growth. tools indicate customer buy-in and empowerment. High opt-in rates suggest that customers perceive value in personalization and trust the business to use their data responsibly. Active use of preference management tools demonstrates that customers feel in control and are engaging with personalization on their own terms.

SMBs should track these metrics to understand customer adoption and engagement with personalization features. These rates are like observing customers actively participating in the bakery’s custom order system, a sign of engagement and trust in the personalization process.

Complaint and Churn Rates Related to Personalization

Imagine a sudden drop in bakery customers complaining about unwanted ads or feeling their preferences are ignored. Conversely, a spike in complaints about intrusive personalization or unwanted data collection is a red flag. Complaint and churn rates directly related to personalization efforts can highlight ethical failures. Increased complaints about privacy concerns, unwanted recommendations, or lack of control signal that personalization is becoming intrusive or manipulative.

Higher churn rates among customers exposed to personalization, compared to those who are not, might indicate that personalization is negatively impacting customer relationships. SMBs need to monitor these rates closely and investigate any negative trends. This is like the bakery noticing a sudden increase in customers complaining about pushy sales tactics or feeling their orders are misunderstood, a sign of ethical personalization gone wrong.

Qualitative Feedback and Sentiment Analysis

Consider listening to customer conversations in the bakery, noting their reactions to recommendations. Quantitative metrics provide numbers, but qualitative feedback offers context and depth. Analyzing customer reviews, social media comments, and direct feedback for sentiment related to personalization provides a richer understanding of customer perceptions. Positive sentiment, expressing appreciation for relevant recommendations or personalized service, indicates ethical personalization success.

Negative sentiment, expressing concerns about privacy, manipulation, or lack of control, highlights areas for improvement. SMBs should actively solicit and analyze qualitative feedback to complement quantitative metrics. This is like the baker engaging in conversations with customers, understanding their nuanced reactions to personalized suggestions, and gaining a deeper understanding of their preferences and concerns.

Ethical personalization success for SMBs is not a destination but a continuous journey. It requires a shift in mindset from simply maximizing engagement to prioritizing customer well-being and building trust. By focusing on transparency, control, value exchange, fairness, and measuring success through ethical metrics, SMBs can harness the power of personalization to create stronger customer relationships and achieve sustainable growth. It’s about building a bakery where the aroma of trust is as enticing as the freshly baked bread.

Navigating Ethical Personalization Complexities

The initial allure of personalization metrics Meaning ● Personalization Metrics for SMBs: Quantifiable measures reflecting tailored customer experiences, driving growth and loyalty. often centers on immediate gains ● increased click-through rates, boosted conversion percentages, and a superficial sense of enhanced customer engagement. Yet, a deeper examination reveals a more intricate landscape. Consider the statistic that while 71% of consumers express some level of frustration with impersonal shopping experiences, a nearly identical percentage, 74%, become equally disillusioned when personalization becomes overly intrusive. This paradox underscores a critical point ● ethical personalization is not merely a technical challenge, but a strategic tightrope walk, demanding a sophisticated understanding of customer expectations and ethical boundaries.

Beyond Basic Metrics ● A Multi-Dimensional Approach

Traditional metrics, while providing a snapshot of immediate impact, often fail to capture the longitudinal effects of personalization strategies. They are akin to measuring the immediate sugar rush from a pastry without considering the long-term health implications. Ethical personalization success necessitates a shift towards a multi-dimensional approach, incorporating metrics that assess not just short-term gains, but also long-term customer relationships, brand reputation, and sustainable business practices. This expanded view requires businesses, especially SMBs aspiring to scale, to move beyond simplistic engagement metrics and embrace a more holistic evaluation framework.

Ethical personalization, when viewed through a long-term lens, becomes a strategic asset, fostering customer loyalty Meaning ● Customer loyalty for SMBs is the ongoing commitment of customers to repeatedly choose your business, fostering growth and stability. and enhancing brand value far beyond immediate transactional gains.

Data Privacy and Security Metrics

In an era defined by data breaches and privacy anxieties, metrics related to data privacy Meaning ● Data privacy for SMBs is the responsible handling of personal data to build trust and enable sustainable business growth. and security become paramount. These are not merely compliance checkboxes, but fundamental indicators of ethical personalization practices. Consider the Equifax breach, which cost the company millions and severely damaged its reputation. For SMBs, even smaller-scale breaches can be devastating.

Metrics should include incident response times, data breach frequency, security audit scores, and customer data access logs. Furthermore, tracking customer inquiries and concerns related to data privacy provides valuable qualitative data. Proactive monitoring and transparent reporting on data security measures build customer confidence and mitigate potential ethical pitfalls. These metrics are analogous to a bakery rigorously adhering to food safety standards ● unseen by most customers, but crucial for building trust and ensuring long-term viability.

Algorithmic Transparency and Explainability Metrics

The “black box” nature of many personalization algorithms raises ethical concerns. Customers are increasingly wary of opaque systems that dictate their experiences without explanation. Metrics that assess algorithmic transparency Meaning ● Algorithmic Transparency for SMBs means understanding how automated systems make decisions to ensure fairness and build trust. and explainability become essential. This includes tracking the clarity and accessibility of explanations provided to customers about personalization logic.

For example, “Why am I seeing this ad?” features, common in many platforms, can be evaluated for their effectiveness and user engagement. Internally, metrics should focus on documenting algorithm design, bias detection processes, and model interpretability scores. Openness about how personalization algorithms function, even at a high level, fosters trust and addresses concerns about manipulation or unfair practices. This transparency is akin to a baker openly explaining the ingredients and baking process ● demystifying the creation and building customer confidence.

Customer Perception of Value and Reciprocity Metrics

Ethical personalization hinges on a balanced value exchange. Metrics must go beyond simply measuring business gains and assess customer perception Meaning ● Customer perception, for SMBs, is the aggregate view customers hold regarding a business's products, services, and overall brand. of value and reciprocity. This involves tracking metrics like perceived personalization relevance scores, customer feedback on the usefulness of recommendations, and engagement with personalized content versus generic content. Surveys can directly ask customers ● “Do you feel personalization enhances your experience?” and “Do you believe the value you receive is fair in exchange for your data?” Analyzing customer language in reviews and social media for sentiment related to value and reciprocity provides further insights.

A positive perception of value and reciprocity indicates that personalization is ethically aligned and mutually beneficial. This is analogous to customers feeling they receive good value for their money at the bakery ● a fair exchange that encourages repeat business.

Fairness and Bias Mitigation Metrics (Advanced)

Building upon the fundamental understanding of fairness, intermediate metrics delve deeper into bias detection and mitigation. This involves not just recognizing potential biases, but actively measuring and reducing them. Metrics should include bias detection rates across different demographic groups, fairness scores for personalization algorithms (e.g., disparate impact scores), and the effectiveness of bias mitigation Meaning ● Bias Mitigation, within the landscape of SMB growth strategies, automation adoption, and successful implementation initiatives, denotes the proactive identification and strategic reduction of prejudiced outcomes and unfair algorithmic decision-making inherent within business processes and automated systems. techniques implemented. Regular audits of personalization systems for fairness, conducted by independent third parties, can provide an objective assessment.

Tracking customer complaints related to perceived unfairness or discrimination is also crucial. Proactive bias mitigation and transparent reporting on fairness metrics demonstrate a commitment to ethical personalization. This is akin to a bakery ensuring its sourcing practices are fair and ethical ● considering the broader societal impact Meaning ● Societal Impact for SMBs: The total effect a business has on society and the environment, encompassing ethical practices, community contributions, and sustainability. of its operations.

Long-Term Customer Relationship Metrics

Ethical personalization is not about short-term transactional gains, but about building enduring customer relationships. Intermediate metrics should focus on long-term relationship health. This includes tracking customer loyalty over extended periods (e.g., 2-year or 5-year CLTV), repeat purchase rates, and customer advocacy metrics (e.g., referral rates). Analyzing customer journey maps to understand how personalization impacts the overall customer experience over time provides valuable insights.

Metrics should also assess the resilience of customer relationships in the face of potential personalization missteps. Do customers forgive occasional errors if they perceive a genuine commitment to ethical practices? Long-term relationship metrics reveal the true impact of ethical personalization on sustainable business growth. This is analogous to a bakery building a multi-generational customer base ● relationships that endure over time, built on trust and consistent quality.

Integrating Ethical Metrics into SMB Growth Strategy

For SMBs aiming for scalable growth, integrating ethical personalization metrics Meaning ● Ethical Personalization Metrics for SMBs: Measuring personalization success responsibly, respecting privacy, and building trust for sustainable growth. into their overall business strategy is crucial. This requires moving beyond viewing ethics as a separate compliance function and embedding it into the core of business operations. Ethical metrics Meaning ● Ethical Metrics, in the context of SMB growth, automation, and implementation, refer to a system of quantifiable measurements designed to evaluate a business's adherence to ethical principles. should be incorporated into key performance indicators (KPIs) and regularly reviewed at strategic decision-making levels. This means not just tracking sales and marketing metrics, but also monitoring customer trust Meaning ● Customer trust for SMBs is the confident reliance customers have in your business to consistently deliver value, act ethically, and responsibly use technology. scores, data privacy metrics, and fairness indicators.

SMBs should invest in tools and technologies that facilitate ethical personalization measurement and reporting. This might include privacy-enhancing technologies, bias detection software, and customer feedback management systems. Furthermore, fostering a company culture that prioritizes ethical considerations in personalization is essential. Training employees on ethical personalization principles and empowering them to make ethical decisions is crucial for long-term success.

Ethical personalization, when strategically integrated, becomes a competitive differentiator, attracting and retaining customers who value trust and integrity. It’s about building a bakery that not only serves delicious pastries but also embodies ethical values, attracting customers who appreciate both quality and integrity.

Navigating the complexities of ethical personalization requires a sophisticated understanding of metrics that extend beyond superficial engagement. By embracing a multi-dimensional approach, incorporating data privacy, algorithmic transparency, customer perception, fairness, and long-term relationship metrics, SMBs can move beyond basic personalization tactics and build sustainable, ethical growth strategies. The tightrope walk becomes less precarious when guided by a comprehensive and ethically grounded metric framework.

The Algorithmic Mandate Ethical Personalization in the Age of Automation

The contemporary business milieu, particularly for Small and Medium-sized Businesses (SMBs), is increasingly defined by the imperative of automation. Personalization, once a bespoke, human-driven endeavor, is now largely algorithmically mediated. This shift, while offering unprecedented scalability and efficiency, introduces a new stratum of ethical complexity.

Consider the statistic that by 2025, AI-driven personalization is projected to influence 95% of all customer interactions. This ubiquity necessitates a rigorous re-evaluation of ethical personalization metrics, moving beyond reactive compliance to proactive, anticipatory frameworks that address the inherent ethical challenges of automated systems.

From Reactive to Proactive Ethics ● Anticipatory Metric Frameworks

Traditional ethical considerations in personalization often operate reactively, addressing ethical lapses after they occur. In the age of automation, this approach is insufficient. The speed and scale of algorithmic decision-making demand proactive, anticipatory metric frameworks. These frameworks must not only measure current ethical performance but also predict potential ethical risks and vulnerabilities embedded within automated personalization Meaning ● Automated Personalization for SMBs: Tailoring customer experiences using data and technology to boost growth and loyalty, ethically and efficiently. systems.

This requires incorporating metrics that assess algorithmic bias proactively, simulate potential ethical dilemmas, and evaluate the robustness of ethical safeguards under diverse operational conditions. For SMBs leveraging automation for growth, this anticipatory approach is not merely a philosophical exercise but a pragmatic necessity for mitigating reputational risks and ensuring long-term sustainability. It is akin to a bakery not just reacting to food safety incidents but proactively implementing rigorous quality control measures to prevent them from occurring in the first place.

Ethical personalization in the age of automation demands a paradigm shift from reactive compliance to proactive anticipation, embedding ethical considerations into the very fabric of algorithmic design and deployment.

Algorithmic Auditing and Explainability Metrics (Advanced)

Building upon intermediate transparency metrics, advanced frameworks necessitate robust algorithmic auditing Meaning ● Algorithmic auditing, in the context of Small and Medium-sized Businesses (SMBs), constitutes a systematic evaluation of automated decision-making systems, verifying that algorithms operate as intended and align with business objectives. and explainability metrics. This extends beyond surface-level explanations to deep, granular analysis of algorithmic decision pathways. Metrics should include algorithmic audit trail depth, decision provenance tracking, and complexity-adjusted explainability scores. Furthermore, metrics must assess the accessibility and comprehensibility of algorithmic explanations for diverse stakeholders, including non-technical users and regulatory bodies.

Advanced techniques like SHAP (SHapley Additive exPlanations) values and LIME (Local Interpretable Model-agnostic Explanations) can be employed to quantify feature importance and decision influence within complex personalization models. Regular, independent algorithmic audits, utilizing these advanced metrics, are crucial for maintaining ethical accountability in automated personalization systems. This is analogous to a bakery undergoing rigorous, independent food safety audits ● ensuring not just compliance but also demonstrating a commitment to the highest standards of quality and safety.

Autonomous Ethical Agents and AI Alignment Metrics

The future of personalization may involve autonomous ethical agents embedded within AI systems, capable of making real-time ethical judgments. This necessitates the development of AI alignment metrics, assessing the degree to which AI agents’ goals and behaviors align with human ethical values. Metrics should evaluate AI agent value alignment scores, ethical decision consistency, and robustness to ethical adversarial attacks. Furthermore, metrics must assess the transparency and explainability of AI agents’ ethical reasoning processes.

Research in AI ethics and value alignment provides a theoretical foundation for developing these advanced metrics. For SMBs exploring AI-driven personalization, understanding and incorporating AI alignment metrics will be crucial for navigating the ethical frontier of autonomous systems. This is akin to a bakery employing AI-driven systems for inventory management and customer service, but ensuring these systems are aligned with the bakery’s core values of customer satisfaction and ethical sourcing.

Dynamic Consent and Preference Evolution Metrics

Traditional consent models, often static and one-time, are ill-suited for the dynamic nature of automated personalization. Advanced ethical frameworks require dynamic consent Meaning ● Dynamic Consent, in the SMB sphere, represents a method of obtaining and managing user permissions for data processing, offering individuals granular control and transparency. and preference evolution metrics. This involves tracking metrics like consent decay rates, preference drift analysis, and proactive consent re-engagement rates. Furthermore, metrics must assess the effectiveness of mechanisms for customers to dynamically adjust their personalization preferences in real-time, based on evolving needs and contexts.

Techniques like contextual bandits and reinforcement learning can be employed to personalize consent requests and preference management interfaces, optimizing for both ethical compliance and user experience. Dynamic consent metrics ensure that customer autonomy is respected in the ongoing, iterative process of automated personalization. This is analogous to a bakery offering customers flexible subscription options and allowing them to easily adjust their orders and preferences based on their changing needs.

Societal Impact and Systemic Bias Metrics

Ethical personalization extends beyond individual customer interactions to encompass broader societal impacts and systemic biases. Advanced metrics must assess the potential for personalization systems to exacerbate existing societal inequalities or create new forms of discrimination at scale. This involves tracking metrics like demographic disparity amplification rates, societal fairness impact scores, and vulnerability mapping for algorithmic harms. Furthermore, metrics should assess the effectiveness of interventions designed to mitigate systemic biases and promote equitable outcomes across diverse societal groups.

Interdisciplinary research, drawing from fields like sociology, ethics, and computer science, is crucial for developing these complex metrics. For SMBs operating in diverse markets, understanding and addressing societal impact metrics is essential for responsible and sustainable growth. This is akin to a bakery considering the broader societal impact of its sourcing practices and ensuring it contributes to a fair and equitable food system.

Ethical Personalization as a Corporate Social Responsibility (CSR) Metric

In the advanced business landscape, ethical personalization transcends mere compliance and becomes a core component of Corporate Social Responsibility Meaning ● CSR for SMBs is strategically embedding ethical practices for positive community & environmental impact, driving sustainable growth. (CSR). Metrics should reflect this strategic integration, assessing the contribution of ethical personalization practices Meaning ● Ethical personalization for SMBs means building customer trust and sustainable growth by respecting privacy and providing value. to overall CSR performance and brand reputation. This involves tracking metrics like ethical personalization CSR scores, stakeholder perception of ethical commitment, and positive brand association metrics linked to ethical personalization initiatives. Furthermore, metrics must assess the return on investment (ROI) of ethical personalization, demonstrating its contribution to long-term business value and competitive advantage.

Transparent reporting on ethical personalization CSR metrics enhances brand trust and attracts ethically conscious customers and investors. For SMBs seeking to differentiate themselves in competitive markets, ethical personalization as a CSR metric provides a powerful strategic advantage. This is analogous to a bakery prominently featuring its ethical sourcing and sustainability practices in its marketing and branding ● attracting customers who value both quality and social responsibility.

The algorithmic mandate of the automation age demands a sophisticated and proactive approach to ethical personalization metrics. Moving beyond reactive compliance to anticipatory frameworks, incorporating advanced algorithmic auditing, AI alignment, dynamic consent, societal impact, and CSR metrics, SMBs can navigate the ethical complexities of automated personalization and build sustainable, responsible, and ultimately more successful businesses. The bakery of the future, powered by AI and automation, must not only bake delicious goods but also embody ethical principles at its algorithmic core, ensuring that every interaction, every recommendation, and every automated process reflects a deep commitment to customer well-being and societal good.

References

• O’Neil, Cathy. Weapons of Math Destruction ● How Big Data Increases Inequality and Threatens Democracy. Crown, 2016.
• Zuboff, Shoshana. The Age of Surveillance Capitalism ● The Fight for a Human Future at the New Frontier of Power. PublicAffairs, 2019.
• Mittelstadt, Brent Daniel, et al. “The ethics of algorithms ● Current landscape, future directions.” Big Data & Society, vol. 3, no. 2, 2016, pp. 1-21.
• Goodman, Bryce, and Seth Flaxman. “EU GDPR and Fairness in Machine Learning.” arXiv preprint arXiv:1708.00963, 2017.
• Shapley, Lloyd S. “A value for n-person games.” Contributions to the Theory of Games, vol. 2, no. 28, 1953, pp. 307-17.
• Ribeiro, Marco Tulio, Sameer Singh, and Carlos Guestrin. ““Why Should I Trust You?” ● Explaining the Predictions of Any Classifier.” Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, ACM, 2016, pp. 1135-44.