Nigerian retailers using AI plus IoT to personalize green engagement report 25–40% higher customer loyalty and measurable cuts in energy and waste; however, improvements are observationally linked to AI-driven insights rather than established through randomized evaluation.
AI and Iot-Based Customer Behaviour Analysis for Business Enhancement in Nigeria
A 12-month Nigerian study finds that firms using AI combined with IoT to tailor green messaging and recommendations are associated with 25–40% higher customer loyalty and measurable reductions in energy, waste and carbon footprint.
Corporate sustainability has evolved from a strategic initiative to a fundamental business imperative, prompting firms to explore new ways to understand and influence customers' purchasing decisions regarding greener products as they strive to maintain their edge in increasingly digital markets where expectations are evolving. This chapter demonstrates how Nigerian companies can utilize artificial intelligence in conjunction with Internet of Things tools to sift through complex streams of customer data in real-time and align their green actions with what shoppers perceive as environmentally friendly. The study employs stratified random sampling across urban shopping centers, suburban retail outlets, and online-to-offline hybrid stores in Nigeria, representing diverse consumer demographics and shopping behaviors. Data collection encompasses retail kiosks, shopping apps, home sensors, and wearables over twelve months. The authors apply machine-learning models, natural language processing, sentiment scoring, predictive dashboards, and clustering techniques to map customer preferences, purchasing patterns, and green program participation. Data analysis combines quantitative analytics with qualitative sentiment analysis, while environmental impact data is collected through IoT sensors measuring energy consumption, waste generation, and carbon footprint metrics. Businesses implementing these insights demonstrate a 25 to 40% increase in loyalty while reducing their ecological footprint through tailored green messages, smarter product suggestions, and targeted eco-marketing aligned with shoppers' values. The article contributes to sustainable change literature by demonstrating that insight-driven engagement drives profit while advancing environmental goals. Results underscore that firms must incorporate data-driven analysis into their sustainability plans to gain actionable insights and develop customer strategies that boost profits while enhancing ecological responsibility.
Summary
Main Finding
AI combined with IoT-enabled sensing and analytics can meaningfully improve customer engagement and sustainability outcomes in Nigerian retail. In the study’s deployments (n = 2,187 final respondents across urban/suburban/O2O stores), firms that used the AI+IoT insights achieved a 25–40% increase in customer loyalty while reducing ecological footprints via targeted green messaging, personalized recommendations, and operational adjustments informed by real‑time environmental monitoring.
Key Points
- Purpose and framing
- Examines how AI + IoT can map customer preferences and link behaviour to sustainability outcomes in Nigerian retail.
- Theoretical basis: integrated Technology Acceptance Model (TAM) and Theory of Planned Behaviour (TPB), extended for environmental concern.
- Sample and setting
- Stratified random sampling across retail environments: urban shopping centres (40%, n≈880), suburban outlets (35%, n≈770), O2O hybrid stores (25%, n≈550).
- Final analytic sample: 2,187 valid responses (99.4% response rate), geographically dispersed across six Nigerian zones.
- Observed effects
- Reported empirical gains at adopters: 25–40% increase in loyalty metrics and measurable reductions in store-level environmental indicators.
- IoT-enabled monitoring informed energy, waste and carbon-reduction interventions; dashboards supported campaign optimization and personalization.
- Technical stack (overview)
- IoT sensors: energy meters, air quality, waste and water sensors, proximity sensors, smart carts, mobile beacons.
- Data sources: POS, mobile apps, social media, loyalty databases, e‑commerce, customer service logs.
- Machine learning tools: NLP (sentiment), computer vision (behaviour), MLP (sales prediction), Random Forest (segmentation), SVM (sentiment classification), LSTM (time series), association rule mining, clustering (K-means) and recommender systems.
- Evaluation metrics: MSE, MAPE, Accuracy/Precision/Recall/F1, AUC, Silhouette score.
- Qualitative analytics
- Sentiment analysis using VADER, BERT classification, facial-expression analysis, LDA topic modeling to surface green preferences and messaging resonance.
- Ethics & limitations
- Data protection: anonymisation, opt-in consent, encrypted transmission, audits.
- Limitations: potential selection bias from partner retailers, 12‑month window, regional heterogeneity in tech adoption, IoT standardization challenges.
Data & Methods
- Design: mixed-methods, longitudinal deployment over 12 months in three phases (IoT baseline; intensive collection/analytics; validation/optimization).
- Sampling: Cochran formula, planned N≈2,200 to allow attrition; stratified across retail types and six geographic zones.
- IoT deployment: store-level environmental sensors + customer-interaction sensors to capture behavioral and sustainability metrics in real time.
- ML pipeline: Data preprocessing → feature engineering (demographics, purchase history, environmental preferences, tech adoption, spatio-temporal features) → model training → validation → deployment.
- Models and analytic purposes:
- Predictive forecasting: MLP, LSTM.
- Segmentation: Random Forest + clustering (RFM + K-means).
- Sentiment & text analytics: SVM, BERT, VADER, topic models.
- Association rules: basket analysis to find green-product co-purchases.
- CV: behavioral observation in-store.
- Environmental metrics:
- Carbon footprint computed via CF = Σ(Ai × EFi); energy efficiency via EEI = (Energy Output / Energy Input) × 100.
- Visualization & delivery: real-time dashboards (Tableau/Power BI) integrating behavioural heatmaps, sustainability scorecards, and campaign performance.
- Validation: model performance assessed by standard ML metrics and operational validation during Phase 3.
Implications for AI Economics
- Firm-level economics
- Value capture: measurable increases in loyalty and targeted conversion suggest meaningful gains to customer lifetime value (CLV) from AI+IoT investments.
- Cost offsets: energy/waste reductions and inventory/forecasting improvements can lower operating costs — important when modeling ROI on analytics platforms.
- Adoption heterogeneity: benefits likely skew toward larger or digitally mature retailers able to absorb upfront sensor and analytics costs; SMEs may face adoption barriers without subsidies or low-cost offerings.
- Competition & market structure
- Data advantage: early adopters can develop superior personalization and inventory efficiency, potentially increasing concentration if data-driven advantages are not diffused.
- Platform effects: integration with payment and mobile channels (mobile-first environment) can amplify network benefits and lock-in.
- Welfare, externalities & regulation
- Positive externalities: reduced carbon and waste are social benefits not fully captured in private returns — potential rationale for public support or carbon-crediting mechanisms tied to verified IoT monitoring.
- Greenwashing risk: AI-personalized green messages increase persuasion power; credible verification (sensor-backed metrics, third-party audits) will be critical for truthful claims and for welfare-preserving regulation.
- Privacy trade-offs: economic models should account for consumer privacy valuations and potential behavioral shifts under stricter data-protection regimes.
- Research & measurement suggestions for economists
- Causal identification: pursue RCTs or staggered rollouts (diff-in-diff, synthetic controls) to estimate causal effects on sales, loyalty, and environmental outcomes.
- Structural modeling: quantify adoption costs, dynamic CLV gains, and pricing responses to personalized green offers.
- Cost‑benefit accounting: monetize energy/waste reductions and include social value of emissions avoided to assess public policy interventions (subsidies, tax credits).
- Distributional analysis: evaluate which consumer segments and firm sizes benefit most, and how labor (store staff, analytics teams) are affected.
- Market-level effects: study competition, entry/exit, and concentration effects from widespread analytics adoption.
- Practical metrics & datasets to collect (recommended)
- Firm metrics: incremental CLV, churn, conversion rates, average basket value, inventory turnover, operating margins.
- Environmental metrics: kWh saved, waste volume reduced, CO2e reduction per store/customer.
- Adoption costs: fixed sensor/IT investment, per-store maintenance, analytics staffing/training costs.
- Privacy/consent metrics: opt-in rates, data retention, opt-out behavior.
- Policy implications
- Consider incentives for SME adoption (grants, shared sensor networks).
- Standardize measurement and third-party verification of IoT-based sustainability claims to reduce information asymmetries.
- Balance data‑use regulations to preserve consumer privacy while allowing socially beneficial analytics (e.g., anonymized aggregate reporting for environmental monitoring).
Recommended next steps for researchers and policymakers: run randomized or phased deployments with pre-specified environmental and economic endpoints, collect granular cost data for ROI estimates, and design verification protocols tying retailer claims to sensor data to inform subsidies or carbon-credit schemes.
Assessment
Paper Typecorrelational
Evidence Strengthlow — Findings rely on observational associations and descriptive ML outputs; the reported 25–40% loyalty gains and environmental reductions are not supported by a randomized or clearly controlled identification strategy, leaving results vulnerable to selection bias, confounding (e.g., early-adopter firms, concurrent initiatives), and measurement/validation concerns.
Methods Rigormedium — The study uses a reasonably rigorous data-collection design (stratified sampling, multi-source IoT and app data over 12 months) and modern analytics (ML, NLP, clustering), but the paper does not document causal-design elements, robustness checks, model validation, sample sizes, or out-of-sample performance in enough detail to rate methods as high rigor.
SampleStratified random sample of consumers and shopping events across urban shopping centers, suburban retail outlets, and online-to-offline hybrid stores in Nigeria; data sources include retail kiosks, shopping apps, home sensors and wearables collected over 12 months, with consumer demographics and shopping behaviors represented across these channels (exact N not reported).
Themesadoption innovation
IdentificationObservational analysis using stratified random sampling across retail environments; machine-learning, NLP, clustering and sentiment scoring to map preferences; firms applied insights and outcomes were compared (apparently pre/post or cross-sectional) without randomized assignment or clear quasi-experimental controls.
GeneralizabilityCountry-specific: results come from Nigeria and may not transfer to other cultural or regulatory contexts, Retail-focused: study centers on consumer retail and hybrid stores, not B2B or other sectors, Digital access bias: reliance on shopping apps, wearables and home sensors may over-represent digitally connected and wealthier consumers, Potential early-adopter bias: firms implementing insights may differ systematically from typical firms, Time-limited: 12-month window may not capture longer-term behavioral or environmental effects
Claims (8)
| Claim | Direction | Confidence | Outcome | Details |
|---|---|---|---|---|
| Businesses implementing these insights demonstrate a 25 to 40% increase in loyalty. Firm Revenue | positive | high | customer loyalty |
25 to 40% increase
0.15
|
| Businesses reduced their ecological footprint through tailored green messages, smarter product suggestions, and targeted eco-marketing aligned with shoppers' values. Consumer Welfare | positive | high | ecological footprint (energy consumption, waste generation, carbon footprint) |
0.3
|
| Nigerian companies can utilize artificial intelligence in conjunction with Internet of Things tools to sift through complex streams of customer data in real-time and align their green actions with what shoppers perceive as environmentally friendly. Adoption Rate | positive | high | alignment of firm green actions with shopper perceptions/preferences |
0.3
|
| The study employed stratified random sampling across urban shopping centers, suburban retail outlets, and online-to-offline hybrid stores in Nigeria to represent diverse consumer demographics and shopping behaviors. Adoption Rate | null_result | high | sampling representativeness / coverage of consumer demographics and shopping behaviors |
0.3
|
| Data collection encompassed retail kiosks, shopping apps, home sensors, and wearables over twelve months. Other | null_result | high | data collection scope and duration |
0.5
|
| The authors applied machine-learning models, natural language processing, sentiment scoring, predictive dashboards, and clustering techniques to map customer preferences, purchasing patterns, and green program participation. Task Allocation | null_result | high | mapping of customer preferences, purchasing patterns, and program participation |
0.5
|
| Data analysis combined quantitative analytics with qualitative sentiment analysis, while environmental impact data was collected through IoT sensors measuring energy consumption, waste generation, and carbon footprint metrics. Other | null_result | high | integrated analytics approach and environmental metrics collection |
0.5
|
| Insight-driven engagement drives profit while advancing environmental goals; firms must incorporate data-driven analysis into their sustainability plans to gain actionable insights and develop customer strategies that boost profits while enhancing ecological responsibility. Firm Revenue | positive | high | profitability and ecological responsibility (dual outcome) |
0.3
|