Data Analytics in Stockholm, Sweden

Fintech / Payments / Data Analytics / Stockholm

A Stockholm-based buy-now-pay-later (BNPL) fintech (SEK 28B in transaction volume annually, 12M active consumers across the Nordics, DACH, and UK, 85,000 merchant partners, 1,400 employees, licensed as a payment institution under Finansinspektionen) needed a merchant intelligence platform to transition from a consumer-facing BNPL provider to a comprehensive merchant commerce platform. The company had built its business on consumer credit analytics (real-time credit decisioning for point-of-sale financing) but had underinvested in merchant-side analytics — the side of the marketplace that generated revenue through merchant fees. Core challenges: (1) Merchant profitability blindness — the company knew total revenue per merchant (transaction fees, typically 3-6 percent of transaction value) but not true merchant profitability. Merchant costs included: consumer credit losses attributable to that merchant's customers, customer service costs for the merchant's transactions, integration support costs, and marketing co-investment. An estimated 22 percent of merchants were unprofitable — generating more in credit losses and support costs than they paid in fees. But without merchant-level profitability analytics, the company couldn't identify which merchants to grow, which to reprice, and which to exit. The sales team was incentivised on merchant acquisition (number of new merchants) rather than merchant quality — leading to aggressive onboarding of merchants in categories with high fraud and high credit loss rates. (2) Merchant churn prediction — the company losing 18 percent of merchants annually through churn. In a competitive market (Klarna, Affirm, PayPal Credit, Afterpay all competing for the same merchants), churn was expensive: each churned merchant representing SEK 12-18 months of customer acquisition cost that was never recovered. Churn signals were invisible: a merchant reducing BNPL share from 15 percent to 5 percent of checkout volume (indicating they were testing a competitor) generated no alert. By the time the merchant formally switched providers, the relationship was unrecoverable. The company needed churn prediction that identified at-risk merchants 60-90 days before churn — enough time for relationship intervention. (3) Consumer credit analytics for merchant underwriting — the company's consumer credit models were mature (real-time scoring with 99.2 percent accuracy) but weren't used to evaluate merchant quality. Some merchants attracted high-risk consumers — not through the merchant's fault, but because certain product categories, price points, and customer demographics had inherently different risk profiles. The company needed merchant-level credit risk analytics: "merchants in this category with this average order value in this market have a credit loss rate of X percent — price the merchant fee accordingly." Without this, the company was cross-subsidising: low-risk merchants (electronics retailers with creditworthy customers) paying the same fee structure as high-risk merchants (fashion retailers with younger, thinner-credit-file customers) — making the company uncompetitive for the best merchants and over-attractive for the worst. (4) Merchant growth analytics — the company wanting to expand wallet share at existing merchants (increasing BNPL share of checkout from the average 8 percent to 15-20 percent). But the company had no analytics on: what drove BNPL adoption at checkout (placement, messaging, payment flow design), how BNPL availability affected merchant conversion rates and average order values, or which merchant categories and consumer segments had the highest BNPL propensity. The merchant success team making recommendations based on general best practices rather than data-driven, merchant-specific insights — "move the BNPL option higher in checkout" rather than "A/B testing shows that for your customer demographic, presenting BNPL as the default payment method with a 4-installment preview increases adoption 34 percent." (5) Regulatory and CSRD requirements — Finansinspektionen increasing scrutiny of BNPL companies: consumer creditworthiness assessment requirements, responsible lending obligations, and capital adequacy reporting. The company needed analytics demonstrating: aggregate credit quality across the portfolio, concentration risk by merchant category and geography, and stress test results showing portfolio resilience under adverse scenarios. CSRD: the company's sustainability reporting requiring analytics for: carbon footprint of operations, responsible lending metrics (consumer outcomes), diversity and inclusion data, and governance transparency. Swedish investor expectations for ESG data quality exceeding minimum regulatory requirements.

We built a merchant intelligence platform over 10 weeks — providing merchant-level profitability, churn prediction, credit-risk-based pricing, growth analytics, and regulatory compliance. (1) Merchant profitability engine: true economics at the merchant level. Cost allocation: calculating actual cost-to-serve for each of 85,000 merchants — consumer credit losses allocated to the originating merchant (using the company's existing credit models to attribute default probability to merchant characteristics), customer service costs allocated by ticket volume per merchant, integration support costs, and marketing co-investment. Merchant P&L: monthly profitability statement for every merchant, segmented by: gross revenue (transaction fees), net revenue (after credit losses, service costs), contribution margin, and lifetime value projection. Results: 24 percent of merchants confirmed as unprofitable (versus 22 percent estimate) — but the distribution was actionable: 8 percent were new merchants still in investment phase (acceptable), 6 percent were repriceable (fee increase would make them profitable), 5 percent were fixable (credit policy adjustments reducing losses), and 5 percent were structural money-losers (exit candidates). Portfolio optimisation: analytics informing a strategic portfolio review — the company developing tiered merchant strategies: "grow" merchants (profitable and expanding, receiving enhanced support and co-marketing), "optimise" merchants (marginally profitable, receiving targeted improvement recommendations), and "manage" merchants (unprofitable, receiving fee restructuring or managed exit). (2) Churn prediction: 90-day early warning system. Behavioural signals: machine learning models monitoring 42 merchant behaviour features including: BNPL share of checkout volume (declining share indicating competitor testing), transaction frequency (decreasing velocity), average order value shifts, customer complaint rates, integration API call patterns, and merchant dashboard login frequency. Competitor intelligence: monitoring merchant websites for competitor BNPL presence — detecting when merchants added Klarna, Afterpay, or other BNPL options alongside the company's offering (a strong pre-churn signal). Prediction accuracy: the model identifying merchants at high churn risk (above 70 percent probability) with 82 percent accuracy, 90 days before formal churn. This created an intervention window: relationship managers receiving weekly "at-risk merchant" reports with specific intelligence on what was driving the risk. Intervention playbook: analytics-driven intervention strategies — different approaches for different churn drivers: "competitor price pressure" (fee renegotiation), "poor BNPL performance" (checkout optimisation support), "integration issues" (technical support acceleration), "business model change" (product adaptation). (3) Credit-risk-based merchant pricing: risk-adjusted fees. Merchant risk profiling: credit analytics adapted from consumer scoring to merchant-level risk assessment — calculating expected credit loss rate for each merchant based on: merchant category, average order value, customer demographic profile, geographic market, and historical loss experience from similar merchants. Dynamic pricing: merchant fees adjusted based on credit risk profile — low-risk merchants (electronics, home goods with creditworthy customers) receiving competitive fees that made the company attractive versus alternatives, while high-risk merchants (fast fashion with younger demographics) paying fees that reflected the true cost of credit risk. Risk-reward alignment: the pricing model ensuring that every merchant relationship was economically sound — removing the cross-subsidy where good merchants overpaid to compensate for losses on bad merchants. This made the company more competitive for its best merchants (reducing churn risk) while appropriately pricing risk on higher-risk merchants. New merchant underwriting: prospective merchants receiving instant fee quotes based on their category, market, and predicted risk profile — enabling the sales team to quote accurately rather than offering standard rates that might be unprofitable. (4) Merchant growth intelligence: data-driven adoption optimisation. Checkout analytics: A/B testing framework for BNPL presentation at checkout — testing: button placement, messaging (instalment preview vs. "pay later" vs. "split in 4"), and default payment method selection. Results segmented by merchant category and customer demographic, providing merchant-specific optimisation recommendations rather than generic advice. Conversion impact: analytics measuring the actual impact of BNPL on merchant performance — "BNPL availability increases your conversion rate by 12 percent and average order value by 22 percent, generating SEK 8.4M in additional annual revenue for your business." These merchant-specific impact numbers provided by relationship managers during renewal negotiations. Category intelligence: benchmarking each merchant against category peers — "your BNPL adoption rate is 6 percent vs. category average of 11 percent — here are the specific checkout optimisations that top-performing merchants in your category have implemented." Seasonal optimisation: analytics identifying seasonal patterns in BNPL adoption by merchant category — fashion merchants seeing 40 percent higher BNPL usage during pre-Christmas (November-December) while electronics merchants peaked during Black Friday and back-to-school. Seasonal merchandising recommendations for BNPL prominence. (5) Regulatory and sustainability analytics: Finansinspektionen and CSRD compliance. Portfolio risk reporting: automated quarterly reporting to Finansinspektionen — portfolio credit quality metrics, concentration analysis (by merchant category, geography, consumer segment), and stress test results. Consumer outcomes: responsible lending analytics demonstrating that BNPL products led to positive consumer outcomes — repayment rates, default rates compared to alternative credit products, and consumer satisfaction metrics. CSRD sustainability: carbon footprint analytics for operations (office, travel, data centre), responsible lending metrics (consumer welfare indicators), and governance transparency data — meeting Swedish investor expectations for comprehensive ESG reporting. Audit readiness: all regulatory analytics with complete audit trail — data sources documented, calculation methodologies transparent, and results reproducible.

Results over 12 months. Merchant profitability: portfolio optimisation generating SEK 180M in annual profit improvement — driven by: fee restructuring for 6 percent of merchants (SEK 42M), credit policy adjustments reducing losses at 5 percent of merchants (SEK 68M), managed exit of 5 percent structural loss-makers (SEK 35M avoided losses), and enhanced growth at profitable merchants (SEK 35M incremental revenue). Churn reduction: merchant annual churn from 18 percent to 11 percent — retaining 5,950 merchants that would have otherwise churned. Estimated lifetime value preservation: SEK 890M. Intervention success rate: 58 percent of predicted-to-churn merchants retained through analytics-driven intervention (versus 12 percent retention rate before predictive analytics). Pricing: risk-adjusted pricing improving competitive position with low-risk merchants — winning 340 new merchants from competitors in Year 1 by offering lower fees than competitors who were still cross-subsidising. High-risk merchant pricing reducing credit loss rate by 14 percent (merchants self-selecting for better credit practices when fees reflected risk). Merchant growth: average BNPL share of checkout from 8 percent to 12.4 percent across the portfolio — a 55 percent improvement. Top-quartile merchants reaching 18 percent BNPL share through checkout optimisation. Incremental merchant revenue from BNPL growth: SEK 420M in additional transaction volume generating SEK 21M in additional fees. Regulatory: full Finansinspektionen compliance with zero examination findings. CSRD reporting completed 6 months ahead of regulatory deadline. Swedish investor ESG expectations met — contributing to successful SEK 2.4B fundraising round where data governance and responsible lending analytics were specifically cited by investors. Financial impact: profitability improvement SEK 180M; churn prevention SEK 89M (annual value, conservative); competitive merchant wins SEK 34M; BNPL growth SEK 21M; credit loss reduction SEK 68M. Total annual impact: SEK 392M (approximately EUR 34M). Development: SEK 8.2M (approximately EUR 720K). Annual platform: SEK 2.8M (approximately EUR 245K). ROI: 47.8x first year.