Wale Shekoni

Fintech unicorn processing 50M+ transactions (about £600 million) a day for 20M+ personal banking customers and 5M+ POS terminals in Nigeria. Reporting to the Head of Growth and the CMO.

• Growth levers: with monthly actives 14% behind plan after Q1, broke the target into sign-ups, conversion and early retention and sized each lever, showing that conversion and early activation were the cheapest to move. Case study →
• Onboarding: identified first-30-day use of everyday payments (airtime, data, bills) as the strongest predictor of retention and revenue, through a cohort study across 13 monthly cohorts. The company rebuilt onboarding around it (live August 2026): MAU growth rose from 2–4% to 8–9% a month, sign-up-to-account conversion +3–4pp and new-user VAS adoption +8–9pp. Cohorts →
• Forecasting: for the CMO, built the 2025 transaction-volume forecast in Prophet from daily card and POS data since 2022; it was taken to the CEO and adopted as the plan baseline, and landed within 1% of the full-year actual. Case study →
• Campaign measurement: for the CMO, measured rewards and TV campaigns with CausalImpact: daily actives +17–19%, volume +16–22%, but no change in frequency, so the next campaign was redesigned around frequency. Reconciled Singular attribution against the warehouse. Case study →
• Customer health: piloted an early-warning score in one region for the SVP of Distributed Network Sales, flagging top business customers falling from their own recent peak for weekly relationship-manager review; delivered customer-360 rankings of the top 10,000 businesses per region.
• Churn or seasonality: when transaction volume fell behind forecast for five straight weeks, the CMO asked for the cause and a remedy. Tested every lever behind the total (forecast by transaction type, retention, churn, frequency, regions against two prior years), ruled out churn and set a dated test for a seasonal rebound. Volume recovered as predicted, so a planned re-engagement campaign was not needed. Case study →
• Diagnostics: for the CEO, showed NPS is a weak churn signal once survey selection is accounted for; for the SVP of Distributed Network Sales, modelled terminal payback after the naira devaluation and mapped terminal density by local government area, flagging 20 states where terminals were serviced from another state; quantified the effect of KYC enforcement on top customers. NPS → Payback → Mapping →
• Verification segmentation: mapped KYC-tier mix by age, gender, region, onboarding channel and occupation against digital benchmarks, showing where verification upgrades lag. Higher tiers carry materially more revenue per customer.
• Acquisition economics: modelled digital acquisition cost per active customer against cost per signup.
• Experiment design: designed (only partly run) a six-experiment activation and reactivation programme for senior leadership: three-arm tests separating the effect of messaging from the effect of incentives, tier-capped cashback, pre-registered success thresholds and gaming controls, plus nine delivery tests (reward floor, push and email frequency, framing, channel, ML targeting) with a full business case.
• Data foundations: rebuilt a multi-year monthly agent summary table (hundreds of millions of rows) with a partitioned, restartable backfill. Built Airflow pipelines from ClickHouse to Google Sheets and Prophet forecasts with macroeconomic regressors.

Commuter carpooling platform. After the September 2026 relaunch, in six weeks: 670+ drivers signed up (79 fully verified), 370+ riders onboarded, 64 routes priced and published.

• Defined 49 KPIs across acquisition, supply, demand, retention and finance, and launched a 25-page analytics site with cohort funnels and weekly retention grids.
• Co-designed the data model with the data engineering team (231 tables across financial and operational PostgreSQL databases) and specified the daily reconciliation checks for its double-entry ledger.
• Designed an A/B/n experimentation framework for onboarding nudges, with sticky assignment, a holdout group and power-based sample sizing.
• For the co-founders, created the Trip Fare Pricing Model from research into fares, journey times and running costs, calibrated by a rule I proposed (four passengers ≈ 1.1× one ride-hailing fare); it prices every published route. Case study →
• Built driver–passenger corridor matching on search data to surface unmet demand and guide driver recruitment.
• Analysed commute intent for the relaunch cohort: destinations converge on the Lagos Island cluster, two-thirds commute five or more days a week, and driver sign-ups outran riders about two to one. The driver onboarding funnel showed email verification as the biggest drop-off. Case study →
• Set up governed BI access: a read-only analytics role with query timeouts, a PII review gate, internal-account exclusion and versioned SQL with shared metric definitions.
• Modelled per-customer analytics tables of lifecycle milestones, with an hourly reconciler that corrects drift against operational data. It fixed drivers who were being nudged in error.
• Validated the migration of legacy users from Firebase to PostgreSQL, with every reconciliation check passing.

Data lead on a two-month client market study; reported to the IFS line manager and presented findings to the client's management team.

• Turned an open research brief into a dashboard specification (metrics, filters, drill-downs) and delivered the dashboard in Looker Studio. Case study →
• Sized the market from transaction and survey data, correcting for multi-platform users; built behavioural segments and a two-year forecast with weekly and yearly seasonality.
• Analysed seven years of exchange data (2018 to mid-2025): account opening and daily, weekly and monthly active users; assets tiered from stablecoins to emerging tokens; rule-based user segments (traders, savers, cash-out users); and deposit and withdrawal flows by asset tier and region.

• Modelled blackout probabilities for electricity networks under climate-change scenarios in MATLAB, combining meteorological data with network models.
• Worked with the research team to analyse complex climate and infrastructure datasets and interpret climate-model outputs.

• Built customer-targeting models that contributed to a 4pp increase in customer retention; segmented customers with K-means on geocoded transactions for marketing; developed supervised and unsupervised fraud-detection models.
• Owned data cleaning, exploration and hyperparameter tuning to improve model accuracy.

• Built live streaming data pipelines with Apache Spark and Kafka, pre-processed structured and unstructured data, and assessed the accuracy of new data sources.
• Led the development team through the redesign and deployment of the beta websites for mobile and tablet, running scrum ceremonies and mentoring the team in agile practices.

• Analysed core-banking (Temenos T24) data and produced executive reports on trends, patterns and forecasts.
• Core member of the 2018 core-banking (T24) migration team, resolving post-migration issues; handled the 2020 IT audit.
• Administered core-banking databases and servers, supported almost 1,000 staff, and trained staff on the core-banking and identity-enrolment platforms.
• Worked with management to prioritise business and information needs, and recommended system and data-governance improvements.