Marketing Intelligence · Nairobi, Kenya · Global

Your data is already
telling a story.
Read it.

Advanced marketing analytics, customer intelligence and predictive modelling for e-commerce brands, growing businesses and nonprofits across Africa and beyond. Real numbers. Real decisions.

$722K+
Revenue Analysed
5
Consulting Projects
87%
Churn Model Accuracy
44x
Email ROAS Achieved
Core Capabilities
  • Marketing Mix Modelling (MMM)Advanced
  • Attribution Model ComparisonMulti-model
  • Customer Churn PredictionML
  • Cohort Retention AnalysisSQL
  • Customer Segmentation (RFM + ML)Clustering
  • A/B Testing & ExperimentationStats
  • Power BI Executive DashboardsBI
  • NGO Impact ReportingM&E
Positioning
Not another agency. A focused analytics team. Full-stack capability. No overhead.
Python · SQL Power BI · Tableau GA4 · GTM E-commerce Predictive Models NGO & Impact East Africa Focus Remote-First

Specialized intelligence for
modern businesses.

We partner with teams that have outgrown basic reporting and need rigorous, actionable data models to drive strategy.

E-commerce Brands
Identify high-LTV customers, optimize multi-channel attribution, and predict churn before it happens.
SaaS & Startups
Understand product usage patterns, optimize the acquisition funnel, and measure true feature impact.
NGOs & Impact
Translate complex field data into transparent, compelling impact reports for donors and stakeholders.
Agencies
White-label advanced analytics capabilities to offer your clients deeper insights and proven ROI.

Full-stack data solutions.

From raw database extraction to executive dashboards and predictive models, we handle the entire pipeline. No hand-offs. No lost context.

01 / Marketing Analytics

Campaign intelligence and channel performance

Most marketing reporting tells you what happened. This pillar tells you why — and which channels deserve next month's budget. We build attribution models, campaign analysis and funnel diagnostics that connect your ad spend to actual revenue outcomes rather than platform-reported clicks.

Marketing Attribution Modelling

Python GA4 SQL
  • Last Click, Linear, Time Decay & Shapley models side-by-side
  • Revenue share comparison across all model types
  • Budget reallocation with dollar-impact quantification
  • Platform attribution reconciliation (Google Ads vs actual)

Marketing Mix Modelling (MMM)

Python Statsmodels Ridge SciPy
  • Adstock transformation per channel
  • OLS + Ridge regression with full diagnostics
  • Channel contribution decomposition + ROAS per channel
  • Constrained budget optimisation with revenue uplift projection

Web Analytics & Traffic Intelligence

GA4 GTM Looker Studio
  • Full GA4 audit and reporting setup
  • Traffic source analysis by revenue contribution
  • Session quality and funnel drop-off by channel
  • Conversion rate analysis end-to-end

Campaign Performance Analysis

Google Ads Meta Ads Power BI
  • ROAS, CPL and CPA by channel, campaign and creative
  • Cross-channel performance benchmarking
  • Ad fatigue detection and creative refresh scheduling
  • Automated weekly performance reporting

Funnel Analytics & CRO

GA4 GTM Python
  • Funnel drop-off analysis by device, channel and segment
  • A/B test design, execution and statistical analysis
  • Checkout abandonment root cause diagnosis
  • Conversion rate optimisation opportunity identification

Content & Social Media Analytics

Python Instagram TikTok
  • Engagement rate analysis and content performance benchmarking
  • Organic vs paid performance comparison
  • Audience growth and reach trend analysis
  • Content calendar optimisation based on performance data
02 / Customer Intelligence

Who your customers are, what they do, and when they leave

Aggregate revenue numbers hide the customer dynamics underneath them. This pillar builds the systems that let you see individual and group-level behaviour — segmentation that drives personalisation, retention analysis that predicts churn before it happens, and lifetime value modelling that tells you which customers are worth acquiring.

Customer Segmentation (RFM + ML)

Python K-Means Power BI
  • Recency, Frequency and Monetary segmentation
  • K-Means clustering for behavioural segments beyond RFM
  • Segment-level revenue contribution and reactivation opportunity sizing
  • Actionable segment profiles with campaign recommendations

Cohort Retention Analysis

SQL Python Seaborn
  • 5-view SQL pipeline — BigQuery and Snowflake ready
  • Retention heatmaps tracking M0 through M12 per cohort
  • Revenue-weighted cohort value analysis
  • Identification of high-value acquisition months and channels

Customer Lifetime Value Modelling

Python SQL Excel
  • Historical LTV calculation by segment and acquisition channel
  • LTV-to-CAC ratio by channel for budget allocation decisions
  • Payback period analysis for acquisition spend
  • Predicted LTV for new customer cohorts

Retention Strategy Analysis

Python SQL Power BI
  • Churn rate by segment, channel and product category
  • Win-back campaign opportunity sizing
  • Repeat purchase interval analysis
  • Post-purchase journey mapping to identify drop-off points
03 / Predictive Analytics & Modelling

What is going to happen, and what should you do about it

Predictive models move analytics from explaining the past to influencing the future. Whether that is a machine learning churn model that flags at-risk customers before they cancel or a demand forecast that tells finance what to expect next quarter — this is where analytics becomes directly measurable in revenue saved or earned.

Churn Prediction Modelling

XGBoost Scikit-learn SHAP
  • XGBoost and Logistic Regression churn models
  • SHAP feature importance — understand why customers leave
  • Customer risk scoring for CRM integration
  • Intervention ROI calculation — cost of saving vs losing a customer

A/B Testing & Experimentation

Python SciPy Statsmodels
  • Test design — power analysis and sample size calculation
  • Statistical significance testing (t-test, chi-square, Mann-Whitney)
  • Bayesian A/B analysis for early decision making
  • Full stakeholder report with business impact and recommendations

Demand Forecasting

Python Prophet ARIMA
  • Time-series forecasting for revenue, orders and inventory
  • Seasonality decomposition and calendar effect modelling
  • Scenario-based projections: conservative, base and growth
  • Confidence interval outputs for planning and finance teams

Predictive Lifetime Value

Python Scikit-learn SQL
  • Predicted LTV for new customer cohorts
  • Segment-level LTV distribution analysis
  • Budget allocation recommendations by LTV band
  • LTV-to-CAC tracking over time
04 / Dashboards & Reporting

The right number, in the right hands, at the right time

Analysis that lives in a notebook is analysis that does not get acted on. This pillar turns model outputs and raw data into live dashboards, automated reports and self-service tools that your team uses every week — not only when an analyst is in the room.

Power BI Executive Dashboards

Power BI DAX SQL
  • KPI overview with revenue, orders, ROAS and customer metrics
  • Channel attribution and campaign performance pages
  • Customer segmentation and retention visualisations
  • Drill-through to product, campaign and cohort level

Looker Studio & Google Dashboards

Looker Studio GA4 BigQuery
  • GA4-connected live dashboards updated daily
  • Cross-platform data blending (GA4 + Sheets + Ads)
  • Shareable, embeddable reports for stakeholders
  • Automated email delivery schedules

Excel & Sheets Intelligence Systems

Excel Google Sheets Python
  • KPI trackers with formula validation and automated alerts
  • Sales pipeline and CRM trackers
  • Financial performance dashboards with scenario modelling
  • Templates built to be maintained by non-technical teams

Automated Reporting Pipelines

Python SQL Pandas
  • Weekly and monthly report generation from raw data sources
  • Data cleaning and transformation pipelines
  • Scheduled exports and stakeholder distribution
  • Alert systems for KPI threshold breaches

What clients say.

Don't just take our word for it — hear directly from the teams we've worked with.

"PA Data Analytics helped us understand that our social ad spend was completely cannibalizing our organic traffic. We shifted budget and immediately saw a 32% lift in overall ROAS."

↑ 32% ROAS improvement
SM
Sarah M.
CMO, E-commerce Brand

"The churn prediction model was a game-changer. We can now proactively reach out to accounts before they cancel. It paid for itself in the first month."

↓ 18% Churn reduction
JK
James K.
Founder, SaaS Startup

Templates & Resources.

Accelerate your analytics journey with our plug-and-play templates.

Marketing Attribution Dashboard Template
A plug-and-play Tableau dashboard template for multi-channel attribution analysis.
Buy for .00

Data transparency for
a better world.

We partner with nonprofits and NGOs to build scalable M&E reporting frameworks, ensuring stakeholders and donors see exactly where their impact is felt.

Impact Dashboards
Automated Power BI & Tableau dashboards replacing manual PDF reports.
Survey Data Pipelines
Cleaning and structuring messy field data (KoboToolbox, ODK) for instant analysis.
Donor Transparency Portals
Public-facing or private web portals showcasing real-time metrics.
Advisory
For smaller teams needing direction.
Consulting

The team behind
the numbers.

Patience Anono — Lead Data Analyst
Patience Anono
Lead Data Analyst

We started PA Data Analytics because we noticed a gap: businesses were buying expensive tools but still couldn't answer basic questions about their marketing ROI or customer retention.

We don't just build dashboards. We build the pipelines that feed them, the statistical models that power them, and the strategies that make them useful.

Full-Stack
From SQL extraction to Python ML models to final presentation.
Commercial Focus
Analytics designed to move the needle, not just look pretty.
  • Based in Nairobi, Kenya. Working with clients globally.
  • Certified in Advanced Google Analytics & Data Engineering.

Let the data tell you
something useful.

Most businesses are sitting on enough data to make significantly better decisions. The question is whether anyone has looked at it properly. Let's find out.

Send us a message