Available for E-Commerce Audit Engagements

Your Margin Leak Is Already Sitting in Your Data.

Built for $500K to $20M D2C and Shopify brands, not pre-revenue startups or enterprise retailers with in-house data teams.

Most e-commerce businesses don't lose profit to bad products or weak marketing. They lose it to discount erosion, dead SKUs, customer concentration, and stockout risk sitting unread inside their own data. I find it, quantify it, and hand you a decision, not a report.

$643K Profit Exposure Surfaced
61% Discount Dependency Diagnosed
3.2x Champion vs. Average Customer Value
Book a 15-Minute Revenue Diagnostic View Case Studies
Zeeshan Akram
Audits Completed
7 Case Studies
BI Dashboards
3+ Delivered
Scroll

About

Where profit actually leaks, and how I find it before it shows up as a quarter of disappointing revenue

Analytical Philosophy

E-Commerce Profit Analyst

Your margin leak is already sitting inside your transaction data. You just haven't found it yet. Most e-commerce businesses aren't losing profit because of bad products or weak marketing. They're losing it because the evidence, discount erosion, dead SKUs, customer concentration, stockout risk, is sitting unread inside their own data.

I specialise in finding it. Every audit traces profit leakage across four areas: pricing behaviour, product-level profitability, customer concentration, and inventory risk, then translates the evidence into a small number of clear, financially quantified decisions. Every engagement ends in one language: revenue recovered, margin protected, operational risk eliminated.

0+

Case Studies

0+

Certifications

4

Profit Leak Categories

Where Profit Actually Leaks

Pricing: Discount models quietly compressing margin from 54% to 49%, affecting 61% of all transactions.
Products: 8 products in a $35M catalog driving 84.5% of profit, with $643K in exposure sitting on a single point of failure.
Customers: Top 20% of customers generating 77.2% of revenue, Champions worth 3.2x average lifetime value.
Inventory: A single SKU carrying 40% of category revenue with critical stockout exposure and no buffer.

Analytical Toolkit

Commercial Analysis
Margin AuditingChurn DiagnosticsRFM SegmentationCLV ModellingCohort Analysis
Data Engineering & Querying
Advanced SQLPostgreSQLMySQLWindow FunctionsCTEs
Business Intelligence
Power BIDAXPower QueryExcel (Advanced)
Data Manipulation & Statistical Validation
PythonPandasNumPySciPyHypothesis Testing

Commercial Analytics Audits

End-to-end investigations into real business performance problems, from raw data to boardroom-ready conclusions

These audits are built on public datasets from Kaggle and the UCI Machine Learning Repository, plus Microsoft's Contoso synthetic retail dataset, chosen for realistic scale and complexity. The same methods apply directly to your live business data.

The Profit Leak Diagnostic, DataCo Supply Chain Catalog
Profit Audit View on GitHub
Flagship Case Study
Case Study 01 Public Dataset, Kaggle

The Profit Leak Diagnostic

A full profitability and operational risk audit of a $35.2M e-commerce catalog spanning 180,519 transactions. Built across SQL data architecture, Python statistical validation, and a five-page Power BI dashboard, the investigation locates exactly which products, shipping tiers, and pricing habits are quietly converting strong revenue into weak or negative profit.

Just 8 products generate 84.5% of all profit, while the other 110 are operational noise riding on their back.
28 products combine weak margins with high delivery failure while producing only 2.2% of total revenue.
First Class shipping arrives late 100% of the time, while the cheapest tier in the catalog is also the most reliable.
A modeled 20% margin shock to the top 8 products alone would wipe out $643,000 in profit, nearly 17% of total company profit.
PostgreSQLSQLPythonSciPyPandasPower BI
Audit Conclusion
Revenue is not the issue. Mistaking revenue for business health is. Profit in this catalog depends on 8 products performing perfectly, indefinitely, with no backup plan if any of them stumble.
Diagnostic Report (PDF) Slide Deck (PDF) Dashboard Walkthrough
See If Your Catalog Has the Same Risk
Retail Discount Profitability Audit, Contoso Synthetic Retail Dataset
Margin Audit
Case Study 02 Contoso Synthetic Dataset

Retail Discount Profitability Audit

A multi-layer commercial audit investigating whether a retailer's discount strategy functions as a pricing lever or a structural margin destructor. Spanning SQL data architecture, Python behavioural analysis, and a Power BI audit narrative, the investigation surfaces exactly where discounting accelerates volume while quietly compressing profit.

PostgreSQLSQLPython PandasPower BI
Audit Conclusion
61% of transactions were sold at a discount, with each discounted line delivering $97 less profit than its full-price counterpart. A full pricing policy review was recommended before any further volume-led growth targets were set.
Read Executive Report (PDF)
Check Your Own Discount Dependency
Olist eCommerce RFM Analysis
Retention Analysis
Case Study 03 Public Dataset, Kaggle

Olist eCommerce: Churn Root Cause & RFM Investigation

Root-cause investigation across 113K+ Brazilian eCommerce orders revealing that logistics failure, not pricing or product, was the dominant churn driver. Delivery delays directly correlated with zero repeat purchases, reframing the problem from a product question to an operations intervention requiring immediate fulfilment process changes.

PythonPandasSeabornRFM Analysis
Business Finding
Logistics failure, not price or product, was the dominant driver of churn. Delivery delays directly correlated with one-star reviews and zero repeat purchases.
Read Executive Report (PDF) View Dashboard Layout (PDF)
UK Retail CLV & RFM Dashboard
CLV Modelling
Case Study 04 Public Dataset, UCI

UK Retail: Customer Lifetime Value & Segment Migration

Analysed £17M+ in revenue across 1M+ UK retail transactions to surface the true concentration of customer value. Built a custom Power Pivot relational model with Power Query ETL to separate VIP behaviour from stagnant cohorts and quantify the revenue opportunity hidden in segment migration.

Advanced ExcelPower Pivot DAXPower Query
Business Finding
Champion customers were worth 3.2x the platform average, and up to 40x an At-Risk customer (£8,767 vs £220). That gap revealed a £390K+ incremental revenue opportunity achievable purely through cohort upgrade, no new acquisition required.
Watch 3-Min Model Walkthrough
Professional Standards & Workflow Signals
GDPR / Data Privacy Conscious
Documented SQL Audit Trails
Git-Versioned Analysis
No Raw Data Retained Post-Engagement
Business-First Reporting Standard
Async-First, US/UK Timezone Compatible
Analytical Finding, Contoso Discount Audit

"The logic applied in this audit successfully isolated a $97-per-line profit gap that was previously hidden in aggregate revenue reporting. By separating discounted from full-price transactions at the order line level, the analysis surfaced a structural margin compression, from 54% to 49%, driven entirely by unmanaged discounting habit, not market pressure. This finding was invisible at the category level and would have remained so under standard dashboard reporting."

Zeeshan Akram, Retail Discount Profitability Audit, Contoso Synthetic Dataset (2026)

What Most Founders Miss

Common patterns where standard reporting obscures the underlying business problem

Revenue grows while profit stagnates

Top-line growth can mask structural margin compression from discounting, customer mix shifts, or rising acquisition costs that don't appear in headline revenue figures.

Discounting drives volume, not value

High discount rates often inflate order counts while simultaneously training customers to never pay full price, creating retention that is entirely price-dependent and operationally fragile.

Churn is attributed to the wrong cause

The actual driver is frequently operational: delivery failures, fulfilment inconsistencies, or logistics performance, and it never surfaces in product analytics.

The top 20% of customers are obscured by averages

Aggregated CLV and ARPU figures hide extreme concentration. Knowing that 18% of your customer base generates 74% of profit changes every acquisition and retention decision you make.

Cohort behaviour is never isolated

What looks like a catalogue problem is often a single acquisition period or campaign cohort with structurally poor fulfilment performance.

Margin is measured at category, not line level

Category-level margin reporting allows loss-making SKUs to hide behind strong performers. Profit destruction is almost always concentrated in specific product lines that aggregate data cannot isolate.

Recognise two or three of these in your own numbers?

Book a 15-Minute Diagnostic Call

My Analytical Process

Every engagement follows the same rigorous framework, from raw question to boardroom-ready recommendation.

1
Define the
Business Problem

Translate a vague business pain into a precise, measurable analytical question.

Problem Statement
2
Collect &
Clean Data

ETL pipelines, deduplication, and outlier handling to build a trustworthy single source of truth.

Clean Dataset
3
Analyze &
Model

SQL joins, Python aggregations, RFM scoring, and cohort breakdowns, digging until a signal emerges.

Key Findings
4
Visualize &
Communicate

Power BI dashboards and Seaborn charts designed for a non-technical executive audience.

Live Dashboard
5
Recommend
& Quantify

Every output ends with a £/$ figure: a specific revenue impact, not a vague "consider improving."

Revenue Impact

The last step is non-negotiable. Any analysis that doesn't end with a quantified business recommendation is just a data exercise, not a decision tool. Every case study in my portfolio closes with a specific £ or $ figure for this reason.

Technical Foundation

Machine Learning (Andrew Ng, DeepLearning.AI & Stanford) · NLP: Classification & Vector Spaces (DeepLearning.AI) · Python for Data Science, AI & Development (IBM) · Data Analysis with Python (IBM) · Data Visualization with Python (IBM) · Databases & SQL for Data Science (IBM) · Crash Course on Python (Google)

Diagnostic Services

Three fixed-scope entry points. One clear output: a quantified commercial finding and an evidence-backed recommendation you can act on.

Entry Point 01

The Profitability Audit

A full-stack review of your transaction data to locate exactly where margin compresses between revenue recognition and net profit, across product lines, discount tiers, customer segments, and geographies. Built for operators who see top-line growth but declining bottom-line performance.

  • SQL-documented audit trail
  • Python behavioural analysis
  • Margin leakage quantified by segment
  • Power BI executive narrative
Book a 15-Minute Revenue Diagnostic
Entry Point 02

Retention & Churn Deep-Dive

Identifying exactly when and why customers stop buying: isolating the lifecycle stage, root cause (operational, product, or pricing), and the precise revenue-at-risk figure that justifies intervention. Outputs a cohort-level breakdown, not a generic churn rate.

  • Cohort-level churn breakdown
  • RFM segmentation & LTV modelling
  • Revenue-at-risk quantification
  • Retention intervention roadmap
Book a 15-Minute Revenue Diagnostic
Entry Point 03

Executive Dashboard Build

Turning messy sales or operational data into a clean Power BI or Excel command centre, built around the business questions that actually drive decisions, not vanity metrics. Designed for founders who need one trusted source of commercial intelligence.

  • Power BI / Advanced Excel dashboard
  • Dynamic YoY time-intelligence
  • Cross-filtered executive views
  • Full drill-down capability
Book a 15-Minute Revenue Diagnostic

Common Questions

What founders usually ask before booking a first call

Is my business data safe with you?

Yes. I work with read-only exports or view-only database access, never full admin access to your store or systems. An NDA is available and signed before any data changes hands, and raw data is deleted once the engagement is complete unless you ask me to retain it.

What if the audit doesn't find anything meaningful?

In practice this hasn't happened. Every business with real order volume has margin sitting in blind spots, in specific SKUs, shipping tiers, or discount habits, because standard dashboards report averages, not concentration. If a scoped audit genuinely turns up nothing actionable, you still keep the clean data model and dashboard built along the way.

How is this different from hiring an agency?

You work directly with me from the first call to the final report, not an account manager relaying findings from a team you never speak to. That means faster turnaround, no markup for layers of overhead, and someone who actually understands your data rather than a templated audit checklist.

How long does a typical engagement take?

Most fixed-scope diagnostics take 5 to 10 business days from data access to final report, depending on data volume and cleanliness. I'll confirm a specific timeline on the discovery call once I understand your setup.

Do I need to be technical to understand the output?

No. Every engagement ends with a plain-English report and a walkthrough call, not a raw dashboard you're left to interpret alone. The technical work (SQL, Python, the modelling) stays behind the scenes. What you get is a ranked list of findings in dollars, and what to do about each one.

What size of business is a good fit?

Most engagements work best for D2C and Shopify brands roughly in the $500K to $20M revenue range. Pre-revenue startups usually don't have enough transaction history for the analysis to be meaningful yet, and larger enterprise retailers typically already have an in-house data team for this kind of work.

Analytical Insights

Published breakdowns of real commercial analytics investigations, written for founders and operators, not data scientists.

E-commerce Churn Analysis Python Code
Churn Investigation

Why 97% of Customers Never Returned: A Root-Cause Analysis of 113K eCommerce Orders

A complete Python analytics case study on the Olist dataset, covering data engineering, memory optimisation, and RFM segmentation to determine whether a near-total churn rate was a product, pricing, or operational failure, and what that distinction means for intervention strategy.

Read on Medium
17M Retail Data Analysis Excel Dashboard
Revenue Concentration Analysis

The 40x Customer Multiplier: How RFM Segmentation Surfaced £390K Hidden in Plain Sight

A breakdown of how processing 1M+ retail transactions via Power Query and a custom RFM segmentation model revealed that a small VIP cohort was generating up to 40x the lifetime value of at-risk customers, and what a migration strategy could realistically recover.

Read on Medium

Start a Conversation

Available for analytics consulting engagements, profitability audits, and commercial analysis partnerships

Book a 15-Minute Revenue Diagnostic

Discuss a Business Problem

If your business is experiencing margin pressure, unexplained churn, or revenue growth that isn't translating to profit, that is exactly the kind of problem worth discussing. Reach out to explore whether an analytics engagement would surface the answer.

Twitter / X

@ZeeshanAkram_DS