The six areas that determine whether your e-commerce business is leaking profit — and how to examine each one in two to three weeks, not six months.
Most e-commerce businesses I work with know their margins aren’t where they should be. The CEO can feel it. The CFO can see it in the P&L. But nobody can point to exactly where the money is going, because the costs are spread across dozens of line items, multiple departments, and a tangle of SaaS subscriptions, supplier contracts, and operational decisions that were made years ago and never revisited.
That’s what a margin audit fixes. Not by producing a 200-page consulting report that sits on a shelf, but by systematically pulling apart your cost structure, identifying where you’re overspending relative to what you should be, and building a prioritized plan to fix it.
I’ve run about fifty of these audits across e-commerce businesses ranging from £5M to £200M in revenue. The specifics vary, but the shape of the problem is remarkably consistent. Here’s how to do it properly.
A margin audit is not a financial audit. Your accountants are already handling that. It’s an operational examination of every cost that sits between your revenue and your EBITDA, with the specific goal of identifying costs that can be reduced without damaging the business.
The key word is “operational.” We’re not looking at the numbers from a distance. We’re getting into the contracts, the vendor invoices, the tech stack, the warehouse processes, the marketing channel data, the org chart. The P&L tells you what happened. The audit tells you why.
A good audit takes two to three weeks of focused work. I’ve seen consultancies stretch this into six-month engagements, billing by the hour while they slowly work through discovery. That’s unnecessary. If you know what to look for and you have access to the data, the picture comes together fast. The hard part isn’t finding the problems — it’s prioritizing which ones to fix first.
This is not something you can do in isolation. An external consultant staring at spreadsheets for three weeks will miss context that only the leadership team has. Why was that vendor chosen? What happened when you tried to renegotiate last year? Which team is already at breaking point?
The audit works best as a collaborative process with the CEO, CFO, and COO (or whoever owns operations). I typically run it as a series of working sessions — one per area — where we go through the data together, I ask a lot of uncomfortable questions, and we jointly assess what’s fixable. The leadership team walks out of the audit not just with a list of improvements, but with a fundamentally different way of thinking about their cost structure. That’s the part that lasts.
If you’re running this yourself without outside help, the same principle applies. Get your department heads in a room. Make it clear this isn’t about blame — it’s about finding money. Most people are surprisingly willing to flag inefficiencies when they know the goal is improvement, not headcount cuts.
Curious what your margin opportunity looks like?
Free Tool
How much margin are you leaving on the table?
Answer 6 questions. Get a personalised margin estimate in under 2 minutes.
Take the Free Margin AuditStart here. COGS is usually the single largest line item, and it’s the one most e-commerce operators have the least visibility into at a granular level.
Pull your product-level margin data. Not category averages — individual SKU margins. You’re looking for products that are selling well but making you almost nothing after landed cost, and products where the margin looks healthy but the supplier pricing hasn’t been reviewed in years.
Red flags: supplier contracts that auto-renew without renegotiation, landed costs that don’t account for the true cost of freight and duties, products where you’re absorbing tariff increases that should have been passed through in pricing, and any situation where you can’t tell me the exact margin on your top 50 SKUs within ten minutes. If you have to go digging, your COGS visibility is a problem in itself.
The fix is usually a combination of supplier renegotiation, better landed cost modelling, and feeding accurate margin data into your pricing decisions. Most businesses are pricing based on markup from cost price, not from true landed cost. The difference can be several points of margin.
The average e-commerce business I audit is running 30 to 60 SaaS tools. About a third of them are genuinely necessary. Another third are doing something useful but overpriced for what they deliver. The final third are zombie subscriptions — tools that someone signed up for eighteen months ago, used for a project, and forgot about.
Pull every recurring technology cost. Every SaaS subscription, every platform fee, every API charge, every hosting bill. Map each one to who owns it, what it does, and what the contract terms are. This exercise alone, before any optimizations, typically surfaces 10–15% in savings just from cancelling things nobody is using.
Beyond the zombie hunt, look at your infrastructure costs. If you’re on AWS or GCP, there’s almost certainly significant savings available through reserved instances, right-sizing, and architectural changes. We’ve seen businesses cut cloud costs by 50% through a structured migration and optimization programme.
Red flags: annual SaaS contracts that renew in different months (making it impossible to get a full picture at any one time), enterprise-tier subscriptions where you’re using 20% of the features, and any tool where the answer to “what would happen if we cancelled this tomorrow?” is a long pause followed by “I’m not sure.”
One thing I’ve been doing more recently is looking at which SaaS tools can be replaced entirely with AI agents. A lot of the middleware and automation tools that e-commerce businesses pay for monthly can now be replicated with custom agents at a fraction of the cost. This isn’t theoretical — we’ve done it repeatedly.
Fulfilment is death by a thousand cuts. No single cost looks outrageous, but the aggregate is often 15–25% of revenue, and there are savings buried in almost every sub-process.
Look at your cost per order fulfilled — the fully loaded number including pick, pack, packaging materials, and carrier costs. Break it down by order type (single item vs. multi-item, standard vs. express). Compare it to benchmarks for your category and volume tier.
If you’re using a 3PL, pull apart their invoices line by line. Most 3PL contracts have dozens of charge types, and most e-commerce businesses couldn’t tell you what half of them are for. Storage charges, management fees, value-added service fees, peak surcharges — every one of these is negotiable if you have data showing what the market rate should be.
We wrote up the specifics of how we approached this in a recent engagement in the warehouse optimization deep-dive. The short version: 14 operational changes, £1M in annualized savings, zero capital expenditure.
Red flags: 3PL contracts that haven’t been renegotiated since signing, carrier rates that are based on projected volumes you never hit, returns processing costs that nobody tracks separately, and any situation where the ops team can’t tell you the fully loaded cost to fulfil an average order.
I’m not here to tell you how to run your marketing. But I am here to tell you that most e-commerce businesses are spending 20–40% more on customer acquisition than they need to, because they’re measuring the wrong things and optimizing for the wrong goals.
The audit focuses on three questions. First, what is your blended customer acquisition cost, and how has it trended over the last 12 months? If it’s going up and nobody can explain why, that’s a problem. Second, what percentage of your marketing spend is going to customers who would have bought anyway? This is the hardest question in marketing, but if you’re not even asking it, you’re definitely overspending. Third, are you suppressing promotions and discounts from customers who don’t need them? We’ve written about discount suppression separately, but the short version is that most e-commerce businesses are giving away margin to customers who were already going to convert at full price.
Red flags: an attribution model that gives 100% credit to the last click, a marketing team that can’t tell you the payback period on paid acquisition by channel, retargeting spend that’s never been incrementality-tested, and blanket discount codes that are available to everyone rather than targeted to customers who genuinely need a nudge.
This is the sensitive one, so let me be direct about how I approach it: the goal is not to cut heads. The goal is to understand whether people are spending their time on work that actually drives value, or whether they’re trapped doing manual tasks that should be automated.
Map your org chart against your key business processes. Where are people spending their time? How much of it is repetitive operational work versus strategic work? I consistently find that 30–40% of time in e-commerce operations teams is spent on tasks that can be partially or fully automated — data entry, report generation, manual order processing, vendor communication, inventory reconciliation.
The right move is usually not to fire those people. It’s to deploy AI agents to handle the repetitive work and redeploy the humans to higher-value activities. The business gets more done without growing headcount, and the team gets to do more interesting work. When done well, it’s one of the few genuinely win-win outcomes in cost optimization.
Red flags: teams where the headcount has grown linearly with revenue (suggesting no operational efficiency gains), roles that exist primarily to move data between systems, and any department where the manager says they need more people but can’t clearly articulate what those people would do that couldn’t be automated.
Pricing is the most powerful margin lever and the most neglected. Most e-commerce businesses set prices based on a fixed markup from cost, adjust them manually when someone remembers to, and then wonder why their margins are volatile.
The audit looks at your pricing methodology, your competitive positioning, your promotional cadence, and your markdown strategy. How often do prices change? What triggers a price change? Who decides? Is there any systematic use of demand signals, competitor pricing data, or margin targets in pricing decisions?
In most cases, the answer is no. Prices are set when products are bought, adjusted during sales events, and otherwise left alone. This leaves enormous value on the table. Dynamic pricing that responds to demand, inventory levels, and competitive positioning can move gross margin by 2–5 points without any impact on conversion rates. We’ve seen an AI pricing engine deliver 77% revenue uplift on tested categories.
Red flags: a single markup percentage applied across all categories, prices that haven’t changed in months on products with fluctuating input costs, no systematic markdown strategy for slow-moving inventory, and promotional pricing that doesn’t account for the true margin impact after factoring in returns.
At the end of a margin audit, you should have three things.
First, a margin bridge: a clear, quantified view of where your money goes between gross revenue and EBITDA, broken down into the six areas above, with each area benchmarked against what “good” looks like for your category and scale.
Second, an opportunity register: a prioritized list of every margin improvement initiative identified, each with an estimated impact (in pounds, not percentages), an implementation difficulty rating, a timeline, and a clear owner. In the audits I run, this typically contains 40–80 individual initiatives. In a recent engagement, we tracked 119 EBITDA-tracked initiatives through to completion.
Third, a 90-day action plan: the top 10–15 initiatives that should be started immediately, sequenced to account for dependencies and resource constraints. This is the document that actually drives change. The margin bridge explains the problem. The opportunity register maps the full scope. The 90-day plan is what you execute against on Monday morning.
If your audit produces a nice slide deck but no prioritized action plan with pound values attached, it wasn’t worth the time.
People are often surprised when I say this takes two to three weeks. They’ve been quoted twelve-week engagements by big consultancies. The difference is approach.
I don’t start with a blank slate. After doing this across dozens of e-commerce businesses, I know where the bodies are buried. The COGS analysis follows a pattern. The tech stack audit has a playbook. The fulfilment cost breakdown uses a model I’ve refined over years. I’m not building frameworks from scratch — I’m applying proven ones to your specific data.
The two-to-three week timeline assumes the business can provide the data reasonably quickly: P&L detail, vendor contracts, tech stack list, org chart, marketing spend by channel, fulfilment cost breakdowns. If that data is scattered across ten different systems and nobody knows where anything is, add a week for data gathering. But even then, a month is the upper bound for a thorough audit.
The speed matters because margin improvement is a compounding problem. Every week you delay is a week of continued overspend. A quick, focused audit that leads to immediate action will always beat a slow, exhaustive one that produces a thicker report.
Book a margin audit. We’ll work with your leadership team to map every cost between revenue and EBITDA, identify the 10–15 highest-impact improvements, and build a 90-day plan to capture them. Two to three weeks. Concrete numbers. No fluff.
We go into businesses and make them permanently more profitable. Every initiative is EBITDA-tracked.