Industry

Digital, software and AI for e-commerce and retail

The third-largest sector for digital transformation spending, and the one where a business can grow revenue while losing money on every order.

Discuss your situation All services

Retail and e-commerce sit behind only manufacturing and professional services in digital transformation spending, and the money mostly goes into acquisition. That is the wrong end. Most struggling online retailers do not have a traffic problem; they have a contribution margin problem that traffic makes worse.

The work that matters is the arithmetic first — what an order actually earns after shipping, returns, payment fees and acquisition — and then the store, the channels and the operations built around what that number will support.

Why this sector is moving now

Acquisition costs have risen across every paid channel while conversion has not, which compresses the margin available to fund it. Retailers who grew when traffic was cheap are discovering that the model was subsidised by the auction rather than by the product.

Local payment expectations are the most consistently underestimated factor for anyone selling across borders. iDEAL in the Netherlands, Bancontact in Belgium, BLIK in Poland, MobilePay in Denmark, Multibanco in Portugal — a card-only checkout underperforms in each, and the cause is invisible from a funnel report.

Returns are the third pressure and the one most often modelled optimistically. A return rate estimated rather than measured invalidates the contribution calculation entirely, and in apparel it is frequently the single largest variable cost after the product itself.

The pressures behind it

Acquisition cost inflation: Paid channels costing more while conversion stays flat, compressing the margin that funds them.
Local payment methods: A card-only checkout underperforming in most European markets for reasons a funnel report cannot show.
Return rates: Frequently the largest variable cost after product, and frequently estimated rather than measured.
Feed quality: Missing attributes and inconsistent variants capping shopping campaign performance more than bidding does.
Operational cost per order: Manual routing and reconciliation that scales linearly, so growth erodes margin.
Marketplace dependency: Volume from platforms that own the customer relationship and take the margin.

Where the work usually starts

The systems and channels a retailers typically needs joined up. Most engagements begin at one spoke and widen only if it earns it.

The unit economics, always, and before anything is built. Contribution margin per order after every cost, and the acquisition cost the product can sustain. Occasionally this produces the conclusion that the business as configured cannot work, which is worth knowing in week one rather than after a quarter of spend.

Then checkout and payment methods, because that is usually the largest recoverable conversion gain. Acquisition comes after the destination is fixed, not before.

Marketing and brand for retailers

Brand Strategy & Development: Direct-to-consumer positioning collapses into the same few claims — sustainable, crafted, honest pricing — which is why so many brands are visually interchangeable. The differentiator that holds is usually a specific customer situation rather than a product attribute, and it changes what the whole catalogue is for.
Brand Management: Retail brand consistency is mostly a product data and photography problem at scale. A catalogue of two thousand SKUs shot by three different people over five years looks like three different brands, and fixing it is a systems exercise rather than a creative one.
Social Media Strategy: Channel choice should follow where the product is actually discovered, which for most consumer categories now means short-form video rather than static feed. The honest recommendation is frequently to run one channel properly instead of four sporadically.
Social Media Management: Volume is the constraint. Paid social optimises by testing variants, which makes creative production a performance input rather than a brand expense, and most in-house teams cannot sustain the cadence the platforms reward.
Content Creation & Creative Production: Product photography is the product online, and the gap between competent and good is directly measurable in conversion. Shooting for every crop and placement in one session is what makes the volume affordable, and it is the planning most retailers skip.
Digital Marketing: The discipline that matters is reallocation on contribution rather than on revenue. A channel producing revenue at negative contribution is worse than one producing nothing, and revenue-based reporting actively conceals that.
Paid Advertising: Shopping campaign performance is driven by feed quality more than by bidding, and feeds are almost always worse than clients expect. Missing attributes, inconsistent variants and poor titles cap performance in ways no amount of bid management recovers.
Search Engine Optimisation: Category and buying-guide pages are the durable asset; product pages rarely rank on their own. The structural work — faceted navigation that does not generate infinite indexable URLs, canonical handling across variants — matters more here than in any other sector.
Email, SMS & WhatsApp Marketing: The highest return per hour in e-commerce and the most under-invested. Abandoned cart, post-purchase, replenishment and win-back sequences produce revenue from customers already paid for, which is precisely why they beat acquisition on margin.
Lead Generation & Prospecting: Rarely relevant for consumer retail and genuinely useful for wholesale and trade accounts, where a retailer sells business to business alongside direct. Those two motions need separating in the CRM or neither is measurable.

IT, software and AI for retailers

Website Design & Development: Speed is a conversion factor and a ranking factor simultaneously, and most storefronts carry weight from apps nobody audits. Checkout design, payment method coverage and correct VAT handling do more than any redesign of the homepage.
CRM & Sales Systems: For consumer retail this is the customer data platform question: one record across storefront, support and marketing. Where a retailer also sells wholesale, that pipeline is a genuine sales process and needs modelling as one.
Business Process Automation: Order routing, stock reconciliation, supplier notification, returns processing and fulfilment updates. Returns automation in particular is under-built because it happens at lower volume, and it is where per-order cost quietly concentrates.
AI Automation Systems: Product data enrichment is the strongest fit: generating and normalising attributes, categorising SKUs, drafting descriptions for review. It is high-volume, bounded, and directly improves feed quality, which is the thing capping paid performance.
AI Knowledge Bases & RAG: Customer service assistants grounded in real product data, sizing guides and policy documents, with citations. The failure mode to design against is confident invention about stock, delivery dates or specifications, all of which have commercial consequences.
AI Voice & Customer Communication: Order status is the obvious bounded use, and it displaces a genuinely high volume of low-value contacts. Anything touching a refund decision should reach a person, because the cost of getting it wrong exceeds the cost of the call.
Custom Software & Platforms: Justified where the product or pricing model does not fit a platform — configurable products, complex bundling, B2B pricing tiers alongside consumer sales. Not justified as a general storefront replacement, and we will argue that case before quoting.
Data Engineering & BI: Contribution margin by product and channel is the report most retailers cannot produce and most need. It requires reconciling storefront, payment, shipping and returns data that currently sit apart, and that reconciliation is the work.
Cloud, DevOps & Infrastructure: Traffic is spiky in a way most sectors are not, and a campaign or a season can multiply load in an hour. Capacity and caching decisions should be made against peak rather than average, and tested before the peak rather than during it.
Systems Integration: Storefront to stock, orders to fulfilment, payments to accounting, marketplace listings to inventory. Idempotency is critical: a duplicated order or a double-decremented stock level in retail is customer-visible immediately.
Digital Transformation Consulting: The audit routinely finds that the acquisition problem is a margin problem and the margin problem is an operations problem. Sequencing runs economics, checkout, retention, operations — with acquisition last rather than first.
Maintenance & Ongoing Support: Storefront platforms update, payment providers change APIs, and feeds break silently. A broken feed costs revenue every hour it is down and produces no error anyone sees, which is why monitoring matters more than availability here.

What is specific to this sector

Cross-border VAT in the European Union runs through the One Stop Shop regime, with distance selling thresholds and invoicing obligations that vary by member state. A checkout that charges the wrong rate creates a liability rather than a rounding error, and several markets — Italy, Romania, Hungary, Portugal — additionally require structured electronic invoicing or transaction reporting that a PDF does not satisfy.

Consumer rights in the EU include a fourteen-day withdrawal period for distance sales, with prescribed information that must be provided before purchase and specific rules on who bears return shipping. Those obligations shape the returns process and the product page, not just the terms document.

The European Accessibility Act reaches consumer-facing e-commerce from June 2025, which for most online retailers makes WCAG conformance a legal question rather than a preference. In the United States, ADA-related litigation over inaccessible storefronts is a live and continuing commercial risk.

Not legal or regulatory advice. Sector rules described here are scoping context, current to our latest review. Confirm what applies to your business with a qualified adviser.

Questions

Why start with the economics rather than the store?

Because the numbers determine what the store, the shipping policy and the acquisition strategy can be. Building first and modelling later means discovering in month three that no realistic acquisition cost makes the configuration profitable.

What is the most common hidden problem?

A card-only checkout in markets that pay another way, and an understated return rate. The first suppresses conversion invisibly; the second invalidates the margin model entirely. Both are cheap to check and rarely checked.

Where does the best return come from?

Usually post-purchase and lifecycle communication, because it produces revenue from customers already acquired. It is less visible than advertising, which is largely why it is under-invested in.

How much does the product feed matter?

More than bidding does for shopping campaigns. It is a boring problem with large commercial consequences, and feed quality is almost always worse than the retailer believes.

Should we sell on marketplaces?

Frequently yes for volume and no for margin, and the honest framing is that you are renting the customer relationship. It works as a channel and fails as a strategy, and the direct channel needs building alongside it.

Can you handle multi-country VAT?

Yes, and it needs designing in rather than added later. Cross-border rates, One Stop Shop reporting and country-specific invoicing rules affect the checkout and the accounting integration simultaneously.

What does it cost?

Quoted per phase after a discovery call, with acquisition work as a separate monthly arrangement from media spend. We do not take a percentage of ad spend, because it rewards recommending more of it.