ANDWHATSMARTERDATAHABITSDOABOUTIT

Pricing is the fastest profit lever a webshop has, and the one most teams act on least. For a typical S&P 1500 company, a 1% price increase with volume held constant produces an average 8.7% rise in operating profit (McKinsey, The Power of Pricing). The same 1% applied to volume or to cost moves the needle far less. Yet Simon-Kucher's 2025 global survey of more than 2,200 leaders across 28 countries found sales volume is still the most-cited profit driver for the next two years, "proving pricing remains an underused profit lever."

The gap is not awareness. 85% of executives told Bain their organisation has significant room to improve pricing (Bain & Company). The gap is ownership. In most webshops, pricing sits between commercial, marketing, and finance, and belongs to none of them. The product manager sets a launch price. The category lead adjusts on promos. Finance reviews margin once a quarter. Nobody owns the day-to-day price.

Here is the uncomfortable part. Pricing initiatives, when actually run, are also the most reliable lever. Simon-Kucher's 2025 data puts pricing's average time to business impact at 7.8 months, against 11 to 17 months for other value-creation programmes, with only 4% failing to meet business case expectations. The risk-adjusted return is high. The activation rate is low.

In most webshops, the mistake shows up as three habits stacked on top of each other: cost-plus on the price tag, competitor matching at launch, and set-and-forget after that. About 30% of e-commerce companies use dynamic pricing today, according to a 2024 Gartner pricing analysis. The remaining 70% rely on one or more of those static methods. Cost-plus has nothing to do with what customers value. Competitor matching commoditises the catalogue against the cheapest player. Set-and-forget treats every customer, every basket, and every traffic source as if they were the same person on the same day.

What happens in practice: a webshop runs a paid acquisition campaign that brings in higher-intent buyers. The price does not move. The discount banner that worked in Q4 keeps running in Q2. The brand sells stock that should be re-priced for end-of-season at the same price as full launch. The shop logs all of this. Sessions, baskets, add-to-cart events, return rates, time-on-page, code-entry events. The data is sitting in the warehouse. The pricing system is not connected to it.

Set the catalogue. Forget the catalogue. Wonder why margin keeps drifting.

When we audit webshop pricing setups, the most common pattern is not a missing tool. It is a missing habit: nobody on the team reviews price-related data on a weekly cycle. The data exists. The cadence does not. Pricing without cadence is gut-feel with a spreadsheet.

Dynamic pricing alone does not solve the problem, and in some cases it makes it worse. A December 2024 Gartner consumer survey found that 68% of consumers feel taken advantage of when brands use dynamic pricing. Plugging a generic pricing engine into a webshop, configuring it on competitor scrape data and a margin floor, and letting it run is the express route to that 68%. The shop trades short-term margin for medium-term trust.

The deeper issue is that most off-the-shelf dynamic pricing tools optimise for one variable: the lowest defensible price against a competitor set. That is not pricing strategy. That is automated competitor matching with a margin guardrail. Generic tools commoditise. The 20% of your business that is genuinely unique, the part where you actually have pricing power, gets ignored.

Connectivity, not intelligence, is what most dynamic pricing tools quietly lack. The model is fine. The connection to your customer data, your basket data, your return data, your loyalty data, is not. Smart pricing, completely disconnected from where the value is actually created.

That is the same pattern we have written about elsewhere on the .digital radar (the MCP explainer makes the case in detail for AI tools that cannot reach the systems where work happens). It applies to pricing tools too. Intelligence with no connection is decoration.

Smarter data habits are about closing the loop between the data the webshop already collects and the price the customer sees. Done well, this is one of the highest-return data investments a mid-market operator can make. Brands that activated first-party data across all four core marketing functions saw a 2.9x revenue uplift versus peers, according to a joint BCG x Google study. Personalisation built on the same data lifts revenue by 10 to 15% on average, and by up to 25% in the top quartile (McKinsey).

Three habits do most of the work.

First, instrument margin, not just revenue. Most webshop dashboards report gross merchandise value, conversion rate, and average order value. Few report contribution margin per SKU, per basket, per channel. Without that, every pricing decision is partial. You can grow revenue and shrink profit at the same time, and most teams find out a quarter too late.

Second, segment by behaviour, not just by persona. The same customer behaves differently at 09:00 on a Tuesday and at 23:00 on a Saturday. They behave differently when they arrive via paid search and via an email re-engagement flow. Pricing built on intent signals (browse-to-basket time, code-entry frequency, return rate by segment) is a different category of decision from pricing built on demographic personas.

Third, pilot on one product family before you procure a platform. The most common procurement mistake we see in webshop pricing is buying a system before the habit exists. A six-week pilot on one category, with one analyst, on existing data, will tell the team more about its pricing capability than a six-month platform rollout. If the habit lands, the platform is easy. If it does not, no platform fixes that.

These habits do not require new tools. They require connecting the tools the webshop already runs.

The leaders are not subtle about how committed they are. More than a decade ago, Amazon was already making over 2.5 million price changes per day, according to a Profitero analysis reported by Quartz in 2013. For comparison, the same study put Best Buy and Walmart combined at about 50,000 changes per month. Amazon's number has only grown since. The pattern is not "we change prices when something changes." The pattern is "the price is the output of a model that ran a minute ago."

Closer to the European mid-market, Zalando publishes engineering posts about its pricing infrastructure. The team runs tree-based and Transformer-based models on Amazon SageMaker against billions of weekly records, across multiple markets, according to a 2024 AWS Machine Learning case study. Industry analysis at retail-data firms like Intelligence Node has noted similar dynamic, behaviour-linked approaches at ASOS, though the brand does not publicly disclose algorithm details.

Most mid-market webshops will not run Transformer models on billions of records, and do not need to. The borrowable principle is not the scale. It is the architecture: the price is a model output, the model runs on the shop's own first-party data, the loop is closed.

A retail-specific BCG analysis published in 2024 found that retailers implementing AI-driven dynamic pricing have increased gross profit by 5 to 10% while sustainably growing revenue. McKinsey's 2024 LLM-to-ROI study on generative AI in retail put cost reduction at around 15% and revenue growth at around 10% for retailers running AI at scale. Two independent consultancy benchmarks land on the same range.

The window is open. McKinsey's State of AI 2025 reports that 88% of organisations now use AI in at least one function, but no more than 10% in any given function report having scaled AI agents. Most pricing AI is still pilot-stage. The mid-market companies that move first, on the back of their own data, will compound the advantage before the market catches up.

Start with the habit, not the platform. A pricing platform without a pricing habit is shelfware. A pricing habit without a platform still moves margin. Four steps, in order:

1. Audit the pricing data you already have. Web analytics, transaction data, basket data, return data, loyalty data, customer service interactions, and (often forgotten) discount-code redemption data. The audit usually finds that 70% of what is needed is already in the warehouse, just not joined up. This is the work people skip because it feels unsexy. It is also the work that determines whether anything that comes after it will function.

2. Instrument margin at SKU and basket level. Move beyond revenue and conversion rate. If contribution margin per SKU, per channel, per customer segment is not a number the team can pull in 30 seconds, no pricing decision after this is grounded.

3. Pick one product family for a behavioural pricing pilot. Six weeks. Existing tools. One analyst. Test elasticity, willingness-to-pay, and price-response by segment on a controlled subset of the catalogue. Measure the change in contribution margin, not headline revenue.

4. Build the cadence before you build the system. A weekly pricing review meeting, with the data joined up, is a stronger predictor of pricing maturity than any tool. Once the habit exists, automation is a question of capacity, not capability.

This is not anti-platform. We are not anti-SaaS. For a lot of webshops, an off-the-shelf dynamic pricing engine, configured against the right data, is the right answer. The point is the sequencing. The habit, the data plumbing, and the pilot come before the procurement. Anyone who has watched a six-figure pricing system sit unused for eighteen months has lived this lesson once.

The trade-off is honest: building the habit takes engineering time the commercial team does not always want to spend. The first quarter looks like analytics work, not revenue work. By the second quarter, if the loop is closing, the pricing decisions made on the back of it are the ones that actually compound.

It changes the urgency, not the destination. In April 2025, Google formally abandoned its plan to phase out third-party cookies in Chrome, leaving the existing "user choice" model in place. That sounds like a reprieve. It is not, in any meaningful sense. Roughly half the open web (Safari and Firefox) is already cookieless by default. Privacy regulation in the EU continues to tighten. Apple, Meta, and Google are all moving infrastructure toward signals that work without third-party identifiers.

For webshop pricing, the implication is concrete: external behavioural data is getting more expensive, less reliable, and more contested. The data that does compound is the data you collect on your own surfaces. First-party data is the only base that survives the next five years of platform and regulatory change. Building pricing decisions on top of it is not a privacy nicety. It is the only model that scales.

What is the difference between rule-based dynamic pricing and AI-driven pricing for webshops?

Rule-based dynamic pricing changes prices based on a defined rule set, typically competitor scrape data, time of day, or stock levels. AI-driven pricing learns from first-party behavioural data (basket composition, browse patterns, return rates, willingness-to-pay signals) and adapts the model itself. According to BCG (2024), retailers running AI-driven pricing have increased gross profit by 5 to 10%, materially above what rule-based dynamic pricing typically delivers.

How much pricing improvement can a mid-market webshop realistically expect?

McKinsey's classic pricing benchmark puts the average operating-profit impact of a 1% realised-price improvement at 8.7% for S&P 1500 companies. For mid-market e-commerce operators, BCG's 2024 retail pricing study shows gross profit lifts of 5 to 10% are typical at scale. The realistic target for a first-year programme is in the lower half of that range, with the habit and the data plumbing landed in months 1 to 3.

Will customers punish a webshop for using dynamic pricing?

They will if the pricing feels arbitrary. Gartner's December 2024 consumer survey found 68% of consumers feel taken advantage of by dynamic pricing. The difference between the brands that get punished and the ones that do not is whether the pricing is anchored in value signals the customer recognises (loyalty, basket size, returning-customer context) rather than in opaque competitor matching. Transparent, value-aligned pricing is the version that builds trust.

How long before a behavioural pricing pilot shows results?

Simon-Kucher's 2025 Global Pricing Study puts the average time to business impact for a pricing initiative at 7.8 months, against 11 to 17 months for other value-creation programmes. A focused single-product-family pilot can show directional results in six weeks. Full programme ROI, with the cadence and the data plumbing landed, typically lands within two quarters.

What is the first investment a webshop should make in pricing?

Not a platform. A weekly pricing review meeting, with contribution margin reported at SKU and basket level, joined to behavioural data the webshop already collects. The habit before the system. Most of the value sits in connecting tools the team already runs.