Five situations.
Five individual perspectives at the table.

In peak week your strategy has to hit the road: assortment, prices, logistics and systems at once. Retailers who deliver reliably win customers the competition is disappointing right then. Peak readiness is a market share goal, not only an IT goal.

Between a calm November and a lost one stands architecture. Systems have to scale automatically, and load tests belong before the season. Anyone who talks about capacity in October can only oversize.

Peaks produce the densest data of the year. What is searched, abandoned and bought shows up nowhere more clearly. Prepare the analysis and you learn more about your customers in two weeks than in two quarters.

Fixed capacity for the peak day is paid for on 365 days. Scaling infrastructure is paid for when it carries revenue. Then there is the risk. A few minutes of downtime in peak week cost more than the annual budget for load tests.

Every peak-week campaign makes a promise that shop and logistics have to keep. The most expensive mistake is paid traffic on a page that cannot cope. Campaign planning and capacity planning belong at one table.

Seven out of ten baskets die on the last few metres. Hardly anywhere in the company does so much revenue sit so close. Checkout excellence is unspectacular and beats almost every growth initiative on value for money.

The most common reasons for abandonment are system decisions: load time, payment methods, forced accounts, form logic. A modern checkout is therefore not a redesign. It is a rebuild of the journey behind it, including payment connections and guest processes.

Abandonment data is an early warning system. Where exactly do customers drop out, on which device, with which payment method? Measure cleanly and you replace opinions about the checkout with facts.

Run one simple calculation: current revenue divided by the conversion rate times one percentage point of improvement. That is the annual potential of the journey. Against that number a checkout project is almost always small.

The checkout is the moment of truth for your brand, everything before it was a promise. Surprise costs at this point damage more than the single order. Transparency converts and stays in memory.

Personalisation is no longer an experiment in retail, it is the standard among the large players. The strategic question is how fast you catch up and where you start. Beginning with recommendations and search usually pays for itself.

The good news: personalisation does not replace your stack. It sits as a layer over shop, app and email and is connected through interfaces. The architecture question is the data feed, not the frontend.

No relevance without data. Behaviour, segments, purchase history. Personalisation is where data quality translates directly into revenue, and the best occasion to finally bring the customer profile together.

The McKinsey figure of 40 per cent more revenue describes the average of those who master it. Calculated conservatively, a fraction of that is enough to carry a testing programme. Demand measurable experiments instead of platform promises.

Relevance is the only form of advertising customers are grateful for. The regular sees new arrivals from their brand, the occasional buyer finds the right entry point. That is brand management inside the product.

Discounts buy revenue, loyalty builds company value. The gap is measurable: three quarters of executives call loyalty strategic, fewer than half consider their own programmes effective. Close that gap and you grow more profitably than the market.

Loyalty usually fails technically on the connection. The programme knows the online purchase but not the one in the store, the app not the newsletter. An engine with clean interfaces to till, shop and CRM is half the job.

Every incentive produces data about what really moves customers. Segmented programmes are a learning system. They show which customers respond to which incentives and make the next campaign measurably better.

The blanket approach is convenient and expensive. German retail loses 300 million euro a year to wrong discount strategies around Black Friday alone. Segmented incentives cost less and last longer. That is margin protection.

A good programme rewards behaviour you want, not purchases that would have happened anyway. Points are interchangeable, fitting benefits are not. That decides whether your programme produces cost or relationship.

The most valuable customers buy across several channels, and a fragmented system does not recognise exactly those customers. One customer view across all channels carries the other four initiatives. That is why it belongs at the top of the agenda.

The goal is not one giant system but a data foundation with clear ownership. Which system leads on customer, product, transaction? After that, integrations are defined routes instead of individual cases.

Start with a question nobody can answer today. For example: what is a store customer worth online? For that question the two or three necessary sources get connected. The foundation grows along real questions.

Fragmented data costs in three places: duplicate maintenance, wrong decisions, unused potential among the best customers. Investing in a foundation pays back through every initiative that becomes cheaper afterwards.

73 per cent of customers use several channels. They spend more when treated as one person instead of three records. Recognition is the simplest form of appreciation, and the systems decide whether it works.