Personalisation for Online Stores Without Being Creepy

Good ecommerce personalization means showing each shopper products, content and messages that are more relevant to them: recommendations based on what they bought, reminders for things they regularly need, local delivery information, or content in the categories they care about. Done well, it saves customers time and feels like good service. Done badly, it feels intrusive, makes odd assumptions or reveals more than people expect a store to know. The practical rule is to personalise using information customers knowingly share with you, in ways that clearly help them, with their consent where required, and to stop whenever a feature stops being helpful.

What personalisation can do for shoppers

The best personalisation works like a good shop assistant: it remembers what you bought, notices what you are looking for and suggests something useful, without making you feel watched. Large catalogues can overwhelm. Personalisation helps shoppers find what is relevant faster and reminds them of things they genuinely need. Useful examples include:

• Recently viewed products, so shoppers can return to items they were considering.
• Related and complementary products, such as cases for the phone they bought or refills for their device.
• Reorder reminders for consumables, timed to typical usage.
• Local information, such as delivery estimates and currency for their country.
• Category-aware content, such as guides related to what they usually buy.
• Size and preference memory, such as remembering their usual clothing size.

Notice what these have in common. Each uses information the shopper has clearly given you, and each makes the next step easier for them. These help shoppers directly and rarely feel intrusive, because the link between their behaviour and the suggestion is obvious.

Ecommerce personalization: where to start

Begin with the simple, high-value options before investing in complex systems. Many stores get most of the benefit from a few basic features done well, and only move to advanced tools once those are working and measured:

Each feature can be tested on its own, so you can see which ones help your customers before adding more. Many platforms and email tools include these features. Our guide to ecommerce email marketing explains how personalised flows work, and our guide to AI in ecommerce covers more advanced recommendation tools.

Avoiding the creepy line

Trust is fragile. One badly judged message can undo months of helpful service, and customers often share examples of creepy marketing with friends. Personalisation becomes uncomfortable when it seems to know too much, makes awkward assumptions or follows people around. Avoid:

• Revealing sensitive interests, such as health, pregnancy or financial difficulty, in emails or on shared screens.
• Referencing fleeting behaviour, such as "We noticed you looked at this for two seconds".
• Endless retargeting of a product someone already bought or clearly rejected.
• Different prices for different people in ways that seem unfair or hidden.
• Combining data from unexpected sources without telling customers.

A useful test: would the customer be surprised or uncomfortable if they knew how this suggestion was made? If yes, rethink it. Another test is to imagine the message being seen by someone else in the customer's household. If that could embarrass them, the message is too personal.

Data, consent and privacy

Privacy is not only a legal question; it is part of the customer experience. Personalisation uses personal data, so privacy rules apply. In many markets you need a lawful basis for processing, clear information about what you do, and consent for certain tracking and marketing. Automated profiling that significantly affects people has extra rules in some places. The UK regulator's guidance on automated decision-making and profiling explains the principles.

• Explain personalisation in your privacy policy in plain language.
• Respect cookie and marketing consent choices.
• Let customers manage preferences and opt out of personalised marketing.
• Collect only the data you need and keep it secure.
• Avoid using sensitive categories of data for marketing.

Being open helps. A short note such as "Recommended because you bought…" explains a suggestion and reassures customers that you are using their purchase history, not something hidden. Customers who understand why they see a recommendation are more likely to trust it.

Our guide to GDPR for online stores covers the European rules in more detail.

Personalising with first-party data

The most reliable data for personalisation is information customers give you directly: their purchases, account preferences, sizes, saved lists, location for delivery and the topics they choose in a preference centre. This data is accurate, clearly connected to your relationship with the customer and less affected by browser privacy changes than tracking across other websites.

First-party data also builds a closer relationship. Customers who tell you their preferences expect you to use them, and they notice when you do. Encourage customers to share preferences by making the benefit clear: "Tell us your size and we'll show items in stock for you" or "Choose topics and we'll only email what interests you".

Measuring whether it works

Personalisation can absorb a lot of time and budget, so it deserves the same scrutiny as any other investment. Personalisation is easy to over-credit. Products recommended to people who were going to buy them anyway look like successes. To measure honestly:

1. Use a control group that does not see the personalisation, and compare results.
2. Measure revenue per visitor and repeat purchase rate, not just clicks on recommendations.
3. Watch for side effects, such as lower basket value or more returns.
4. Review regularly, because what worked last season may not now.

Look at qualitative signals as well: customer comments, unsubscribe reasons and complaints that mention irrelevant or repetitive suggestions. They often reveal problems before the numbers do.

Personalisation along the customer journey

Home page

The home page is often the first place stores try personalisation, and the place where mistakes are most visible. Returning visitors can see recently viewed items, categories they browse most and new arrivals in those categories. First-time visitors should see a clear, general home page; personalising for someone you know nothing about usually means guessing.

Category and search pages

In large catalogues, sorting can gently favour products related to a shopper's past purchases or preferred sizes, while still showing the full range. Remembering a shopper's last filters, such as their clothing size, saves them repeating the same choices on every visit.

Product pages

"Frequently bought together" and "complete the set" suggestions help shoppers find accessories they would otherwise forget. Keep them relevant and few; a long row of loosely related products distracts from the main decision.

Cart and checkout

Small, relevant add-ons, such as batteries for a device or a cleaning kit for leather shoes, can be helpful in the cart, especially when they are inexpensive and clearly connected to what is already there. Keep checkout itself clean; this is not the moment for long lists of suggestions that distract from paying.

After purchase

Care tips for the product bought, accessories that fit it and timely replenishment reminders are among the most welcome forms of personalisation. Our guide to customer retention explains how these messages bring buyers back.

Personalisation for business and international customers

Trade buyers

For business customers, personalisation is often about efficiency rather than discovery. A purchasing manager wants to find the usual items quickly, not be inspired: their own price list, their regular products at the top of search results, saved order lists, quick reorder from history and the payment terms agreed with their company. These features save buyers time on every order. Our guide to B2B ecommerce portals covers them in detail.

International shoppers

Showing the right language, currency, delivery options and local payment methods is the most basic and valuable form of personalisation for customers in other countries, and it is often overlooked in favour of more fashionable recommendation features. Suggest these based on location, but always let shoppers change them, since travellers and gift buyers may want something different.

Getting your team and tools ready

Personalisation depends on good data and clear rules, more than on clever software.

• Clean product data: accurate categories and attributes make recommendations sensible.
• Connected systems: your store, email platform and customer data need to share purchases and preferences reliably.
• Clear rules: decide which categories are too sensitive for personalised marketing, and how long data is used.
• An owner: someone who reviews what customers see, checks results and switches off features that do not help customers or the business.

Start small, review real examples of what customers are shown across different accounts, and expand only what proves useful. Customers notice odd or repetitive suggestions quickly, so a regular look at actual recommendations is just as important as the numbers in any monthly report.

Frequently asked questions

Want personalisation that customers welcome? Talk to our team about recommendations, personalised emails and privacy-respecting data for your store.