E-commerce Personalisation: What Works and What It Returns

Ecommerce personalisation means tailoring what each shopper sees (products, recommendations, offers, content and messages) to their behaviour and stated preferences. Done well, it lifts revenue and lowers acquisition costs; McKinsey puts typical revenue gains at 5% to 15%. The tactics that work for most retailers are not elaborate: relevant recommendations, useful reminders, consistent experiences across channels and respect for privacy.

Why personalisation matters to shoppers

Expectations are high. McKinsey's Next in Personalization 2021 report found 71% of consumers expect companies to deliver personalised interactions and 76% get frustrated when that does not happen. Faster-growing companies derived 40% more of their revenue from personalisation than slower-growing peers.

In Australia, those expectations now span channels. Salesforce's 2026 ANZ research found 88% of B2C respondents agree customers expect the same personalisation in store as online, and personalisation was among the top improvements reported by organisations using agentic AI.

AI assistants are raising the bar further. ChatGPT's shopping research asks follow-up questions about budget and priorities and, with memory on, factors in what it already knows about the user. Shoppers used to that will notice a store that treats every visitor the same.

What the returns look like

McKinsey's explainer on personalisation summarises its research: personalisation can reduce customer acquisition costs by as much as 50%, lift revenues by 5% to 15%, and increase marketing ROI by 10% to 30%.

Recommendation engines show up in peak trading data too. Salesforce reported that during Cyber Week 2025, AI and agents influenced 20% of global orders, driving US$67 billion in sales through personalised product recommendations and conversational customer service.

These are averages across large datasets. Your results depend on traffic volume, catalogue size and data quality.

Tactics that work for most retailers

1. Complementary recommendations, not repeats. McKinsey's diary study of shoppers found people value relevant recommendations they would not have thought of themselves. Showing the same item they already viewed can feel intrusive; suggesting what goes with it (the cover for the outdoor setting, the filter for the purifier) feels useful.

2. Behaviour-based email and SMS. Abandoned cart, browse abandonment, back-in-stock and replenishment messages use data you already hold and usually outperform generic campaigns.

3. Personalised on-site search and category sorting. Re-rank results using what the shopper has clicked, bought or filtered on. See our guide to site search.

4. Location-aware delivery and stock messages. Showing realistic delivery dates and nearby pickup options for the shopper's postcode is personalisation that directly affects conversion.

5. Know when to stop. McKinsey recommends tracking impressions and stopping ads to customers who have not responded. Frequency caps protect your brand.

6. Consistency across channels. A shopper who buys on a marketplace and later visits your store should not be offered an introductory discount on a product they already own. That requires order data from every channel in one place.

Measuring whether it works

Personalisation is easy to switch on and hard to prove. Before scaling any tactic:

• Hold out a control group. Keep 10% to 20% of visitors or subscribers on the non-personalised experience so you can compare conversion, order value and revenue per visitor.
• Measure margin, not just clicks. A recommendation widget that pushes heavily discounted items can lift conversion while lowering profit.
• Watch for fatigue. Rising unsubscribe and complaint rates are an early sign that messages feel intrusive rather than helpful.
• Check every channel. If marketplace orders are missing from your data, your segments will be wrong and your offers will misfire.

Personalisation and privacy

Personalisation depends on data, and Australians are wary of how it is used. The OAIC's 2023 community attitudes survey found 84% want more control over the collection and use of their information, and 89% consider online tracking, profiling and targeted advertising aimed at children unfair and unreasonable.

Practical rules: personalise from data customers knowingly gave you (purchases, preferences, sign-ups), explain why you are recommending something, and avoid sensitive inferences. Our guide to first-party data covers how to build that data properly.

Key figures

• 71% of consumers expect personalised interactions; 76% are frustrated when they do not get them (McKinsey, 2021).
• Personalisation can cut acquisition costs by up to 50% and lift revenue 5% to 15% (McKinsey, 2023).
• AI and agents influenced 20% of global Cyber Week 2025 orders, worth US$67 billion (Salesforce, 2025).
• 88% of B2C respondents in ANZ say customers expect the same personalisation in store as online (Salesforce, 2026).
• 84% of Australians want more control over their personal information (OAIC, 2023).

How eComChannels helps

Personalisation is only as good as the data behind it and the effort it takes to run. eComChannels reduces both problems.

• Order management gives you one view of orders from every channel, so you know what each customer has bought, whether on your store or a marketplace, and can avoid irrelevant offers.
• Branded storefronts are connected to the same live catalogue and stock as your marketplaces, so what you promote to each shopper on your own site is something you can actually ship.
• AI social media outreach drafts posts and content calendars tailored to each platform's audience, from Instagram and TikTok to Pinterest and RedNote, so tailored outreach does not mean more hours of manual work.

Relevance that lifts conversion, with lower operating overhead. Book a demo or start a 30-day free trial.

FAQ

Do I need a lot of traffic to personalise?

Rule-based personalisation (complementary products, replenishment emails, location-aware delivery dates) works at any size. Machine-learning recommendations improve with more data, so smaller stores should start with rules.

What should I personalise first?

Start where intent is highest: product page recommendations, abandoned cart messages and search results. They are simple to measure.

Can I personalise on marketplaces?

Not directly. Marketplaces run their own recommendation and search engines. What you control is the quality of your listings, which determines how well those engines match your products to shoppers, plus personalisation on your own store, email and social channels.