Personalisation can influence almost every stage of an eCommerce journey. Retailers can change product recommendations, search results, offers, website content, emails and post-purchase communications according to what they know about a customer. But once personalisation becomes widespread, a deceptively difficult question emerges: Is it actually making customers more likely to buy?
Our first article, eCommerce Personalisation Software: What Retailers Should Compare, explored the technology required to deliver relevant experiences.
The second, eCommerce Personalisation Strategy: Turning Customer Data into Relevant Experiences, looked at where and how those capabilities should be applied.
The next challenge is measurement. Because a customer interacted with personalised content before purchasing does not necessarily mean the personalisation caused the purchase.
Separate Interaction from Impact
Imagine a customer regularly buys running shoes from a retailer.
They return to the website, see a personalised recommendation for running shoes and make another purchase.
The personalisation platform may attribute revenue to that recommendation.
But would the customer have purchased anyway?
That is the question measurement needs to answer.
Personalisation Principle
Don’t only ask how much revenue touched personalisation. Ask how much additional value the personalisation created.
This distinction between attributed and incremental performance is crucial.
Start with a Clear Hypothesis
Every meaningful personalisation test should begin with a reason for making the change.
For example:
If we recommend complementary products on the basket page, average order value should increase.
Or:
If returning customers see products from categories they previously browsed, conversion should improve.
That gives the retailer something specific to test.
Without a hypothesis, teams can end up deploying large numbers of personalised experiences and reporting whichever engagement metrics move afterwards.
Use Control Groups
One of the most useful ways to understand incremental impact is to compare customers receiving personalisation with an appropriate control group that does not.
For example:
Group A: personalised product recommendations.
Group B: standard recommendations.
Performance can then be compared across metrics such as:
- Conversion
- Revenue per visitor
- Average order value
- Click-through
- Add-to-basket rate
This provides stronger evidence than simply observing that customers clicked personalised content.
Measure According to the Objective
Not every personalisation should be judged by the same metric.
A homepage recommendation might aim to improve product discovery.
A basket intervention might aim to increase order value.
A post-purchase message might encourage repeat purchase.
Useful measures can therefore include:
- Conversion rate
- Revenue per visitor
- Average order value
- Add-to-basket rate
- Repeat purchase
- Retention
- Margin
- Customer lifetime value
Measurement Insight
The right metric is determined by the customer behaviour you were trying to change.
Don’t Optimise Clicks at the Expense of Purchases
Personalised recommendations can attract clicks without necessarily creating meaningful commercial value.
An eye-catching recommendation may increase engagement but distract customers from completing the purchase they originally intended to make.
Retailers should therefore follow the journey beyond the initial interaction.
Impression → click → basket → checkout → purchase → repeat purchase
This makes downstream outcomes more important than engagement in isolation.
Measure Revenue and Margin
Revenue growth does not always equal profit growth.
A personalisation strategy driven heavily by discounts may increase conversion while reducing margin.
Retailers should consider whether personalisation is:
- Increasing full-price purchases
- Encouraging unnecessary discounting
- Improving average order value
- Moving unwanted stock
- Increasing margin
- Changing returns
This is particularly important where algorithms are optimised around conversion alone.
Commercial Principle
The easiest customer to convert with personalisation may not be the most profitable customer to convert.
Test Product Recommendations Properly
Recommendation engines are one of the most visible applications of personalisation.
Providers such as Alchemy Network, Magebit, On Tap OÜ, SOZO Design and VINSINFO operate across areas of eCommerce technology, development and optimisation where recommendation and customer-experience strategies can form part of the wider digital proposition.
Retailers can test different recommendation strategies such as:
- Frequently bought together
- Similar products
- Recently viewed
- Category affinity
- Bestseller recommendations
- Complementary products
The important point is not to assume that a more sophisticated algorithm automatically delivers better commercial performance.
Test it.
Evaluate Personalised Messaging
Personalisation extends beyond the website.
Dotdigital and Brevo, for example, provide marketing and customer-engagement platforms supporting segmentation, automation and personalised communications.
Retailers can evaluate:
- Personalised subject lines
- Product recommendations
- Browse abandonment
- Basket abandonment
- Replenishment reminders
- Post-purchase journeys
Again, the most useful question is not simply whether personalised emails receive more clicks.
It is whether they change meaningful customer behaviour.
Measure Offers Carefully
RevLifter operates in the area of intelligent incentives and personalised offers, highlighting another measurement challenge.
If a customer was already prepared to buy at full price, giving them a discount may increase attributed conversion while reducing the value of the transaction.
Retailers should therefore consider:
Did the incentive cause the purchase, or merely subsidise it?
Testing different customer groups can help identify where incentives genuinely influence behaviour.
Segment Results
An overall average can hide important differences.
Personalisation may perform differently for:
- New customers
- Existing customers
- High-value customers
- Mobile users
- Particular product categories
- Different acquisition channels
A strategy producing a small average improvement might deliver substantial gains for one segment and no benefit for another.
Providers such as Culina, CTI Digital, ICS-digital and Merx work across different elements of digital commerce, customer experience and technology where understanding these customer and channel differences can support optimisation.
Watch for Personalisation Fatigue
More personalisation is not necessarily better.
Customers may become less responsive if every page, message and offer attempts to adapt itself around them.
Over-personalisation can also become repetitive or intrusive.
Testing should therefore ask not only:
Which personalised experience performs best?
but occasionally:
Does personalisation outperform doing nothing here?
That control is easy to overlook.
Account for Privacy
Personalisation often depends on analysing customer behaviour and preferences.
The ICO defines profiling as automated processing of personal data used to evaluate or predict aspects such as preferences, interests and behaviour, and UK data-protection requirements remain relevant when organisations use profiling for marketing and personalisation.
Retailers should understand:
- What data is being used
- The lawful basis for processing
- What customers have been told
- How long profiles are retained
- How customers can exercise relevant rights
The measurement strategy should not encourage collection of customer information simply because it might eventually prove useful.
Avoid Testing Everything at Once
If homepage content, recommendations, offers and email journeys all change simultaneously, identifying what caused the result becomes difficult.
Retailers should prioritise the areas where personalisation is most likely to solve a meaningful customer or commercial problem.
A useful sequence is:
hypothesis → control → test → measure → learn → scale
Successful experiences can then be expanded.
Unsuccessful ones can be changed or removed.
Use Personalisation Platforms as Learning Systems
The objective should not be to reach a point where personalisation is “finished”.
Customer behaviour changes.
Product ranges change.
Traffic sources change.
Economic conditions change.
What worked six months ago may not remain the best approach.
Arsenalia and other digital transformation specialists within the supplier group illustrate the wider role of technology, data and continuous optimisation in commerce environments.
The strongest personalisation programmes therefore develop a culture of ongoing experimentation rather than deploying rules once and leaving them untouched.
A Practical Personalisation Measurement Checklist
eCommerce teams should ask:
- What customer behaviour are we trying to change?
- What is our hypothesis?
- Do we have an appropriate control?
- Which metric represents success?
- Are we measuring beyond clicks?
- Is revenue incremental or merely attributed?
- What happens to margin?
- Do different customer segments respond differently?
- Are discounts genuinely influencing behaviour?
- Does personalisation outperform the non-personalised experience?
- Are customer data and profiling appropriately governed?
- Are successful tests being scaled and poor performers removed?
Frequently Asked Questions
How should eCommerce personalisation ROI be measured?
Measurement should reflect the objective of the personalisation and can include incremental conversion, revenue per visitor, average order value, margin, repeat purchase and retention.
What is incremental revenue from personalisation?
Incremental revenue is additional revenue that would not otherwise have occurred without the personalised experience. Control groups can help retailers estimate this more reliably than attribution alone.
Should every customer receive personalisation?
Not necessarily. Different customers may respond differently, and retaining non-personalised control groups can help establish whether personalisation is creating value.
Can personalisation hurt conversion?
Yes. Irrelevant recommendations, excessive offers or intrusive experiences can create distraction or friction, which is why testing is important.
Product Guide
Alchemy Network Ltd
Supports businesses with eCommerce and digital technology solutions focused on online customer experiences and growth.
Website: https://www.alchemynetwork.co.uk/
Arsenalia
Digital consultancy and technology group supporting organisations with digital transformation, commerce and customer-experience initiatives.
Website: https://www.arsenalia.com/
Brevo
Customer relationship and marketing platform providing email, automation, segmentation and other digital engagement capabilities.
Website: https://www.brevo.com/
CTI Digital
Digital agency delivering commerce, experience and technology services for organisations across multiple sectors.
Website: https://www.ctidigital.com/
Culina
Supports brands and retailers across areas of digital commerce and customer experience.
Dotdigital
Customer experience and data platform supporting marketing automation, segmentation and personalised digital communications.
Website: https://dotdigital.com/
ICS-digital
International digital agency providing digital marketing and online growth services across global markets.
Website: https://www.ics-digital.com/
Magebit
eCommerce agency specialising in the design, development and optimisation of online retail platforms.
Website: https://magebit.com/
Merx
Digital commerce specialist supporting businesses with eCommerce technology and online customer experiences.
On Tap OÜ
eCommerce technology specialist supporting retailers with development, integrations and optimisation.
RevLifter
Technology provider focused on intelligent incentives and personalised offers designed to influence customer purchasing behaviour.
Website: https://revlifter.com/
SOZO Design
Digital agency providing web, eCommerce and digital marketing services.
Website: https://www.sozodesign.co.uk/
VINSINFO
Technology and digital solutions provider supporting organisations with eCommerce and wider software requirements.
Website: https://www.vinsinfo.com/
From Personalisation Capability to Personalisation Proof
Across this three-part series, the journey becomes:
choose → apply → prove
The first challenge is selecting technology capable of delivering relevant experiences.
The second is deciding where personalisation genuinely improves the customer journey.
The third is proving that those experiences create value.
That final step matters because personalisation can become self-justifying: the platform reports that personalised experiences generated revenue, so the programme appears successful.
The stronger question is whether customers behaved differently because of the personalisation.
The eCommerce Forum connects senior eCommerce and payments professionals with carefully selected suppliers through pre-arranged one-to-one meetings, providing an opportunity to explore personalisation, customer experience, digital commerce and wider retail technologies.
Related Reading
This article follows eCommerce Personalisation Software: What Retailers Should Compare and eCommerce Personalisation Strategy: Turning Customer Data into Relevant Experiences.
Together, the three articles cover technology selection, practical application and performance measurement, providing a complete framework for developing an evidence-led eCommerce personalisation programme.
Sources
- Information Commissioner’s Office – Automated Decision-Making and Profiling – https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/individual-rights/automated-decision-making-and-profiling/
- Information Commissioner’s Office – Direct Marketing Guidance: Collect Information and Generate Leads – https://ico.org.uk/for-organisations/direct-marketing-and-privacy-and-electronic-communications/direct-marketing-guidance/collect-information-and-generate-leads/
Image credit: https://unsplash.com/photos/illustration-of-smartphone-application-screenshots-weRQAu9TA-A
