Website personalization examples that sell, by visitor intent
4 min read · updated
The website personalization examples that move revenue respond to what a visitor is doing right now, not to who they are. Someone who read the returns policy before the cart needs return terms beside the buy button. Someone comparing three products needs the difference between them. Someone returning to the same product needs it waiting for them. Each example below names the signal, the change and the number to watch, and each should run against a control group so you know it helped.
Why organise personalization examples by intent?
Most lists of personalisation ideas are organised by technique: geolocation, weather, name in the headline. Those are easy to build and rarely move revenue, because they answer a question nobody asked. A visitor does not need to be told which city they are in.
What changes a decision is removing the thing that is stopping it. So the examples here start from what a visitor is visibly doing, which tells you what they probably need, and only then say what to change.
What can you show a visitor who hesitates at checkout?
Checkout hesitation is the highest value moment on most stores, because the person has already decided what they want. The usual causes are cost surprises, delivery doubt and payment worry.
- Signal: reached checkout, went back to the cart. Change: show the full delivered price and the delivery date beside the pay button. Watch: checkout completion.
- Signal: opened the shipping page before checkout. Change: a one line delivery promise with the pin code check inline. Watch: orders from visitors who saw it versus control.
- Signal: an Indian visitor on the payment step. Change: state clearly whether cash on delivery is available and whether it costs extra. Watch: payment step drop off.
What helps a visitor comparing products?
Someone who opens three products and returns to the first is not undecided about buying; they are undecided about which. A generic banner does nothing for them.
- Signal: viewed two or more products in the same category in one visit. Change: a short comparison block on the product page naming the one real difference. Watch: add to cart rate.
- Signal: back and forth between two variants. Change: show the review count and the most useful review for each. Watch: time to add to cart.
- Signal: opened the size guide or specification page. Change: put the fit or compatibility answer next to the size selector. Watch: returns as well as conversions.
How should a site treat a returning visitor?
A returning visitor who comes back to the same product two or three times across visits is one of the strongest buying signals there is. They should not have to find it again.
- Signal: same product viewed in two or more sessions. Change: a welcome back line on the home page linking straight to it. Watch: conversion of returning visitors.
- Signal: cart left more than a day ago. Change: show the cart contents and any stock warning on arrival, without a discount. Watch: recovered carts versus control.
- Signal: already bought in the last thirty days. Change: stop showing acquisition offers and show care instructions or the matching refill instead. Watch: repeat purchase rate and unsubscribes.
What works for visitors who are worried about price?
Reading the pricing page more than once, or for a long time, usually means the price is not the problem so much as what the price includes. A discount is the most expensive answer and often the wrong one.
- Signal: pricing page opened twice in one visit. Change: a short what is included line above the plans. Watch: plan selection rate.
- Signal: long read of the pricing table. Change: point the main button at a conversation with a person rather than a form. Watch: qualified conversations started.
- Signal: arrived from a campaign that promised a specific offer. Change: repeat that offer, word for word, at the top of the landing page. Watch: bounce rate on campaign traffic.
Which personalization examples backfire?
Three patterns lose more than they win. The first is showing a visitor you know too much: naming their company, city or browsing history back to them reads as surveillance. The second is a discount to everyone who hesitates, which teaches your best customers to hesitate. The third is personalising without a control group, because then a change that hurt conversion looks like a success for months.
A fourth, quieter one: personalising so much that nobody can say which version a customer saw when they complain. Keep a record of what each visitor was shown.
How do you measure a personalization example?
Pick one number per example before you start, usually the conversion step right after the change. Keep a random share of eligible visitors, 10 to 20 percent, on the unchanged page. Compare eligible visitors who saw the change with eligible visitors who did not, and read the difference with its uncertainty. If you have fewer than a few hundred visitors in each group, the result is preliminary however good it looks.
| Checkout reassurance | Checkout completion among visitors who reached checkout |
| Comparison block | Add to cart rate on compared products |
| Welcome back | Purchase rate of returning visitors |
| Pricing clarification | Plan selection or conversation started |
| Campaign message match | Bounce rate and conversion on that campaign's traffic |
Can these examples run without a developer?
With IntellQ, mostly yes. You mark the elements that may change once, with a data-intellq-slot attribute, and the agents choose between these kinds of changes per visit from the signals above, keeping a control group on each. Changes are limited to setting text, showing, hiding and same site links on marked elements, so a mistake cannot rewrite the rest of the page. Anything beyond that, such as a new comparison component, still needs building once.
Questions
What is the simplest website personalization to start with?
Repeating a campaign's promise on the page that campaign lands on. It needs only the campaign tag in the link, it is easy to measure on that traffic alone, and it fixes the most common reason paid visitors leave immediately.
Is showing a visitor's city personalisation?
Technically, but it rarely helps a decision unless delivery or price actually depends on location. Show the delivery date for their pin code instead of the city name.
Should personalisation include discounts?
Sparingly. A discount for every hesitant visitor trains people to hesitate. Try reassurance, clearer costs and delivery dates first, and test any discount against a control group that does not get it.
How many personalisation rules should a store run?
Fewer than you think. Five well measured changes on the highest intent pages beat forty unmeasured rules that nobody can explain. Retire any change that does not beat its control.