When you log into Amazon, the homepage tends to feature recommended products. And no two homepages are exactly the same. The recommended products vary from person to person.
First, there is the “Pick up where you left off” section. Here, Amazon shows you a selection of products that you looked at but didn’t buy. Then Amazon has featured categories. These are also inspired by your shopping and browsing habits.
Beneath the featured categories, there is the “Inspired by your shopping trends” section. Then beneath that is the “Related to items you’ve viewed” section— which as the name suggests, contains a selection of products similar to ones you have expressed interest in before.
As the largest e-commerce site in the world and one obsessed with pocketing every last dollar, Amazon clearly knows what they’re doing. They have mountains of data backing up their homepage design choices.
Why personalization works
A survey by VB insights found that 87% of people who personalized their websites saw at least a 5% rise in their key metrics (conversion rates, browsing time, newsletter sign-ups, etc.)
But the question is, why is personalization so effective? The answer is simple. People like to feel special.
A fairly generic website will appeal to a wide audience but the inherent risk of trying to appeal to everyone is that you will end up not appealing to anyone at all. Bars don’t stock milk and cookies because if they become hangouts for milk drinkers, the whiskey drinkers won’t show up.
But with personalization, a bar can create separate sections (or more preferably, separate locations) for milk drinkers and whiskey drinkers. This might be tough for a physical store to do but a lot easier for an e-commerce store.
Imagine how much worse the likes of Facebook, YouTube, or Tiktok would do if they showed the same content to everyone. If they tried that, no one would use those platforms anymore. But because they personalize every feed, they end up with engaged users who browse the sites for hours every day.
Personalization works the same for e-commerce. If I go to a site and have to browse through t-shirts, hoodies, and socks, I may not stay that long. They are broad-spectrum products that could appeal to anyone but what if I have enough socks? What if I don’t wear hoodies? What if I don’t want a t-shirt?
Sometimes, by being as wide with your targeting as possible, you end up missing the mark entirely. Now, obviously, you’ll hit your mark sometimes. Targeting a wide selection of customers isn’t a bad thing in itself. If you have a thousand visitors an hour, maybe 300 would be interested in the stuff you’re recommending. That’s not bad. But if you get granular with your recommendations and target each of the 1,000 visitors individually, you can do even better and end up engaging 80% of your visitors instead of just 10%.
The biggest problem with personalization is obviously where to start. Most of your traffic is probably anonymous. For you to personalize content to a user, you have to know them on some level. This is easy for signed-in customers since you have details from their previous visits and purchases to inform your personalization efforts.
But you can still personalize your site for first-time visitors to an extent. You can personalize according to:
- Geographical location – You can easily tell this from the IP address. For example, a user in southern Florida will probably not be interested in heated jackets even if it’s the middle of winter because it rarely freezes over there. On the other hand, a user from Minnesota will very much be interested in that heated jacket because a Florida winter is what Minnesotans call summer.
- Device – This needs no explanation.
- Referral source – If you’re running ads or an influencer campaign, then you already have some demographic information you can use to personalize your landing pages. Certain influencers only appeal to certain demographics. Gym bros are probably not watching eyebrow-styling tutorials and the eye-brow styling-tutorial watchers are probably not interested in pre-workout protein mixes. If you’re targeting both demographics, personalize accordingly. The same applies to Facebook and Google ads. These platforms allow very granular forms of targeting. You can and should use the targeting information from your ad campaign to personalize the landing pages for anyone who comes to your site from campaign pages.
How to collect data
Privacy laws have made third-party data collection very hard, so if you need to collect user data for personalization purposes, you have to do it in-house. This can be tough and costly for small operations but right now it’s the only way to do it without running into legal problems, especially in Europe. And California.