What dynamic content is
Every element of a static site is the same for everyone. Dynamic content inverts that: a banner, a headline, a widget or a whole section renders differently depending on who is looking at it.
In practice that means:
– A new visitor sees a banner offering 10% off their first order
– A loyal customer with history in the Sports category sees new sports arrivals
– A visitor with an abandoned cart sees a reminder of what is in it
All on the same page, with no separate URLs.
Implementation: three models
| Model | How it works | When it fits |
|---|---|---|
| Client-side | JavaScript in the browser swaps the content after load | Simple swaps; risk of content flicker |
| Server-side | The server returns personalized HTML | Complex logic, SEO-critical pages |
| Edge-side | Personalization on a CDN node before delivery | Minimal latency, A/B at the infrastructure level |
For e-commerce the standard to aim at is server-side or edge-side: no substitution delay, no flicker, and the base version still indexes correctly.
E-commerce use cases
The homepage. The highest-priority surface: the hero banner, the featured categories and the top recommendation block all change per segment. A new visitor gets bestsellers and social proof. A returning one gets the continuation of an unfinished journey.
Promo banners. One banner slot × several audiences: shoppers of women’s clothing see the new collection, electronics shoppers see a sale on accessories.
CTAs and headlines. Not signed in: “Create an account and get a discount.” Signed in: “Your personal picks are ready.”
Tip: start with two segments — “new visitor” versus “returning visitor with purchase history”. It is the simplest split with the largest potential lift.
Common mistakes
- Personalization with no fallback. When there is no data on the visitor, the system has to show the base variant, not an empty block.
- Personalization that is too aggressive. A “you viewed this product” reminder on every page of a session reads as surveillance rather than service.
- Ignoring anonymous traffic. Most visitors are not signed in. Even minimal personalization by traffic source, device or geo already moves the numbers.