# E-commerce personalization

> E-commerce personalization means showing each visitor different products, content and offers based on their past orders, the products they browsed and what similar customers did. Corrplus combines your store and marketplace order data, connects a recommendation system to your site, email and WhatsApp campaigns, and measures the effect with controlled tests.

URL: https://corrplus.net/solutions/ecommerce-personalization/
Updated: 2026-10-22
Corrplus Team

## Symptoms

- Every visitor sees the same products on the home page and product pages.
- Campaign emails go to the whole list with the same content.
- Most first-time buyers never come back.
- Customer data from marketplaces and your own store sits in different places.
- Related product fields are filled in by hand and go stale.

## What waiting costs

Showing every visitor the same shop window means customers leave without ever seeing the product they would have bought. As paid traffic gets more expensive, that loss costs more.

## Our approach

- **Combine the data**: We bring order, catalogue and browsing data from your store platform and marketplace panels into one structure.
- **Start with rules**: Simple recommendations such as best sellers, frequently bought together and in-stock alternatives go live first.
- **Recommendation model**: If there is enough data, a model that learns from customer behaviour is built and compared with the rule-based version.
- **Connect the channels**: Recommendations feed the site, email and WhatsApp messages, with marketing consent and cookie consent built into the flow.
- **Controlled testing**: Every change is compared against a control group, and its effect is measured before it becomes permanent.

## What to measure

- Average order value
- Visit-to-order conversion rate
- Repeat purchase rate
- Share of sales from recommended products
- Click rate on campaign emails and messages

## Where should personalization start?

Personalization often brings to mind a complex AI system. In many stores, though, the first gains come from much simpler steps: showing products that are actually bought together, suggesting an in-stock alternative when an item is sold out, or reminding a first-time buyer about a related category.

Most e-commerce companies in Türkiye sell both on their own site and on marketplaces such as Trendyol and Hepsiburada, and many European sellers do the same on Amazon or Zalando. Customer data ends up in pieces. The first job is to bring those pieces together and see which products are bought together, and by whom. Whether you can use a marketplace customer's contact details for marketing depends on the marketplace's rules and data protection law, so we draw that line at the start.

The same logic applies in tourism and hospitality. Putting villas with pools first for a guest who looked at family rooms, and central rooms first for a business traveller, is the same catalogue ordered differently for different people.

## Which setup option fits you?

| Option | When it fits | Data needed | Control |
|---|---|---|---|
| Platform's built-in recommendations | Small catalogue, new store | Low | Limited |
| Off-the-shelf personalization service | Quick start, standard needs | Medium | Medium, data sits with the vendor |
| Custom recommendation model | Large catalogue, several channels, data must stay with you | High | Full, needs upkeep |

Whichever you choose, we test every change against a control group. Whether a recommendation box really adds sales only shows when you compare it with a group that did not see it.

## How we start

We review your store and marketplace data together and find where personalization will matter most: product page, basket, email flow or WhatsApp. The channel and campaign side runs through our [customer experience and digital marketing](/services/customer-experience-digital-marketing/) work, and the on-site build through [WordPress and WooCommerce development](/services/wordpress-web-development/). If you want an assistant that recommends products in conversation, an [AI sales assistant](/services/ai-sales-assistant/) is the natural next step.

After the first call you get a short proposal covering which data is ready, which channel to start with and which measure will judge success. If we think you do not need a model yet, we say so.

## Frequently asked questions

### Can we personalise our marketplace storefront?

No. Marketplaces such as Amazon, Trendyol or Hepsiburada control their own storefront and recommendations. You can still combine marketplace orders with your own customer data to improve product content, campaign planning and offers to customers who come to your own store.

### Which store platforms does it work with?

Platforms with an API, such as WooCommerce, Shopify, and Turkish platforms like ikas and Ticimax. For a custom-built store we work through the database or data exports.

### Is it worth it for a small store?

With little order data, a complex model has nothing to learn from. Rule-based recommendations and well-designed email flows are usually enough at that stage. A model becomes the next step as data builds up.

### What about consent for personalised marketing?

Marketing messages need recipient consent; in Türkiye that consent must also be registered in the national message management system (İYS). Browsing data needs cookie consent, and profiling needs a privacy notice under GDPR and Türkiye's data protection law, KVKK. We settle the details with your legal counsel.

### Does an AI chat assistant count as personalization?

It can. A sales assistant that answers a customer's question with the right product from your catalogue is personalization in conversation. Our own product, Promptor, is an AI chat and sales assistant for WordPress and WooCommerce sites.