We are excited to share an innovative approach to optimizing Google Shopping feeds using Feedoptimise and OpenAI's ChatGPT. We focus on leveraging AI to significantly boost traffic and conversions by optimizing item titles.
Strategy Overview:
The goal is to identify high-potential items and experiment with three different title versions. We'll analyze each version's performance and automatically adopt the title that delivers the best results.
We'll target items with a high conversion rate (numerous conversions per click) yet a low Click-Through Rate (CTR). This usually indicates that while these products are popular upon viewing the product page, they are often overlooked in ads, potentially due to suboptimal titles.
Process Steps:
1. Identifying High-Potential Items and set the first condition to initiate A/B testing:
The above condition has 4 rules:
- Keeps the final best-performed title post-testing.
- Maintains the current title during the testing cycle (lasting 7 days).
- Generates a new title for testing or for the subsequent cycle when applicable.
- Defaults to the original title as a baseline.
2. Generating Title Variations:
- Version 1 - Basic Optimization: Utilizes OpenAI's Title Optimizer to craft a straightforward title combining the product title, brand, and category.
- Version 2 - Enhanced Attributes: This version incorporates additional attributes such as color, material, and a dynamically rotating variable for the season (Winter, Spring, Summer, and Autumn).
- Version 3 - Custom Creative Approach: Here, we take a more unique route by crafting a custom prompt. The goal is to create an optimized title for the Google shopping feed, particularly appealing to younger audiences, using the given title and item description.
Prompt: Use the below item title and description to build the best possible optimised title for Google shopping feed that will bring the attention of young people | title: “{title}” and item description: {description}
3. Performance Measurement and Selection
We automatically record each title's performance over a 7-day testing period into an override sheet using our reporting functionality and Google Ads integration. At the end of the test, we implement a rule to adopt the title with the highest performance. This approach, in certain instances, has led to over 700% improvement in CTR, significantly boosting sales for the best-converting items.
This method offers a scalable, fully automated solution that can be effortlessly replicated across multiple campaigns and websites. By harnessing the power of AI, we're not just optimizing titles – we're revolutionizing the way products are presented and discovered in Google Shopping.