[K]ampaignMate – AI-powered campaign optimization with ChatGPT
We developed [K]ampaignMate, an AI-driven solution powered by ChatGPT, to automate and optimize large-scale campaign structures – delivering impressive results across all key metrics.
Mission & Solution
To broaden a campaign structure and find new relevant keywords to a category or a product, we came up with an AI-driven solution to source keywords based on product meta data such as description, specifications and more.
[K]ampaignMate can create a large scale account setup in hours rather than days. We let the large language model AI ChatGPT find both the best performing keywords based on the products meta-data and add those to the products ad group.
[K]ampaignMate created hyper specific ads for products. Ads that then are shown to potential customers where the product should be the most relevant.
Key Metrics
Higher
Click Through Rate
Lower
Cost Per Acquisition
Higher
Profit On Ad Spend
Results
The test was a BIG success, resulting in an impressive performance increase across all key metrics:
31%
CTR increase
52%
CPA decrease
100%
POAS increase
How It Works
The solution leverages automated longtail keyword generation combined with AI-powered keyword relevance scoring. Product descriptions are fed into ChatGPT which generates relevant keywords with search volume. These keywords are then automatically added to the campaign structure, creating hyper-targeted ads at scale.
Example: Coolstuff Norge – Sokker med eget foto
To make the flow concrete, here is how [K]ampaignMate processes a single product from Coolstuff in the Norwegian market – the product Sokker med eget foto. The same logic is then repeated across the entire product catalog.
Fetch product meta-data
The script pulls the product's title, description, category and key specifications directly from the PDP.
AI generates keyword candidates
ChatGPT analyzes the meta-data and proposes a broad list of relevant keyword candidates in Norwegian.
Validate against search volume
Each candidate is checked against actual monthly search volume. Keywords below 100 searches/month are filtered out.
Build hyper-specific ad
The 5–6 most relevant keywords with >100 searches/month are selected, and an ad is built with a landing page going straight to the product.
Selected keywords (NO)
For this product, the script kept the following keywords – each with more than 100 monthly searches in Norway:
- sokker med eget bilde100+ /mo
- personlige sokker100+ /mo
- sokker med eget foto100+ /mo
- sokker med trykk100+ /mo
- personlige sokker med bilde100+ /mo
Resulting ad
From those keywords, [K]ampaignMate generates a hyper-specific ad that lands the user directly on the product page – no detours through category pages:
Sponsored
www.coolstuff.no/p/sokker-med-eget-foto
Sokker med eget foto – design dine egne | Coolstuff
Lag personlige sokker med eget bilde eller trykk. Rask levering i hele Norge. Perfekt gave – bestill dine egne sokker hos Coolstuff i dag.
Multiplied across thousands of products, this is what drives the campaign-level results for Coolstuff's personalized socks and the rest of the catalog – a 31% CTR increase, 52% lower CPA and 100% higher POAS.