<p>Recommendation engines are becoming increasingly associated with new on-line marketing strategies. This is primarily due to its capacity to display to the user products that excite his interest and capture his attention. As advertisers are not required to bid for their ad to be displayed by the recommendation engine, this further contributes to increasing the advertisers' sales without adding to their burden. A recommendation engine's ability to successfully suggest items to users depends on the item selection made. This paper aims to present an adaptive filtering-based technique that will improve results and increase publisher (search engine) and advertiser profit. When making recommendations, this method considers the user's previous browsing history and the keywords entered into the search. Matrix factorization and an agglomerative hierarchical clustering technique produce the findings. The suggested system and the experimental findings are discussed in the paper to demonstrate the enhanced performance of the indicated task.</p>