Find out how PrettyLittleThing achieved:
Attraqt’s existing search engine optimises the vast majority of PrettyLittleThing’s on-site searches. PrettyLittleThing, in collaboration with Attraqt, wanted to research options into how to deliver more accurate and relevant shopping experiences for the most complex and unstructured search requests. Long-tail searches are difficult to optimise though they make up 15% of the overall search volume. Often, these low confidence searches represent significant untapped revenue to the business.
Attraqt worked with PrettyLittleThing to review search patterns on their UK and French sites. They focused on the 15% of searches that were identified as complex and longtail.
Attraqt deployed its new AI-powered search approach, collecting and harnessing all available product data, from all signals, and transformed search requests made on the PrettyLittleThing site into relevant search results for the shopper.
Attraqt’s AI-powered search works by scanning all available information about a product before combining natural language processing, deep learning and computer vision to build a comprehensive semantic representation of the products.
These self-learning algorithms now work in the background, tailoring each query to the PrettyLittleThing customer’s intent and behaviour, ensuring continuous optimisation of search results and the discovery experience.
“Attraqt has been a trusted partner of ours for many years, so when the opportunity rose to implement their newest technology for Search we couldn’t pass this up. We were particularly impressed by how quick and simple the implementation of the technology was, and how well it worked with our existing data feeds. We started seeing significant improvements in affected searches pretty much straight away and the results achieved over just few weeks of the algorithms running speak for themselves.”Leanne Clancy, IT Director, PrettyLittleThing