50 -100% improvement in search relevance from Hyper-local Language Models
ChatGPT has permanently changed our expectations of technology. Before it, people might not have known that computers can spit out fully formed, seemingly thoughtful answers, to complex questions. It showed the world the endless possibilities that open up when people can interact with large bodies of information on their own terms – speaking in plain language, asking questions as they might to a relative, friend or teacher.
People have been talking to Alexa and Siri for a few years now, and maybe some even know that both of those technologies are, at their heart, search engines – just like ChatGPT. Google can quickly get decent answers to complex questions. Our everyday interactions with machines are starting to become more and more like the conversations we have with each other.
So, why is the keyword search that powers so much of the web outside of Google and Amazon so awful? If the natural human language that instructs Alexa to turn on lights, or Siri to schedule a wake-up call could be used for guiding product discovery, our cumbersome and irritating process for finding and buying products could be improved immeasurably. Plain language in…great results out.
Go beyond keyword Search and reimagine product discovery
If we could search on most ecommerce websites the way we find and discover things in real life, our experiences would not only be better, the merchants who run those sites would enjoy more sales. In real life, we might walk into a hardware store and ask an agent for a tool to accomplish some task. In a popular hardware store, a query for “backed up toilet,” produces these results:
In their everyday interactions, people use meaning to find people, things and knowledge. Semantic Search lets people engage in the same way online. The result is instant relevance, which is a far better experience for users, but is also far better than keyword search for sales conversions.
Going beyond keyword search has until now mostly been an expensive and time-consuming affair that only the most deep-pocketed could afford. Usually, the platforms which promised semantic understanding were jargon-laden, scary and unapproachable. Under the hood, they were also the same old keyword search, based on good ol’ Lucene. Sprinkling a little bit of “vector search” pixie dust on text-based keyword search doesn’t automatically get to the promised land, however.
In order for semantic search to truly work for everyone, it needs to be easy, quick, cost-effective, transparent and integrated. If it did all of these things, then it would be possible for any retailer to incorporate meaning-based search, and truly reimagining product discovery. We’re going to show it’s possible to get a lot more out of current search platforms – all without breaking the bank. First, dig into the benefits.
Better product discovery = better conversions
Searching by meaning is good for the user because it lets a person use the search box to input queries just as she might to another human being. She doesn’t need to know the exact name of the product she might be looking for, can use slang or other colloquialisms, describe what she needs to get done instead of what the name of the thing is and more. The result is an intuitive relevance with clear benefits.
- Eliminate zero search results. One of the most persistent annoyances from keyword-based search is when users can’t find any results for their query terms. When users don’t find anything, they don’t buy anything, and they usually get frustrated and leave the site too. Semantic Search solves multiple problems that lead to zero search results. Misspellings, mismatches in vocabulary, slang and thematic expressions are all handled out of the box, resulting in 90%+ reduction in zero search results.
- Fix common search problems. Finding irrelevant products is almost as bad as not finding anything at all. With Semantic Search, retailers can instantly improve some of the frequently poor-performing search problems including thematic searches, non-product queries, slang, abbreviations and symbols, product-type queries. These improve search-driven click-through rates up 20%, and increase search-originated orders 30%
Get more bang from your existing search platforms
Semantic Search can be a drop-in integration into your current search engine. As long as your search engine can serve vectors, we can easily integrate with it and your ecommerce tech stacks. We even bring the data and pre-train models with your own data. There is virtually no implementation time, and you can enjoy the quick revenue boost from instant relevance improvements.
With vector embeddings for Solr and Elasticsearch, you can get drop in a preloaded set of embeddings into Solr or Elasticsearch and be up and running in minutes. We offer a fully managed service, take care of operations, capacity and scaling.
Contact us today for a free assessment and find out how we can help you improve relevance instantly.
Speakers Description In this webinar we discuss MC+A’s new solution and approach for Intelligence Experiences for investigation use cases using LLM technology and machine learning. We demo how the solutions can act as a catalyst in expediting investigative processes. LLM technology assist with investigation due to its ability to understand, interpret, and analyze vast swathes of data, thereby aiding in
In this article, we’ll compare the performance of OpenAI’s API versus Microsoft’s API when utilizing GPT-4.
Join our panelists for a webinar where they discuss approaches for improving relevance for e-commerce search. They will cover ELAND and ELSER, promising to reshape the relevance landscape with vector-based search. Don’t miss out on an interesting discussion that could change your approach to e-commerce search relevancy.
Go Further with Expert Consulting
Launch your technology project with confidence. Our experts allow you to focus on your project’s business value by accelerating the technical implementation with a best practice approach. We provide the expert guidance needed to enhance your users’ search experience, push past technology roadblocks, and leverage the full business potential of search technology.