Blog

Announcing Quepid For Elasticsearch

vectors are fun

Carefully prune the bonsai tree with Quepid for Elasticsearch 🙂

Good news! We’re proud to announce our test-driven search toolbench Quepid now supports Elasticsearch.

Quepid solves fundamental problems that come with improving search results. Search tweaks are easy, yet getting all your searches returning the right results can be maddeningly difficult. You can never be sure you’re moving forwards or mired in quicksand of constantly breaking search.

Quepid helps by bringing test-driven principles to tuning search results – what we call Test-Driven Relevancy. It helps define what good search results by incorporating your own business expertise from colleagues that know users the best: your merchandisers, domain and content curators. With this intelligence encoded in a tool, it gives the developer the ultimate test-driven tool for improving search transparently in a way the whole team can understand.

For example, you’re in charge of the online retro VHS store Movie’s R Us. You need to be sure that when users type “Rambo” they get back the right results, including Rambo I-III and most importantly the classic Rambo movie “First Blood”. But then requirements change, as they do for all kinds of software. Users come in and type “Sylvester Stallone,” expecting to see his films. Suddenly you need to solve a different problem. Can you boost and tweak search and ensure “First Blood” still comes back first for “Rambo”? Can you measure the quality of the search results as you tune? What happens when you add more search forms? Other search queries? Different use cases?

Quepid Screenshot

While your juggling all these different use cases, you’re effectively engaged in programming work. Yet its a very peculiar kind of work. You’re programming a conversation with your users. The one your app has every day, far more frequently than your in-person conversations! It’s a conversation with your specific user audience, your content, and your own specific business aims. Moreover, it’s a kind of programming where possibly only about 20 lines of “code” (queries, analyzers, mappings and the like) have an impact on every use case. Going backwards somewhere is almost guaranteed.

You need tools to make the whole process transparent and measurable to the whole enterprise. Quepid shows the whole team exactly why Elasticsearch thinks certain results are the right area. It measure the impact of many queries simultaneously, showing what’s improving and what’s getting worse. In short, it makes the whole process measurable and transparent. It’s crystal clear exactly where you’re succeeding in meeting customer expectations and where you’re failing – and why!

We hope you’ll give Quepid a spin. Get in touch if we can help you in anyway with Quepid or your tough search problems. We love problems like this one or this one and probably yours as well 🙂 Happy Searching!