The Trouble With Search

For users, the best search systems are essentially invisible. With a query in mind, a user types in keywords, and if search results provide the answers they’re looking for, no consideration is given to the mechanism behind the machine. Nothing is more aggravating than zero results -- except for way too many.

For search practitioners, creating a system that intuitively knows what each individual user needs requires extensive experimentation and tweaking. Endless query analysis, rule-writing and manually creating dictionaries and thesauri were the traditional way to connect a query writer’s intent with applicable, personalized responses. 

Four ACTIVATE speakers shared challenges they face as search practitioners -- and how they overcame them.

1. Big Goals, Limited Resources - Mark Desormeau, Lenovo

   “Our mission for search encompases many different uses cases, everything from supporting ecommerce transactional goals, to branding recruiting and prompting thought leadership across many industries. This is a very broad                             mission that we have to support with a small team.”

 

2. Shifting User Expectations - Ankit Patil, StubHub

  “My biggest challenge is building a robust model immune to unpredictable and rapid changes in user expectations”

 

3. Longer Implementation Cycles - Candace Marks, Getty Images 

  “Managing stakeholder expectations of turnaround time on DS efforts is challenging. The time it takes to produce models, generate data, and iterate is longer than typical agile sprint cycles.”

4. Turning Ideas Into Reality - Sava Kalbahou, Lucidworks

   “The biggest challenge is to productionalize state-of-the-art algorithms, moving them from research / proof of concept stage to the real end-user product.”

 

Want to know how these obstacles were overcome? At ACTIVATE, experts in search, Solr, and Fusion will provide real world solutions. Explore all of the sessions by topic and sign up for ACTIVATE by September 6 to save $200 on your pass.