Last summer KMW technology formally announced a professional services partnership with Lucidworks. The partnership was an obvious choice. Lucidworks is the company most invested in the maintenance, growth and success of the open-source Solr and Lucene Apache projects and KMW has long had a reputation as a premier firm for building amazing things on Solr. We’ve also known a lot of the people at LucidWorks for ages and we like working with them.
In addition to commercially supporting Solr, Lucidworks created the Fusion search platform which is a serious player in the eCommerce and Enterprise Search markets. In the last decade, we’ve seen a substantial shift in the search market as Fortune 500 companies started migrating en masse from the traditional commercial search vendors (Fast, Endeca, Autonomy) to open-source software projects like Solr/Lucene. The interesting play Lucidworks made was to create a commercial Search Platform around the open-source search engine (Solr). One of the most compelling selling points is that a company can keep an existing Solr server/collection “as is” and deploy Fusion to simplify cluster management and leverage a component set which enables the customer to more easily enhance the existing search index. It’s a strong approach to providing a commercial grade product while embracing the use and prolificacy of Solr as an open source search server.
Fusion offers all the components that Solr is missing that one would find in a traditional search platform; content ingestion and enterprise connectors, query manipulation and enrichment, monitoring, security, administration, etc. For a lot of companies this comes down to the classic enterprise IT build vs buy decision. If you’re looking to buy, Fusion is a solid option. Fusion version 5 is expected later this year and promises more cloud native capabilities especially around containerization and scaling.
At KMW we have supported many commercial search platforms (buy) but we also recognize the need for some companies to enhance home grown systems (build). Sometimes we consult with companies on that build vs buy decision, a lot of time we focus on the implementation of that decision to make search systems that work. When considering a buy option it’s helpful to consider the true market differentiators for a product. For Fusion a key differentiator is “Fusion AI” as an out of the box offering which enables capturing and learning from user Signals.
Signals refers to small “events” that occur as users interact with the search system. Examples include clicking on a search result and adding items to shopping carts. Machine learning techniques are then used on these signals to aggregate the information so that it can be used to inform relevancy calculations at query time. The end result is perpetually better search results for users based on what they or other users have done in the past. he concept of signals is not a new or unique approach to improving relevancy – at KMW we have implemented custom implementations of similar functionality as “adaptive relevancy” because relevancy gets better as more users interact with the system. What makes Fusion AI unique is by offering Signals based relevancy out of the box along with fully baked APIs for the collection of signals, ETL for transforming and enriching that signal data for maximum effect, ready to use (and customizable) Spark jobs for machine learning and hooks for leveraging signals data at query time. These are advanced relevancy capabilities suitable for an eCommerce giant made part of a commercial search engine. Signals / Adaptive relevancy is a major trend in search to watch. In addition to the partnership with Lucidworks, KMW was a sponsor at last year’s Activate conference in Montreal. At the KMW booth we got a chance to talk with a lot of Fusion users and saw a large representation of major retailers working to improve the relevancy of their eCommerce sites. In a world where Amazon’s search experience is the existential threat, it’s no wonder retailers are arming themselves with the best relevancy solution they can find. This trend is not limited to eCommerce either and we expect the tools and machine learning techniques that drive signals based relevancy to be a major need across industries in the next five years. It’s 2019, user’s expect excellent search results and business must meet this expectation or suffer the consequences, we at KMW can help.