What’s So Personal About Elasticsearch?
In this day and age, people expect a lot from the web. In order to keep up with the ever-changing needs of internet users, the websites of tomorrow need to provide a lot more than just content, their users will need to see “context” that surrounds the content! That’s where the “Big Data” solution comes in; it allows you to start analyzing your own content, user data and search behaviors deeper than ever before. If you’re new to “Big Data” this article from Doug Laney on the 3-V’s, for Volume, Velocity and Variety does a great job at explaining the “Big Data” approach in its entirety.
This technology has been around for years, but up until recent years only the big hitters such as Google, Amazon and Netflix have been using it’s concepts. And for the most part they have used Elasticsearch for search, analytics and logging across billions of records. However, it now has more powerful tools, from searching between multiple languages, geolocation, contextual did-you-mean suggestions and autocomplete. With just the search feature alone you can start to provide for your web audience with search refinement, which is a lot more powerful than simply searching across server error logs.
Elasticsearch vs Google
So what’s the difference between Google and Elasticsearch? Google provides a broad search engine for crawling websites and looking mainly at the data structure of HTML markup. Plus, Google is very limited on how much search refinement the end user can have. Google is also built from a money machine model, which means it’s trying to give you the top 10 results and then provide advertising that relates. Additionally, if you’re a website owner trying to make it into that top 10 you either have to pay for it or make sure your website is being constantly SEO optimized in an organic way.
Elasticsearch on the other hand is a powerful search engine built with the “Big Data” concepts in mind. Overtime, it’s designed to become a powerful matching machine with powerful refinement tools. Elasticsearch uses a schemaless architecture, meaning you can throw any form of data at it, and it will be able to crawl it fast. Elasticsearch is so fast that it can crawl through 500k records in a matter of seconds, see it’s power in action with this great demo. In this example, we can search and refine through thousands of artifacts, some of which have up to 70 fields and subfields of data.
Do I need Elasticsearch for my website search?
Ideally, it all depends on your project and your business needs. For most of our clients it’s not necessary; we can usually just depend on WordPress’s default search engine or Google’s Custom Search Engine approach. However, in recent months we’ve identified clients with a multitude of content and various channels of content on the web. We’ve also discovered with certain clients that they have a large array of different data types that we’d ideally like to search across. So, for them an optimized search process is an act of necessity. They can then start to dig deep and really analyze their end-users search behaviors to form new hypotheses on how they can better serve up their own content and the way they present it. Our clients are finding new ways to form design decisions with their website initiatives to make it a more personal experience!
Personalization and the Future of the Web?
Developers are a lot like scientists: we’re constantly learning, evolving and trying out new technologies. At Westwerk our development team is no different, and we’ve dreamed up one big idea that we feel could pack a lot of punch for your brand website and messaging. We call it “Custom Persona Experiences.” Our vision is for Elasticsearch to watch your user’s search behavior and create statistical models that will help identify trends within your “Big Data” collections.
You’ll discover the key data that will give you a better picture on what your visitors are looking for. By doing that you can create personalized experiences that will speak volumes to your audience. Right now we give our web audiences a search bar; why not give them a “matching engine” interface that asks them questions to enhance their experience? Where do these questions come from? We use the search terms as building blocks to form questions for the matching engine. This then creates an experience with more personalized direction on what they’re looking for.
Interested in learning more about Elasticsearch and the benefits it could provide your site? Drop us a line, we’re happy to chat!