Top Trends in Search for 2017

By on November 29, 2016

Search is changing.

For the past few years search has become much more relevant. There was a time when savvy users had to scroll past the first few results habitually, when composing the right query was as much of an art as it was about learning a secret parser grammar. Now, users are accustomed to looking at the first 3 results at most and generally just click on the first result. Search has gotten smarter and users have become accustomed to those smarts.

Some of these smarts have come because search has become much more personal. Search that captures the signals a user gives about their preferences and tunes future results accordingly tends to produce highly relevant results. For example, in online retail when I search for shoes, I expect it to bring up mens shoes first because I’ve been searching on men’s clothing. I expect products that come in my size to be boosted to the top.

That where we are today, as we look forward to 2017 what trends can we look forward to?

Search will drive big data

image00

The last few years of Big Data have largely been a freakshow compared to what is on the horizon. “Let’s Create a Data Lake!” this unindexed dump of data on a distributed filesystem… then load every bit of it into memory every time you want to analyze it. Sure your analytics might be fast now with Spark, but you still have to wait on I/O. Having the data indexed is the ticket.

Search will become more anticipatory

At Lucidworks we’ve long explained that search is more than the search box. Using signals and personalization and new techniques in machine learning, search will start to bring you results before you ask for them. When I go to a webpage about to click in a search box and realize that what I want is already staring me right in the face. There are many techniques for this. A tried and true method involves using a simple clustering algorithm to look at what similar users have searched on (or purchased) and boost those results for similar users.

Search will become more conversational

image02

Google “What is the distance to mars” and you get an answer. The answer isn’t a command but a specialized form of search, a boosted result and a special UI template. It looks almost like Google composed something for you. This is really just the start of more to come. In 2017, we will start to see this become ubiquitous even through enterprise search. We will start to see grammatical forms beyond “Who is, what is, how do I, etc.” Users will also come to expect this. Gone are the days of carefully crafted keywords and strange queries like +mars +“distance from earth.” Search is a query and a query is a question and the result is an answer. In talking to a person you don’t expect them to answer such a question with a list of webpages, just an answer. Why should it be different just because you asked a machine?

Search will continue to become speech

image01

Whether it is your phone, the integrated voice search on Google or Enterprise or the microphone button on your favorite e-commerce site, speech will continue to become speech. In 2017 this will accelerate.

Search will become even more ubiquitous

Whether it is your smartphone, your smart TV, you Amazon Fire Stick or your Amazon Echo, voice search is popping up everywhere and search will continue to become a bigger part of your life. Long gone are the days of the unanswered question that you forget on the long drive home.

I’m looking forward to 2017, the Year of Search. The year we index big data, have our customer needs anticipated, just ask the question we need answered using our voice and anywhere we damn well want to. If you’re interested in getting a jump on the new year by more relevant search results, enabling conversational search and voice search, I suggest checking out Lucidworks Fusion our Solr-powered platform for building and deploying powerful search apps.

Share on LinkedInShare on FacebookTweet about this on Twitter

Related Posts

Article: A programmer’s guide to big data: 12 tools to know

Top Posts

Using Word2Vec in Fusion for Better Search Results

Data Science for Dummies

The Data That Lies Beneath: A Dark Data Deep Dive

Visualizing Search Results in Solr: /browse and Beyond

Implementing Apache Solr at Target

Posted in Big Data, Topics and Technologies

Your email address will not be published. Required fields are marked *

*