Which Search is Best for eCommerce?

Tait Weblog Leave a Comment

Experience counts for everything, and that goes equally for the user experience as for any other connotation of the word.

For the eCommerce business hoping to thrive, an effective search function acts as a critical cornerstone of any website.

Choosing wisely equips your site with a powerful tool to boost conversion rates and draw greater profits, while the wrong choice, providing irrelevant results or poor performance, practically guarantees customer dissatisfaction and soaring bounce rates.

With over 20 search engines to choose from, however, we’re faced with quite the challenge of choice. Which is why we’ve put together this comparison between 5 of the top contenders in the field.

Next to selecting a reliable inventory management software, the right site search is crucial for any eCommerce business to reach their full potential, but the decision should not be made lightly. You’re in this for the long haul, and many of these searches boast similar features. It’s in the minutiae the magic lies, but we’re here to make finding your perfect fit as easy as it can be.

The Lay of the Land

Before you can build, you have to know what you’re working with. So before we get into the searches, let’s look over some cold hard facts.

  • Search is the main navigation method for mobile users, the platform of choice for the majority of online shoppers today
  • The main aim of the search is retrieving the most relevant matches to user queries while omitting other generic website content
  • A secondary aim of the search engine is supplying suggestions based on user journeys and experience
  • 95% of product searches never go beyond page 1
  • 80% of visitors abandon a site whose search performance is poor
  • 40% of online purchases are influenced by the site search
  • 30% of site activities are swayed by site search

Now, we’ll take a glance at the most popular features shared by a number of search engines, so that we know which bells and whistles really stand out from the crowd.

  • Full-text search; the ability to handle everything from simple words and phrases to multiple forms of a word or phrase
  • Fuzzy search; recognition of typos, etc
  • Synonym search; identifying synonyms to best serve the query
  • Faceted search; a count of attributes and variations to help customers spot specific sizes, models or colours, etc
  • Filters; to narrow results for desirable parameters only
  • Highlighting; a visual signpost among the results of words used in the search
  • Multifield search
  • Auto-complete
  • Spelling corrections

Sounds familiar, right?

It should do. We’re all so accustomed to this kind of shopping experience that anything less is unacceptable.

For this comparison of searches, we’ll be looking at 5 of the most reliable and efficient providers you could ask for; Sphinx, Solr, Elastic, SLI and Fact-Finder.

First up, we have 3 open-source solutions, each with exceptionally competitive performance and features, and then to finish we’ll explore the 2 paid alternatives.

Sphinx

Those familiar with the search experience on sites like Craigslist, Joomla or Tumblr have felt the power of the Sphinx search solution first hand.

Like its namesake, Sphinx has been around for some time now, however, that should hold no weight in your decision as such time can lead to entrenchment in days gone by, rather than innovation.

Sphinx search supplies nearly real-time results and boasts the largest registered volume of indexing at an estimated, whopping 25 BILLION documents.

With this powerful search at your disposal, you can process over five hundred queries per second and reap the benefits of each of the aforementioned most popular features.

Their extensive experience covers handling the 300 million daily queries on Craigslist and delivering content to over 400 million monthly users on China’s biggest video site, Youku Tudou, which can peak at volumes of fifteen thousand queries per second.

Sphinx is truly deserving of its mighty name and longevity. They’ve honed in on their search function over the years to fully deliver in every expected user function.

One of their greatest weaknesses could almost be construed a strength, depending on your business goals. While this level of search focus delivers rapid results and queries huge data volumes and documents with minimal computing resource, their tunnel-vision like approach leaves their back-end data analysis somewhat lacking through what can only be assumed as simple neglect.

Which means, the business owner who values their data and the readily-accessible insights it offers might want to reconsider Sphinx as a suitable choice, and explore the alternatives.

Sphinx search is also not adept at handling unstructured data, such as .doc or .pdf files, images, videos or music. It will take developers much time and effort which, alongside other configuration challenges, brings the Sphinx solution in behind its other open-source competitors.

Solr

Solr boasts some formidable names in its roster of clients, from Magento to NASA, MTV, Disney and Apple, there’s not many of us who wouldn’t have experienced the Solr search capabilities in our time.

This open-source search tech uses the Apache Lucene library and comes with a rich set of highly configurable features.

Their rich full-text feature works straight out-of-the-box, and their solution is capable of supporting a whole host of suggester implementations, highlighting functions and spell checkers which even our pick of the bunch is lacking.

Solr search shines with its strong faceted searching, which makes it the perfect partner for fashion eCommerce sites whose users expect to whiz through hundreds of thousands of styles in seconds flat.

Example of the faceted search function on a fashion website.

Being one of the few search engines able to read rich content documents such as .pdf, .doc or XML files, Solr is ideal for businesses with a need to scan vast swarms of these document types. This means if your site processes contracts, learning materials or resumes, for example, Solr could be a fantastic fit.

Unlike Sphinx, Solr allows, through the use of external tools, the monitoring of events and logs in their dashboards, delivering those data insights we all know and love. This also means reduced stress on manual workloads, so your team will be happy.

Solr’s main downfall comes courtesy of its caching method. The search uses global caches, which means even the minutest change to the cache demands a full index refresh. To those in the know, this is a remarkably time-consuming undertaking.

For these reasons, and more we’re about to dive into, Solr only ranks a close second for our liking among the search giants of this article.

Elasticsearch

Rarely has a service lived up to the imagery of its name so thoroughly as with Elasticsearch. In fact, we’d go as far as to call the adaptiveness of Elasticsearch truly unrivalled.

Such a statement can easily be overlooked until laying eyes on their impressive user list and considering the varied services offered by those giant adopters.

From Netflix to TripAdvisor, Uber, Shopify, Etsy, SoundCloud, Yelp and the titan of industry that is eBay, it’s clear that Elasticsearch must be doing something remarkably right.

With near real-time indexing, you’re looking at less than 1 second to index rapidly changing data (a drastic improvement on Solr’s offering already!). This makes Elasticsearch perfect for projects with a database in constant flux.

Just ask Uber, who benefit immeasurably from the search’s powerful ability to aggregate business metrics on fluid pricing and availability at the drop of a hat. Arguably, their phenomenal service would be distinctly lacking without this formidable search engine.

Thousands of queries per second during peak times is child’s play to this robust tool, and you can be assured of maintaining those high-speed results as your database grows thanks to exceptional scalability. Expedia would surely struggle with managing their tremendous swells of users without Elasticsearch backing their every move.

Beyond its unequalled search abilities as an indexer, Elasticsearch can also act as additional data storage for logs, further boosting those rapid results.

But perhaps their finest strength lies in how they provide back-end data. This is where big names such as Netflix and The Guardian really reap the rewards of this sublime solution. It’s obvious when these areas have been built with the user experience, your experience, at the forefront of their mind, and Elasticsearch truly stands as an example for emulation.

With data visualisation clearly designed for supreme ease of access and real-time tracking, the savvy business owner can observe the most crucial data at a glance, from total or unique visitors to popular queries, most requested pages, devices and browsers used, timestamps and traffic logs – the whole unseen world of your eCommerce site is laid at your fingertips with Elasticsearch.

How Netflix would stay ahead of the curve without this ability to study their enormous droves of data and understand viewing activities, error logs, performance and diagnostics is a nightmare I’m certain they’re thankful doesn’t plague them.

And if that wasn’t enough, their real-time security analytics are guaranteed to plug any holes in the boat, helping you identify server problems, broken links or external attacks before they ever become a true threat.

In fact, Dell attributes a 30% server load reduction and a security boost to Elasticsearch.

With machine learning plugins and an Amazon service which operates and scales this search in the cloud, without any need to configure your own servers, Elasticsearch’s only real weakness is in its youth. But with youth comes ambition, and that’s something these guys have in spades.

SLI

SLI stands for Search, Learn, Improve, and that’s exactly what this search engine delivers most strongly on.

This AI-driven predictive search engine delivers results with an enhanced level of personal touch by focusing heavily on learning from customer behaviours and preferences when they visit your site. A handy tool indeed for those hoping to create a sense of familiarity and loyalty among their audience.

With its personalised results, engagement tools and strong mobile solution, SLI is highly adept at spiking conversion rates by helping customers find exactly what they’re after and converting window shoppers into paying customers.

What’s more, SLI comes with an integrated SEO strategy which gathers data from your customers to help you rank higher in Google, placing you in front of those who are ready to buy. This however is possible with data gathered from the other contenders.

While they cannot claim as illustrious a client list as Elasticsearch, those familiar with Everlast, Bestway or City Chic, to name a few, have experienced the power of this search engine up close.

Despite the attentive strength of its behavioural learning and personalisation, SLI doesn’t offer enough in our eyes to stand apart from the open-source alternatives capable of the same (and more).

Fact-Finder

Similarly to the search engines, we’ve covered already, choosing Fact-Finder will land you with a powerful onsite product search and navigation solution. 

Their search is able to deliver consistently regardless of language, though favours European countries; features admirable behavioural learning and personalised results, capturing customer data and serving up suitable products with targeted relevance; and offers a terrific recommendation engine to suggest complementary products and services.

These traits set it nicely in line with SLI’s offering, although Fact-Finder earn a leg up in the competition by offering a free trial for their software.

This, combined with their predictive basket, which takes AI learning to the next level, offering reminders of potentially overlooked items, and drag and drop usability, sets Fact-Finder as a stronger choice between the 2 paid search engines. And, it’s many out-of-the-box features make setup and configuration a relative breeze.

Hobbycraft certainly place great faith in Fact-Finder, attributing a boost of up to 170% conversion rates to the search engine.

But it is enough to stack up to the might of any of the open-source alternatives? Well, that entirely depends on your primary business goals. From where we sit, though, the choice is fairly clear.

Conclusion

Of course, there’s a lot to consider when making such a critical choice for the future of your eCommerce business.

As many of these search engines share functionality with one another, we’ve also only highlighted certain standout traits of each, along with some of their more famous users to give each service a more relatable feel.

For ease of reference, we hope this comparison table will help you reach your final decision.

 SphinxSolrElasticsearchSLIFact-Finder
Open SourceYesYesYes
PaidYesYes
Real-time IndexingYesYesYesYesYes
Data VisualisationYesYesYesYes
Machine LearningYesYesYesYes
Scalabilityhighhighhighhighhigh
Full-text SearchingYesYesYesYesYes
Auto-complete SuggestionsYesYesYesYesYes
Faceted SearchYesYesYesYesYes
Multi-fieldYesYesYesYesYes
SynonymsYesYesYesYesYes
FuzzyYesYesYesYesYes
HighlightingYesYes
SpellcheckerYesYes
Unique FeaturesLargest registered # of indexingRich content reading (.pdf/.doc etc)Security analytics, Amazon service, Doubles as storageAccount ManagedAccount Managed
Corporations AdoptedCraigslist
Tumblr
Disney
Apple
NASA
MTV
AOL
Netflix
eBay
TripAdvisor
The Guardian
Etsy
SoundCloud
Yelp
Everlast
Bestway
City Chic
Hobbycraft
Lidl
Esprit

From our perspective, Elasticsearch stands proud and tall, head and shoulders above the rest, and they don’t look set to slow down any time soon.

Leave a Reply

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.