In this post from our Magento Site Search series we will check another search solution that has cloud deployment can be integrated with Magento eCommerce – Algolia Site Search for Magento.
General search quality
Search relevance and ranking
Algolia has Magento demo website with default installation of Magento quite fresh version of Magento eCommerce with sample products. Let’s try to search for “dress” there.
The results are quite good – we see actual dresses there and they are at the beginning of the list. For some reason product thumbnails are missed, but this is demo store glitch.
Plural form search, “dresses” returns the same results, so another tick there.
What we don’t see at the search result page however is product category called “Dresses&Skirts” that exists in the sample store, only the products are listed. We’ll return to this, when review auto-complete functionality.
Also, what we’ve noticed about Algolia, at least at the demo instance is that it doesn’t look into product detailed description at all when conducts the search.
So if you look for “feminine frames glasses” the product above will be found. However such search as “Acetate frame glasses” brings no results.
In our opinion the better practice is giving detailed description text less importance than product name or short description, but do not ignore it at all.
Understanding of typos and mixed word orders
As you may see on the screenshot above Algolia Search for Magento understands typos quite well, even if the mistake is done at the begin of the word. We’ve searched for ‘cress’ having ‘dress’ in mind and got list products where words as ‘dress’ and ‘cross’ were encountered.
Mixed word order
Our text searches for “cotton shirt”/”shirt cotton” and “dress maxi”/”maxi dress” returned the same results, good work here.
Understanding of jargon and synonyms
This test was failed by all solutions we’ve tried before, so it’s interesting to see is Algolia Search better here.
Unfortunately the test has been failed. We tried to search for “pants” and ‘trousers” and the results we got were completely different as you may see on the screenshots above. Search for ‘strides’ (Aussie slang word with the same meaning) doesn’t bring any results at all. Another try with searching for ‘jumper’ as synonym of ‘sweater’ also wasn’t successful. So, the fail here.
No good surprise here too, as well as competing solution Algolia Site Search for Magento hasn’t worked out our abbreviation test. We searched for “8 gigabyte”, but didn’t see in search results matching product – “8 GB memory card”.
Search with ” (inches) like “20” x 20″ pillow” doesn’t bring us expected results, moreover, quotes character was translated in not very user-friendly form in search refine box as you may see on the screenshot below.
The overall impression from Algolia Site Search quality and relevance is mixed: search relevance in general, work with mixed word orders and understanding of typos are quite good, but other little things that make a perfect search are missed. Not having other than products type of content at search result page is a flaw in our opinion, you can’t rely only on auto-suggest box.
Auto-complete and auto-suggest
Auto-complete and auto-suggest works together in search box for Algolia Site Search.
As you may see from the screenshot above it works across CMS and categories, not only products, interesting difference from search results page. Algolia Search auto-complete starts to work from the first letter you type (with other solutions it typically starts from 2nd or 3rd letter) and tries to smart match you with the most appropriate content. Another interesting feature, that it is auto-complete and auto-suggest all-in-one, so you’ll see possible typo matches in the search box, which sometimes may look counter-intuitive: you may have longer list of suggested results for longer word than for shorter one.
Search query pre-fill on search results page, faceted search
As we mentioned above there is a flow in search results page – only product-related results are displayed – no CMS pages and categories. Other than that it looks good – there is a special field to refine your search with pre-filled original query as well as layered navigation/faceted search functionality.
Non-functional information about Algolia Site Search for Magento
Algolia Search can be deployed in two ways, as classical cloud solution and at dedicated infrastructure (but in Algolia’s data centres) for Enterprise clients.
Algolia Search as cloud solution has multiple price tiers, including free one, that may work for smaller stores. The price depends on the size of search index, number of operations and additional features such as regional replications. Paid plans starts from $49 per month and can be as high as $1785 per month.
For Enterprise clients who need more powerful solution than public cloud can handle, Algolia offers dedicated server infrastructure with starting price of $3,490 per month.
Ease of installation
Algolia offers free Magento extension to make integration easier. For Enterprise clients and more expensive cloud plans UX/UI guidance and review service is included in price. According user reviews, the installation is comparatively easy.
Magento extension popularity and review score
Algolia Search for Magento currently has 725 Popularity Score (23 November 2015). The review rating is 4 stars (out of 5).
Last release of Magento extension
The last version of Algolia Search Magento extension, # 1.4.8 was released 19 October 2015.
Company or individual behind the solution
Algolia Site search is developed by Algolia, a privately held company with offices in Paris, France and San Francisco, USA.
Stores that use this solution
Algolia is implemented on number of high profile Magento eCommerce websites, including Quicksilver and Birchbox.
In spite of some flows, Algolia site search for Magento is a decent solution with quite good search quality. It’s not very expensive either, so many websites with moderate traffic and small database can use it for free or on lower-cost plans, which is good for start-up businesses.
It would be good if the vendor can improve search results page and see will the search relevance still be good if the broader set of data is indexed.