Algolia being a SaaS hosted service API, its pricing is 100% based on usage of their service.
Remember, Algolia is an API, and as such you will have to pay for whatever you do with it. Including indexing, querying, or even getting statistics.
Algolia pricing changed a lot during the past few years, and even if they decided to be more transparent and easier to forecast, there are still some pitfalls remaining that could be a bad surprise at the end of the month.
Pricing is based on so-called “UNITS“. A “UNIT” can be free (each plan provides 10 units for free), or cost $1/month to $1.50/month, depending on your current plan.
1 UNIT = 1,000 search requests AND THE CAPACITY TO INDEX UP TO 1,000 RECORDS.
Be aware that indexing is taken on your UNIT queries quotas: if you have 1,000 records, and reindex them every day, you will consume 1 unit every day.
Also, notice that sorting with Algolia is very costly, as every sort order is cloning your main index in a new optimized index. Therefore, if you sort by price ascending and price descending, you will use 1 UNIT for your main index, and 2 other units for your 2 extra cloned indexes.
But then, does Algolia worth its pricing?
The answer depends on your needs of course.
If your business requires a first class search, with typo tolerance in 60+ languages, then Algolia provides just that out of the box. This is especially worth it if you do not have a lot of data to index frequently, let say up to 50K documents. At this level, Algolia is not only good, but can also be less expensive than some simple Elasticsearch or Solr plans, with the advantage of “ready to use” of course. This is so true that we decided to use Algolia here, at wpsolr.com, instead of Elasticsearch. The quality of suggestions is very nice, and Algolia does not cost much at all for our few thousand of posts and forums topics.
If you have a WooCommerce with 500K products, Algolia will be expensive. But it could still remain worth the price, depending on how profitable is your search to your business.
“Ultra fast”, “Super easy to implement” and “Easy, fast and relevant installation” are the key factors for which developers consider Algolia; while “Powerful API”, “Excellent Search Engine” and “Open Source” are the main reasons Elasticsearch or Solr is favored.
Out of the box relevancy: Algolia wins
Elasticsearch (and Solr), or more exactly Lucene engine, is a great tool for big data analysis, but it is very difficult to achieve great relevance with it in WordPress research.
You can try adding logic on top of Elasticsearch or trying to rearrange the results of some queries, but it’s a tedious job that still needs to be sorted out.
Algolia on the other hand focuses on very good relevance with minimal configuration. While not optimal for all use cases, it makes it particularly suitable for WordPress searching.
Open source: Solr wins
Algolia is a black box, their software is proprietary and no one officially knows what their engine look like behind the scene.
This is inconvenient compared to Apache Solr and Elasticsearch, both being Open source (respectively Apache-license and SSPL license).
The main drawback of proprietary is to be locked-in to a provider: what will happen if you cannot afford anymore to pay for Algolia’s service?
Well, there is a good news for you: WPSOLR can run you current WordPress search with al three engines, which means you can just reindex your data in WPSOLR and instantly switch from Algolia to Elasticsearch or Solr (or the other way round).
Of course, you would loose the out of the box relevancy of Algolia, but nothing prevents you from tuning your new index to your liking and get as close from Algolia’s quality as you want.
Analytics: Algolia wins
Algolia is pretty much out of reach when considering reporting and Analytics.
Out of the box, Algolia provides not only Analytics report on your indexes and queries, but also let you act on your results to improve your results. For instance, you can find which queries yield no results to prevent visitors bounce. But also one can reorder search results query by query to display some results on top based on arbitrary reasons, like advertising or best sales.
Algolia analytic summary
Algolia analytic searches
Algolia analytic searches without results
Algolia analytic filters
On this subject, Apache Solr delivers absolutely nothing.
And Elasticsearch provides very nice reporting tools, but mostly oriented to multi-dimentional analysis, like data-mining security logs in search for patterns. Search analytics is pretty much left empty.
NLP (Natural Language processing): Algolia has 70 languages OOTB
List of supported languages with their associated language ISO code:
Create and delete indexes with one click.
Automatically install the index configuration, facets, searched fields and settings.
By default, the English analyser is configured.
Connect to your Algolia admin dashboard to control your search.
Set your own languages, synonyms, stop words, typos.
Choose post types, taxonomies, and custom fields to search.
Manage custom field types, and conversion errors.
Choose among several indexes.
Useful for multi-languages, or testing.
Click on one button to index, re-index, or delete your data.
Choose Incremental or full reindexing.
Index in real-time, or not.
Index by batches, and select the batch size.
Show debugging informations for troubleshooting.
Choose post types and taxonomies to index, re-index, or delete.