Overall Satisfaction with Algolia
Algolia powers all of our search functionality on our browse page and from the header across the site. Specifically, we use the react instant search package with a custom hierarchical menu for category selection and react locations for location selection and the default pagination. We also use Algolia for our search auto suggestions and recommendations. We also use Algolia to solve interesting business problems and tasks such as categorizing user-generated item titles.
- Extremely fast search results
- Very easy to set up, no maintenance needed from our side
- Efficient and easy to adapt ranking formula
- Customising the react packages can be difficult. Easy to send an excessive amount of requests
- Customising ranking is restricted to prioritising factors, rather than weighting each factor individually
- Costs can escalate dramatically when you expand your traffic
- Decreased SEO landing page load speeds, due to faster search results.
- Increased user conversion from search
- Reduced server costs
Algolia was by far and away the easiest of the three to implement. PostgreSQL has many search modules that can be used on top of your usual database, however, none are particularly efficient and can quickly become overwhelmed at scale. Equally, they do not handle business metric weightings well. Elastic search is what Algolia is built on and works well, however, it would require a lot more manual setup to ultimately get to a solution that is similar and likely worse than Algolia offer. Both Algolia and elastic search suffer from the problem of having to keep two items databases in sync.
Any consumer-facing business that needs instant search results for user searches. Excellent fits would be e-commerce solutions and marketplaces. If your traffic is low it is also a relatively cost-effective solution than using your own solution (e.g. elastic search). Perfect for teams that do not have the architectural knowledge to set up their own search solutions.