Founded in 2013, Unbxd enables their customers to build and deploy search and discovery solutions for e-commerce. Customers laud the product’s ability to help them boost ROI for their commerce operations.
Unbxed was acquired by Indian-based provider Netcore Cloud in 2022, and while the future isn't yet clear how it will affect the search service, customers may already be looking for a new solution.
While Unbxd is capable of delivering good results, it comes at the customer's expense. Unbxd's relevance, AI, and ranking features are mostly manual, meaning customers must spend a considerable amount of time writing relevance rules, adding synonyms, and manually ranking results. Or, Unbxd may handle this, but ultimately the work is manual and subject to human decision-making which isn't as accurate as true AI-powered search.
Unbxd is built on Solr, which itself is built on top of Lucene, the "grandfather" of open source search libraries. Lucene-powered engines are good for sites with static content. However, they're less reliable for e-commerce use cases. See our write-up on Lucene to learn more.
If you're struggling with Unbxd and looking for something new, we've compiled a list of four alternative solutions to Unbxd below.
Search.io is a SaaS site search solution that's been built from the ground up and takes advantage of a cloud-native architecture for elastic scale. For non-technical users, Search.io offers a no-code UI that allows for deep customization and configuration. For more technical users, there is a simple, flexible YAML-based config interface.
Search.io includes all the basic search and discovery features — faceted search, spell checking, autocomplete, natural language processing (NLP), merchandising, personalization, etc., — but also offers unique AI capabilities and dynamic ranking to boost revenue and conversions. This includes:
- Neuralsearch®: A proprietary search technology that delivers highly relevant results in milliseconds. It eliminates much of the hard work previously required to make search work well. For instance you no longer need to add synonyms, complex search rules, keyword-stuff titles, or plug-in additional language packs.
- Reinforcement learning: Search.io was the first site search engine to offer reinforcement learning, an approach to ranking optimization. Different ranking algorithms can even be A/B tested to determine which one(s) deliver the best performance.
- Dynamic boosting: No-code settings that allow customers to rank results even higher based on positive actions such as shopping cart ads, purchases, etc.
Unlike traditional search engines that have layered AI on top of their search algorithms, Search.io has included Neuralsearch and reinforcement learning as part of its core search algorithm.
Algolia is a general purpose search engine built for technical users. Algolia has flourished as a new type of search engine relying on a pre-sorted Radix Trie index. By choosing the ranking (sorting) logic upfront and storing the information in the ideal sort order, the index is very fast. This has some downsides but it is incredibly cheap and very simple for people to understand.
The sorting operation is defined by the person who designed it and doesn’t offer any additional intelligent sorting. Yes you can run A/B tests, but aside from that it is only as good as the rules and sort logic that someone wrote. For search this is typically a bad idea, as one person’s opinion is notoriously bad at representing a broad group of customers.
While Algolia enjoys popularity due to its ease of use and fast retrieval speed, it has its critics too, particularly around complexity for managing custom rules and pricing. For example, anytime Algolia needs to re-index your content — such as for A/B testing — it can double or quadruple your usage, which can quickly exceed monthly quotas.
Like Unbxd, Searchspring is a search engine built for retailers. It is a custom enterprise solution that is built-to-fit for each client. Once configured, customers have reported being extremely satisfied with reporting, personalization, and other features. However, unlike the pure software solutions from Search.io and Algolia, which can be more easily customized, Searchspring is not as automated.
- Adjustments are not automated, keyword, synonyms, and search mappings must be entered manually — a time consuming and expensive process.
- Re-indexing can be slow and results may not be reflected on the site for a while.
- Customizing facets and filters may require professional support rather than self service.
Like Searchspring, Klevu is built for e-commerce use cases. Also like Searchspring, there are no self-service options; Klevu builds custom, bespoke solutions through customer success services. The company gets rave reviews from users on their customer service, professional services, and support.
Klevu is built on top of Solr, which itself is built on top of Lucene. While these platforms are very flexible for building search, it also has the disadvantage of limiting Klevu's ability to finely optimize its search algorithms for speed and adding native AI capabilities. Customers are essentially buying a product with three abstraction layers — Lucene, Solr, and Klevu — which slows application speed and affects performance.