Table of Contents

1. Ecommerce Search and Product Discovery
2. Algolia vs Searchspring vs ExpertRec: At a Glance
3. Algolia: API-First Search for Highly Customised Experiences
4. Searchspring: Search Combined With Merchandising
5. ExpertRec: Search, AI Relevance and Ecommerce Controls without Building the Stack From Scratch
6. Search Relevance Is More than Keyword Matching
7. Search Analytics: The Part Merchants Should Not Ignore
8. Which Search Architecture Fits Your Ecommerce Stack?

Algolia vs Searchspring vs ExpertRec: At a Glance

Algolia: API-First Search for Highly Customised Experiences

It is basically an API-first search and discovery layer. A standard implementation can be described as follows:

Such architecture offers engineers granular control over their search experience.

On the technical side, Algolia encompasses such concepts as hybrid search, semantic retrieval, AI ranking, query categorization, personalization, facets, synonyms, rules, and merchandising.

Its current AI Search combines keyword accuracy with semantic/vector methods and moves beyond literal keyword matching.

Its merchandising function enables the user to pin particular products, boost or bury categories, configure filters, reorder facets, and apply dynamic re-ranking rules.

For instance, if an electronics retailer receives

“4K monitor for video editing at the price tag of ₹35,000”

this advanced search implementation should recognize:

• Product type = monitor
• Resolution = 4K
• Use case = video editing
• Price constraint = ₹35,000
• Potentially relevant attributes = refresh rate, panel type, and color accuracy

The engineering team is now positioned to investigate the influence of these signals on retrieval and ranking. This flexibility helps when search is an element of a larger composable commerce architecture. However, the downside is that it means more work regarding indexing, front-end integration, ranking configuration, event instrumentation, and continuous relevance management in general.

Searchspring – Search Combined With Merchandising

Searchspring’s approach can be defined through e-commerce merchandising. It combines search, merchandising, personalization, recommendations, and analytics.

The API can be used with search result pages, category pages, and content results, while the Beacon API records events like searches, clicks, items added to cart, and purchases.

The layer of merchandising is essential to e-commerce teams, as Searchspring helps with campaigns affecting search and category results through product boost, pinning, banners, and contextual campaigns.

Let’s take an example of a furniture shop searching for:
“compact writing table for less than ₹8000”

The search engine has to provide products that fulfill this functional query; after that, the merchandising layer can determine which commercially viable products are presented first. This can be any product meeting such criteria as stock availability, margin, or promotion.

Additionally, Searchspring gives data related to search revenue, the most popular searches, spelling corrections, zero search results, and product performance. Its dashboard gives an efficiency metric in computing the performance of searches, based on existing consumers who used search results versus those who opted out of using the search option entirely.

One case study of Searchspring clients, BaxterBoo, tells us that 18% of visitors happened to use the site search, resulting in 44% of all revenue generated through search. These are unique outcomes from one client, not an overall standard, but they show the importance of search analytics

ExpertRec: Search, AI Relevance and Ecommerce Controls without Building the Stack From Scratch

ExpertRec revolves around simplifying complex searches and product finding for eCommerce teams instead of asking them to create the whole search architecture from scratch.

Its present capabilities cover semantic search, autofill, error correction in keywords, configuration of the search field, filters, synonyms, personalization, AI identification of intentions, visual search, A/B testing, analytics, and API/headless integration.

The technical benefit is becoming more evident when analyzing the product data.

For example, a jewellery shop sells a product:
“A925 Sterling Silver Necklace – CZ Pendant – 18 Inch.”

Shoppers might search the phrase: “silver pendant chain for no more than ₹3,000”

Should the search engine consider the product name only in a word-for-word manner, the search output will be weak. A properly built-up search channel will leverage structured fields such as:

• title,
• SKU,
• brand,
• category,
• tags,
• metafields,
• description,
• price,
• and availability.

With ExpertRec, merchants are able to decide which features of products become searchable, such as title, SKU, category, tags, and custom fields. This functionality is especially useful in cases where SKU is a signal of intent on its own. For instance, “LAP-7842” should bring up exactly that variant instead of many products sharing similar descriptive text.

Search Relevance Is More than Keyword Matching

One way to assess these platforms is to analyze their whole relevance pipeline:

The query itself can be something like:
“wireless keyboard for Mac under ₹ 2,500”
has so many signals in it:

• Category of products: keyboard
• Type of connectivity: wireless
• Compatibility: Mac
• Limitations in terms of price: ₹2,500

So, the search engine should search before ranking starts.

The difference between retrieval and ranking is crucial: while retrieval determines which options can be considered for ranking, ranking orders possible results.

Algolia has adjustable ranking and AI re-ranking options. Searchspring has merchandising tools to change the order of results. ExpertRec has search weight, semantic search, intent detection, and merchandising options for making retrieving relevant results easier.

Search Analytics: The Part Merchants Should Not Ignore

Search tools can provide merchants with useful information on why searches are not yielding results. More than search rate should be analyzed; other metrics should be monitored:

• Searches with zero results
• Searches not clicked
• Search hit rate
• Search-to-cart performance
• Search-to-purchase performance
• Popular filters
• Frequent spelling mistakes made by users when searching
• High-volume queries without sales made
• Products that get views but no clicks
• Income coming from searches

ExpertRec platform provides tools for gathering search metrics, tracking queries that give no results and clicks, as well as making corrections, analyzing filters, and calculating conversion rates with integrated results in the GA4 platform

That allows creating an optimization loop:

For instance, let’s say the store has 20,000 search sessions monthly that give a search conversion rate of 4% and an average order value of ₹2,500:
20,000 × 4% = 800 sales
800 × ₹1500 = ₹20,00,000 revenue generated by search
If it turns out that Controlled Testing allows us to raise search conversion by 0.4%, the difference will be:
880 × ₹2500 = ₹22,00,000
Thus, the difference will be ₹2,00,000.

Which Search Architecture Fits Your Ecommerce Stack?

The choice of search architecture mainly relies on how much of the search architecture your organization would like to own.

Algolia is the right pick for those organizations that want to benefit from a highly programmable and API-first search architecture and have engineering resources to create customized experiences.

Searchspring offers the best search and e-commerce merchandising and personalization experience since the merchandising layer is the highlight of the platform.

ExpertRec builds powerful search, AI-based relevance, merchandising, and analytics capabilities into one solution and minimizes the hassle of building and maintaining the underlying search.

When evaluating search options for various shopping platforms such as Shopify, WooCommerce, Magento/Adobe Commerce, or custom-built shops, it is important to look at the following questions:

The best search implementation is not always the one that boasts the longest list of features. It is the one in which all elements – product information, retrieval, ranking, merchandising and analytics of customers’ activity – come together to form an efficient revenue-generating process.