SpotSaaS is a software comparison and recommendation platform that helps businesses evaluate SaaS products through reviews, feature comparisons, and buying guides. With over 40,000 software products, 500+ categories, and more than a million reviews, the platform needed significant performance improvements to deliver fast and reliable software discovery experiences.
On average each company relies on more than 20 SaaS applications to run day-to-day operations. Selecting the right software for a use case is difficult when faced with over a thousand options, and the failure mode is quiet: a suitable product exists, but insufficient research leads a buyer to a poorly suited one first. The platform meant to solve that had become too slow to use for it.
Each company relies on more than 20 SaaS applications spanning recruitment to sales, and choosing correctly is hard when a single category holds over a thousand options. The failure is rarely that no suitable product exists. It is that one exists and the buyer never reaches it, then lives with the alternative for years as switching costs accumulate immediately.
Numerous videos and blogs offer insight on software selection, but keeping that content current is a significant challenge, since products change pricing and features continuously. Influencers also exhibit bias towards preferred solutions, and a buyer generally cannot see the commercial relationship behind a recommendation they are reading.
SpotSaaS developed its MVP in-house, which was right for reaching market, but the platform hit performance problems from handling extensive data. This is the characteristic failure of a successful data product: an MVP validated against a small catalogue behaves entirely differently against 40,000 products and a million reviews.
API response times reached approximately 6 seconds and page loads around 9 seconds. Comparison is inherently iterative, since a buyer filters, adjusts, and compares repeatedly before deciding. At 6 seconds per call, trying five comparisons costs half a minute of waiting, so the platform was least usable during the exact behaviour it existed to support.
The reflexive fix is more infrastructure, but the causes here were structural rather than volumetric, so scaling up would have bought a temporary gain at permanently higher cost. Eternalight attacked performance at three levels instead: what the system holds in memory, how the APIs are shaped, and how the database is queried. Each addresses a different cause, and fixing only one would let the others reassert themselves.
A robust structural caching system was implemented using Redis and MongoDB, with multiple APIs rearchitected to store caching data and burst strategies added. A comparison catalogue caches unusually well, since product data, categories, and review aggregates are read constantly and change rarely, so most requests never reach the database at all.
More than 40 APIs and their input parameters were refactored, introducing pagination and additional filters for richer frontend functionality. Pagination changes the shape of the problem, not its size: an endpoint returning a whole result set does work proportional to the catalogue, so it gets slower as the business succeeds, while a paginated one stays constant.
Legacy queries were streamlined by eliminating multiple cross-joins on tables and implementing multiple indexes on MongoDB. Cross-joins are the pattern that turns a manageable dataset unmanageable, since work multiplies rather than adds as tables grow. Without indexes, finding records means examining every one of 40,000 products.
SpotSaaS built a tool letting companies search for software solutions and their features and compare features and pricing across vendors. This answers the bias problem rather than the performance one, since comparison presents underlying facts for the buyer to weigh, where a recommendation asks the buyer to trust the recommender instead.
API response times fell from approximately 6 seconds to 400 milliseconds, a fifteen-fold gain that significantly increased user engagement and active time on the website.
After the comprehensive API refactoring, page load time dropped from about 9 seconds to 1.5 seconds, further improving the experience for buyers working through comparisons.
The improvements produced a 12 percent reduction in database costs alongside notable gains in query efficiency, which is the signature of fixing causes rather than adding capacity.
Eternalight rebuilt SpotSaaS on Node.js and Express.js with a React.js frontend, written in TypeScript, over MongoDB for the catalogue and Redis for caching. Work covered structural caching with burst strategies, more than 40 APIs refactored with pagination and filters, and legacy queries streamlined by removing cross-joins and adding MongoDB indexes to cut both latency and database cost.
Brainstorming
Wireframe
Development
User Feedback
Launch
Integration
A cross‑platform, open‑source JavaScript runtime environment for building fast, scalable server‑side applications.
React: A component‑based UI library for building fast, interactive web interfaces with a virtual DOM.
Lightweight framework for building backend APIs.
A document‑oriented NoSQL database storing JSON‑like data for flexible schemas
Type-safe frontend development for improved reliability and maintainability.
An in‑memory data store for caching, queues, and real‑time analytics with sub‑millisecond latency.
Send us a message, and we’ll promptly discuss your project with you.
