Overview
Nextens, part of LexisNexis Risk Solutions, is a leading Dutch provider of tax solutions and digital products for fiscal and financial professionals. Its offerings span tax software, knowledge products, and tools used daily by accountants, tax advisors, and finance teams. Search is one of the primary ways users access that value.
The platform is used for tax research and advisory work, where users depend on fast access to accurate, up-to-date information.
As Nextens expanded its corporate offering, search quality became a meaningful product and business concern. Customers relied on search to find current, relevant fiscal guidance, but were encountering outdated results, duplicate editions, and lower-value content in prominent positions. Internally, Nextens had already connected search and discoverability to retention and growth, and recognized that without improvement, this would impact the product’s long-term adoption.
Searchplex was brought in to audit the system, identify the root causes, and help restore confidence in a retrieval experience that had become central to customer value.
Context
Nextens’ knowledge products combined multiple content types — laws, rulings, expert articles, almanacs, and practical materials. Many of these assets were versioned or updated annually.
Users expected the system to:
- return the most relevant current result by default
- surface a specific yearly edition when the query indicated that intent
- avoid returning multiple near-identical versions of the same content
- support simple queries from users without specialist search knowledge
These expectations were already visible in internal product thinking before the engagement began.
Challenge
The visible issues were consistent:
- irrelevant results ranking too highly
- older editions outranking current content
- duplicate versions appearing as separate results
- weak query assistance
- inconsistent metadata and unclear scoring behavior
Incremental tuning had already been attempted, but it did not resolve the underlying problem. The team could not reliably explain why results ranked the way they did, which made systematic improvement difficult.
What appeared as a search-quality issue was, in practice, affecting product trust, day-to-day usage, and retention.
What Searchplex Found
Searchplex conducted a detailed audit across data structures, metadata usage, ranking logic, and query handling.
The core issue was not a single ranking flaw. Nextens was operating a high-trust, versioned knowledge experience in a search setup that did not provide sufficient clarity or control over retrieval behavior.
The existing implementation, based on Azure Cognitive Search, had evolved from an earlier proof-of-concept path. Over time, it had become difficult to reason about scoring behavior, metadata impact, and query handling in a consistent way.
The audit made one point clear: improving relevance required a more controlled and inspectable retrieval foundation.
Approach
Searchplex approached the engagement as a retrieval redesign and replatforming effort.
Retrieval Audit
A structured analysis of the full search system to separate symptoms from root causes before proposing changes.
Architecture Recommendation
Searchplex recommended moving to an owned Lucene-based search architecture. This provided stronger control over indexing, ranking behavior, APIs, and ongoing evolution, and aligned with the need to deliver improvements quickly and reliably.
Search-Oriented Data Model
A data model designed for retrieval, with improved handling of versioned content, metadata quality, and ranking signals.
Functional PoC
A functional validation of the new approach before full migration.
Migrating Production Search
Full corpus migration, search API implementation, ranking improvements, and query suggestions delivered as a complete working system.
The focus throughout was to establish a system the team could understand, operate, and improve with confidence.
Outcome
Within roughly six months, Nextens moved to a production-ready retrieval system aligned with its product needs.
The engagement resulted in:
- improved handling of versioned content
- more consistent and predictable ranking behavior
- cleaner search results with reduced duplication
- query suggestions to support discoverability
- a replatformed search foundation with stronger operational clarity
The result was not only improved search behavior, but a retrieval layer the team could reason about and build on.
Why This Matters
In knowledge-intensive products, search quality often becomes visible as a business issue before it is fully understood as a technical one.
In this case, search relevance influenced trust in the product, day-to-day usage, and retention in a key segment. Addressing the problem required more than tuning — it required a retrieval foundation that supported clarity, control, and continuous improvement.
Tax & Regulatory Technology
Azure Cognitive Search
Search Audit, Retrieval Redesign, Replatforming
Versioned knowledge system with limited visibility and control over retrieval behavior
Audit, architecture redesign, Lucene-based replatforming, and full implementation
Improved relevance, clearer ranking behavior, and a retrieval system the team could operate and evolve with confidence
If this case reflects the kind of retrieval issues your team is dealing with, start with a structured diagnostic before jumping into a migration or rebuild.