Running a Search Platform, Not Just an Index
Search support has barely changed in a decade. You configure an index, point a crawler at your content and hope for the best. When relevance degrades, someone raises a ticket. It is time for a better model.
The "Set and Forget" Problem
Traditional search management is a cost centre disguised as infrastructure. Organisations configure a search engine during a website build, tune it once based on a handful of test queries, and then leave it alone until complaints reach a critical mass. The conversation is always about index size and query latency rather than whether users actually find what they need. Teams treat search as a feature rather than a platform. There is no backlog of relevance improvements. There is no regular review of query analytics. There is no governance framework that connects search performance to business outcomes. Search stagnates. Users learn to work around it. The investment in search technology delivers a fraction of its potential value. Everyone senses this is not working, but the alternative has never been clearly defined. What does it actually look like to run search as a managed platform rather than a configured feature?
Run: Query Performance, Index Health, Crawl Coverage
The Run layer covers everything needed to keep a search platform healthy, fast and comprehensive. Query performance monitoring means tracking response times under real load, not synthetic benchmarks. Index health means verifying that document counts match expectations, that schema changes have not silently dropped fields, that stale content is being expired on schedule. Crawl coverage means confirming that every content source is being discovered, that new pages are indexed within defined SLAs, that redirects and canonical tags are being respected. This also includes security and access control: ensuring that authenticated content is not leaking into public search results, that PII is not being indexed where it should not be, that API keys and crawl credentials are rotated on schedule. This is not glamorous work, but it is the foundation everything else depends on. If users cannot trust that search results are current, complete and fast, no amount of relevance tuning will recover their confidence.
Improve: Analytics That Surface Relevance Gaps
The Improve layer is where value compounds. Each month we review search analytics and identify the highest-impact opportunities for relevance improvement. Zero-result queries tell you what users expect to find but cannot. High-refinement queries tell you where the first set of results is not good enough. Low click-through on top results tells you where ranking is failing. Search exit rates tell you where users gave up entirely. These signals are gold for anyone willing to act on them. Improvement might mean adding synonym mappings so "pricing" and "cost" return the same results. It might mean boosting recent content for time-sensitive queries. It might mean restructuring facets so users can narrow results more effectively. It might mean implementing promoted results for high-value queries that consistently underperform. The point is that improvement is systematic, prioritised by measured impact and delivered continuously. Not a wish list that sits in a backlog until the next redesign.
Govern: Findability Metrics Tied to Business Outcomes
Governance is the layer most organisations are missing from their search operations. Monthly operational reports that show crawl coverage, index health, query performance and relevance metrics. Quarterly reviews that connect search analytics to business outcomes: did improved findability for product pages correlate with conversion improvements? Did reducing zero-result rates for support queries reduce call centre volume? Are content gaps identified through search analytics being addressed by the editorial team? Without governance, Run becomes routine maintenance and Improve becomes ad hoc tuning that nobody can prove was worthwhile. With governance, search becomes a managed business asset with clear metrics, defined ownership and demonstrable impact. That is the difference between running an index and running a search platform. The index stores documents. The platform delivers findability. And findability, measured and governed, is a conversation worth having at the executive level.