Can you explain why that document ranked there?

Vespa.aiOct 27, 2026

Hi John,

How do you know why one document outranked another? How should different parts of a complex query influence ranking? And how do you investigate relevance when the results aren’t what you expect?

These are some of the questions we’ll explore in Vespa Now: Fall 2026 Product Update.

We’ll look at how new Vespa capabilities can help you:

Understand and customize ranking. Access statistics behind BM25 to investigate ranking behavior and build custom ranking functions.
Compose more sophisticated queries. Label query terms and subqueries so their contributions can be scored independently and combined in ranking.
Control the recall/latency tradeoff. Give ANN search an explicit time budget rather than relying only on fixed HNSW exploration parameters.
Work with richer data. Bring structured relationships into tensor-based ranking.

We’ll also cover updates to Vespa Cloud provisioning and OpenTelemetry-based telemetry export.

Register for Vespa Now

October 29
EMEA: 11:00 AM CET | Americas: 1:00 PM ET

Submit your technical questions for Q&A.

The Vespa Team