Prepare to Tune Search Relevance

Search relevance improves faster when you start with the right inputs. Before you adjust Search Fields, Ranking, Rules, or Promotions, confirm that SearchStax has useful queries, focused Search Fields, reliable structured fields, and indexed data to work with.

Use this article before you tune Ranking or troubleshoot result order. You will create a small test set, review Search Fields, check structured fields Ranking can use, and map each business goal to data SearchStax can actually use.

Build a Test Query Set

Start with real search behavior. Choose a small set of searches that represent what visitors commonly look for, what your business cares about, and what you need to validate during onboarding or migration.

Use sources such as analytics, stakeholder input, support cases, migration requirements, or known customer requirements.

Include:

• Common searches that many visitors use.
• High-value searches tied to business goals.
• Broad searches, such as events, programs, tuition, or admissions.
• Specific searches where you know which page or content type should appear.
• Migration searches you need to compare with a previous search experience.

For each query, write down the likely visitor goal, the content types that should appear, and any known pages that should be easy to find. Run the queries in the Search Profile you plan to tune, then record the current top 5-10 results. This baseline helps you tell whether later changes helped.

Review Search Fields

In the Search Profile, review the configured Search Fields before you adjust Ranking. Search Fields control where SearchStax looks for text matches. A field that is not selected as a Search Field cannot contribute text matching through Search Fields, even if the field exists in the index.

Useful Search Fields are usually short and focused:

• title
• description
• main headings, such as the page's H1 text
• short summaries
• page-specific text fields that describe the content

Body content can help visitors find pages that do not have strong titles, descriptions, or headings. It can also create noisy matches because long body fields contain many topics, repeated page elements, and common terms. If body content has too much influence, short searches can return many technically related results that are not useful enough for top positions.

When you review Search Fields, ask:

• Does this field describe the main topic of the page?
• Is it populated consistently?
• Does it help visitors find useful results, or does it create many weak matches?
• Do you need it for keyword matching, or only for display, filtering, or faceting?
• Does it support the queries in your test set?

If users need to filter by a value and search it as text, keep field type in mind. Filters and facets usually work best with string fields. Keyword matching usually requires tokenized text fields in Search Fields.

Check Structured Fields for Ranking

Some fields do not make good Search Fields but can still help Ranking. Structured fields can express business priorities more cleanly than broad body text.

Common structured fields include:

• content type or page type
• date or recency fields
• audience, category, topic, department, or program metadata
• location or service-line metadata
• URL or path information, when supported by your configuration
• other custom fields that represent business priorities

Use structured fields when they identify the content group reliably and the boost should apply broadly within the Search Profile. Check a sample of results that should receive the boost and a sample that should not. Confirm that the field value separates those groups consistently.

For example, if event pages should be more visible across an events-focused Search Profile where event-related content is important, a consistent content_type value is easier to test and maintain than a fragile URL pattern or a broad body-text match. If the behavior should apply only when a visitor enters a specific query or small query group, use a Rule instead.

Validate Indexed Data

Before you rely on a field, confirm that it exists in the Site Search index and contains reliable values.

Look for:

• fields that exist but are empty on many results
• important metadata that is missing entirely
• inconsistent values, such as Blog, blog, and blogs
• stale, missing, or inconsistently formatted date fields
• content types or categories that do not match how visitors think about the content

If the data is incomplete or inconsistent, fix the source data, content model, crawler or connector mapping, or indexing process before you rely on that field for relevance tuning. If you cannot inspect indexed fields directly, work with the team that manages your crawler, connector, or indexing configuration.

Map Goals to Fields

After you choose test queries and review fields, connect each business goal to a field or value SearchStax can use.

Ask:

• Which searches does this goal affect?
• Which content should be easier to find for those searches?
• Which field or value identifies that content reliably?
• Is that field useful for text matching, Ranking, or both?
• Should this change apply across many searches, or only to a specific query or campaign?

For example, a university may want prospective students to find admissions, application, program, tuition, financial aid, and event content more easily. Search Fields such as title, description, and main headings can help those pages match. Structured fields such as content_type and date_modified can help Ranking apply business priorities.

This mapping keeps relevance work grounded in user behavior. It also helps you avoid boosting a field only because it exists, rather than because it supports a specific search goal.

Confirm Readiness to Tune

You are ready to tune when you can name the queries you will test, the results you expect to see, the fields SearchStax should search, and the structured fields Ranking can use. If any of those pieces are missing, fix the data or narrow the goal before you change boosts. This keeps each tuning pass measurable and easier to explain.

It also gives reviewers a clear baseline when they compare result order before and after a change.

Next, use your test queries and field review to tune Search Fields and Ranking. See Improve Result Order with Search Fields and Ranking.

For background about how SearchStax uses Search Fields, field statistics, and boosts to calculate result order, see How SearchStax Calculates Relevance Scores.