Implementing Site Search Through Ditto Output: What Actually Works

This topic repeated a pattern we have seen many times in old Evolution projects: the site frontend was heavily driven by Ditto, but the desired search behavior was described as “search through Ditto output”.

The Important Distinction

Ditto is a renderer. Search needs searchable source data. Once those two concerns are separated, the problem becomes much easier to solve.

What Usually Works Better

• search the source resources that Ditto is rendering
• index the relevant TVs explicitly
• use filtered datasets instead of expecting the renderer to become the search engine
• tune AjaxSearch or another search layer around the real content fields

Why Teams Get Stuck

On content-heavy sites, Ditto can make the frontend feel like it is the content source. But it is only assembling what already exists in resources and TVs. If those records are hidden, fragmented, or incompletely indexed, search quality falls apart.

The practical takeaway is consistent with other old search topics: model the data clearly first, then let the search layer query that model. Do not try to search the final HTML that a listing snippet produced.