diff --git a/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentDensityResult.java b/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentDensityResult.java
index 521feca5..8e695589 100644
--- a/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentDensityResult.java
+++ b/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentDensityResult.java
@@ -19,4 +19,9 @@ public record DocumentDensityResult(
         List<MonthBucket> buckets,
         LocalDate minDate,
         LocalDate maxDate
-) {}
+) {
+    /** The "no documents match the filter" result, with no buckets and null date bounds. */
+    public static DocumentDensityResult empty() {
+        return new DocumentDensityResult(List.of(), null, null);
+    }
+}
diff --git a/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentService.java b/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentService.java
index 3f0d5ece..cca184d8 100644
--- a/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentService.java
+++ b/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentService.java
@@ -147,29 +147,41 @@ public class DocumentService {
      * parallel native-query path.
      */
     public DocumentDensityResult getDensity(DensityFilters filters) {
-        String text = filters.text();
-        boolean hasText = StringUtils.hasText(text);
-        List<UUID> rankedIds = null;
-        if (hasText) {
-            rankedIds = documentRepository.findRankedIdsByFts(text);
-            if (rankedIds.isEmpty()) {
-                return new DocumentDensityResult(List.of(), null, null);
-            }
+        List<UUID> ftsIds = resolveFtsIds(filters.text());
+        if (ftsIds != null && ftsIds.isEmpty()) {
+            return DocumentDensityResult.empty();
         }
+        List<LocalDate> dates = loadFilteredDates(filters, ftsIds);
+        return aggregateByMonth(dates);
+    }
+
+    /**
+     * Returns the FTS-ranked document IDs when {@code text} is non-blank, or {@code null}
+     * when no full-text query is active. An empty list means the FTS query ran but
+     * matched zero documents — the caller short-circuits on that signal.
+     */
+    private List<UUID> resolveFtsIds(String text) {
+        if (!StringUtils.hasText(text)) return null;
+        return documentRepository.findRankedIdsByFts(text);
+    }
+
+    /** Loads matching documents and projects to non-null {@link LocalDate}s. */
+    private List<LocalDate> loadFilteredDates(DensityFilters filters, List<UUID> ftsIds) {
+        boolean hasFts = ftsIds != null;
         Specification<Document> spec = buildSearchSpec(
-                hasText, rankedIds, null, null,
+                hasFts, ftsIds, null, null,
                 filters.sender(), filters.receiver(),
                 filters.tags(), filters.tagQ(),
                 filters.status(), filters.tagOperator());
-
-        List<LocalDate> dates = documentRepository.findAll(spec).stream()
+        return documentRepository.findAll(spec).stream()
                 .map(Document::getDocumentDate)
                 .filter(Objects::nonNull)
                 .toList();
-        if (dates.isEmpty()) {
-            return new DocumentDensityResult(List.of(), null, null);
-        }
+    }
 
+    /** Buckets {@code dates} into one {@link MonthBucket} per YYYY-MM and computes min/max. */
+    private DocumentDensityResult aggregateByMonth(List<LocalDate> dates) {
+        if (dates.isEmpty()) return DocumentDensityResult.empty();
         Map<String, Integer> counts = new java.util.TreeMap<>();
         for (LocalDate d : dates) {
             counts.merge(YearMonth.from(d).toString(), 1, Integer::sum);