diff --git a/web/src/padelnomics/admin/routes.py b/web/src/padelnomics/admin/routes.py
index 996133b..138ed69 100644
--- a/web/src/padelnomics/admin/routes.py
+++ b/web/src/padelnomics/admin/routes.py
@@ -1181,9 +1181,11 @@ async def email_compose():
 @role_required("admin")
 async def audiences():
     """List Resend audiences with local cache + API contact counts."""
-    audience_list = await fetch_all("SELECT * FROM resend_audiences ORDER BY name")
+    # Cap at 20 — Resend free plan limit is 3 audiences, paid is more but still
+    # small. One API call per audience is unavoidable (no bulk contacts endpoint).
+    audience_list = await fetch_all("SELECT * FROM resend_audiences ORDER BY name LIMIT 20")
 
-    # Enrich with contact count from API (best-effort)
+    # Enrich with contact count from API (best-effort, one call per audience)
     for a in audience_list:
         a["contact_count"] = None
         if config.RESEND_API_KEY and a.get("audience_id"):
@@ -1263,21 +1265,20 @@ async def audience_contact_remove(audience_id: str):
 @role_required("admin")
 async def templates():
     """List content templates scanned from disk."""
-    from ..content import discover_templates, fetch_template_data
+    from ..content import count_template_data, discover_templates
 
     template_list = discover_templates()
 
-    # Attach DuckDB row counts
-    for t in template_list:
-        count_rows = await fetch_template_data(t["data_table"], limit=501)
-        t["data_count"] = len(count_rows)
+    # Single query: article counts for all templates — avoids N SQLite round-trips
+    counts_raw = await fetch_all(
+        "SELECT template_slug, COUNT(*) as cnt FROM articles GROUP BY template_slug"
+    )
+    article_counts = {r["template_slug"]: r["cnt"] for r in counts_raw}
 
-        # Count generated articles for this template
-        row = await fetch_one(
-            "SELECT COUNT(*) as cnt FROM articles WHERE template_slug = ?",
-            (t["slug"],),
-        )
-        t["generated_count"] = row["cnt"] if row else 0
+    # One DuckDB COUNT(*) per template (N queries, but cheap vs SELECT * LIMIT 501)
+    for t in template_list:
+        t["data_count"] = await count_template_data(t["data_table"])
+        t["generated_count"] = article_counts.get(t["slug"], 0)
 
     return await render_template("admin/templates.html", templates=template_list)
 
diff --git a/web/src/padelnomics/content/__init__.py b/web/src/padelnomics/content/__init__.py
index e86ad7e..469a0ad 100644
--- a/web/src/padelnomics/content/__init__.py
+++ b/web/src/padelnomics/content/__init__.py
@@ -128,6 +128,15 @@ async def fetch_template_data(
     )
 
 
+async def count_template_data(data_table: str) -> int:
+    """Return the row count of a DuckDB serving table. Returns 0 if unavailable."""
+    assert "." in data_table, "data_table must be schema-qualified"
+    _validate_table_name(data_table)
+
+    rows = await fetch_analytics(f"SELECT COUNT(*) AS cnt FROM {data_table}")
+    return rows[0]["cnt"] if rows else 0
+
+
 def _validate_table_name(data_table: str) -> None:
     """Guard against SQL injection in table names."""
     assert re.match(r"^[a-z_][a-z0-9_.]*$", data_table), (
diff --git a/web/src/padelnomics/worker.py b/web/src/padelnomics/worker.py
index 4c9762b..4561b2d 100644
--- a/web/src/padelnomics/worker.py
+++ b/web/src/padelnomics/worker.py
@@ -18,6 +18,7 @@ from .core import (
     init_db,
     send_email,
     setup_logging,
+    transaction,
     utcnow,
     utcnow_iso,
 )
@@ -589,18 +590,41 @@ async def handle_send_supplier_enquiry_email(payload: dict) -> None:
 
 @task("refill_monthly_credits")
 async def handle_refill_monthly_credits(payload: dict) -> None:
-    """Refill monthly credits for all claimed suppliers with a paid tier."""
-    from .credits import monthly_credit_refill
+    """Refill monthly credits for all claimed suppliers with a paid tier.
 
-    suppliers = await fetch_all(
-        "SELECT id FROM suppliers WHERE tier IN ('growth', 'pro') AND claimed_by IS NOT NULL"
-    )
-    for s in suppliers:
-        try:
-            await monthly_credit_refill(s["id"])
-            logger.info("Refilled credits for supplier %s", s["id"])
-        except Exception as e:
-            logger.error("Failed to refill credits for supplier %s: %s", s["id"], e)
+    Uses two bulk SQL statements instead of N×3 per-supplier queries:
+    1. INSERT INTO credit_ledger SELECT ... for all eligible suppliers at once
+    2. UPDATE suppliers SET credit_balance = credit_balance + monthly_credits
+    """
+    now = utcnow_iso()
+    async with transaction() as db:
+        # Batch-insert ledger rows for all eligible suppliers in one statement
+        await db.execute(
+            """INSERT INTO credit_ledger
+                   (supplier_id, delta, balance_after, event_type, note, created_at)
+               SELECT id,
+                      monthly_credits,
+                      credit_balance + monthly_credits,
+                      'monthly_allocation',
+                      'Monthly refill (' || tier || ' plan)',
+                      ?
+               FROM suppliers
+               WHERE tier IN ('growth', 'pro')
+                 AND claimed_by IS NOT NULL
+                 AND monthly_credits > 0""",
+            (now,),
+        )
+        # Update balances and refill timestamps in one statement
+        result = await db.execute(
+            """UPDATE suppliers
+               SET credit_balance = credit_balance + monthly_credits,
+                   last_credit_refill = ?
+               WHERE tier IN ('growth', 'pro')
+                 AND claimed_by IS NOT NULL
+                 AND monthly_credits > 0""",
+            (now,),
+        )
+    logger.info("Monthly credit refill complete — %d suppliers updated", result.rowcount)
 
 
 @task("generate_business_plan")