Fix: Responsive Search Ads with fewer than 4 descriptions

A Responsive Search Ad with only 2 or 3 descriptions hands Google's combination engine half a deck — every dropped slot removes ad variants from auction-time testing and pushes Ad Strength toward "Average" [1][2]. Descriptions have only 4 slots, so each missing one is worth proportionally more than a missing headline. Fill all four with distinct angles.

Why this matters

RSAs are not a single ad — they are a generator that assembles eligible combinations of headlines and descriptions at auction time, conditioned on query, device, and audience signal. Description slots are scarce: only 4 exist, versus 15 for headlines, so each empty slot is a much larger share of the rotation surface. With 2 descriptions you give the engine 2 possibilities; with 4 you give it 4 distinct angles for the ML to test — value, proof, urgency, CTA — and the description chosen for any impression depends on what the system predicts will close that specific query [1].

The mechanics differ from headlines in three ways that matter operationally. First, only one description is guaranteed to appear — the second "isn't guaranteed to appear in your ads" and shows only "when there's enough space" or when the auction predicts it will lift performance [1]. Underfilling compounds: not only are there fewer combinations, the few you have spend more time as the only description visible. Second, descriptions sit lower in the ad and historically influence CTR less than headlines, so description quality reads as a tie-breaker in mid- and late-funnel queries where the headline already won the click. Third, Ad Strength explicitly weights description count and uniqueness — Google tells advertisers to "create unique headlines and descriptions that highlight various selling points and calls to action" [2].

Practitioner data sharpens the picture. Optmyzr's analysis of roughly 20,000 accounts reports that descriptions in the 61-70 character band deliver the highest CTR (around 12.3%) and the strongest ROAS, while very short descriptions (0-50 chars) win on cheapest CPA but underperform on engagement, and 81-90 char descriptions show diminishing returns with the highest CPC [3]. Filling all four slots is not a cosmetic completeness target — it is the only way to give the ML enough mid-length variants to find the band that works for your query mix. For the headline side of the problem, see Fix: RSAs with fewer than 8 headlines; for the rating layer that surfaces both, see RSA Ad Strength rated Poor.

How to verify the issue

1. Open Campaigns → Ads, filter Ad type to "Responsive search ad", and add the columns "Headlines" and "Descriptions" to the table. Sort by Descriptions ascending — every RSA with fewer than 4 surfaces at the top.
2. Click into any flagged RSA. The edit panel shows description slots 1-4; empty or duplicate slots are visible immediately.
3. Open the asset report: Campaigns → Ads → select the RSA → View asset details. Description assets show performance ratings (Best, Good, Low, Pending, Learning). If you only have 2-3 descriptions, you also have fewer than 4 rows here — the absence is the diagnosis.
4. Check pinning. Open the same RSA and look for the pin icon on any description. If descriptions are pinned to a single position, the combination math drops — pinning a description forces it to a fixed slot and removes other descriptions from that position pool [4].
5. Cross-reference with Ad Strength. An RSA with 2 descriptions and "Good" Ad Strength is rare but possible if headlines are strong; an RSA with 2 descriptions and "Average" or "Poor" Ad Strength is the common case and confirms the finding has real downstream impact.

How to fix it

1. Open the under-filled RSA. Navigate Campaigns → Ads → select the RSA → Edit. The descriptions block sits below the 15 headline slots.
2. Plan four distinct angles before writing. Descriptions are scarce — repetition is the most common mistake. Cover at minimum: a value/benefit description, a proof/credibility description (numbers, awards, social proof), an offer/CTA description, and a feature/use-case description. The four angles must be substantively different — Google's "avoid redundancy" assembly logic skips combinations the system reads as repetitive [1].
3. Aim for the 61-70 character band on at least two descriptions. Practitioner data shows this band is the sweet spot for CTR and ROAS [3]. The 90-char ceiling is available but should be used selectively — long descriptions get the highest CPC and weakest CPA in aggregate. Use one short description (<60 chars) for placements where space is tight, two mid-length descriptions for the main rotation, and one long description with detail for high-intent queries.
4. Leave descriptions unpinned unless legal or brand mandate requires it. Pinning a description to position 1 means only that description ever appears as description 1, and Google's Ad Strength documentation flags pinning as a downgrade signal because it constrains the combination engine [2][4]. If a legal disclaimer must appear, pin two or three descriptions to that position rather than one — Google allows multiple assets per position and recommends "pin 2 or 3 headlines or descriptions to each position so that any of them can show in that position" [1].
5. Save and re-check Ad Strength. The right-hand panel recomputes within minutes. Target "Good" minimum, "Excellent" where the angle inventory permits. If Ad Strength stays at "Average" after filling all 4 description slots, the bottleneck is on the headline side — see Fix: RSAs with fewer than 8 headlines and RSA Ad Strength rated Poor for the headline and rating dimensions.
6. Wait 7-14 days before judging performance. The combination engine needs impressions to learn which description works for which query. Description-level asset ratings stay in "Learning" until enough auctions accumulate, then move to Best/Good/Low. Pull underperformers only after the rating exits Learning.

How to confirm the fix worked

• All RSAs on the account show 4 descriptions in Campaigns → Ads view within 24h of save.
• Ad Strength on edited RSAs moved up at least one tier (Average → Good, or Poor → Average) immediately after save.
• No description is pinned unless a documented legal or brand mandate is on file in the account runbook.
• After 14 days, the asset report shows each description with a rating (not stuck in "Pending"). At least 2 descriptions per RSA are rated Best or Good.
• Description-level CTR variance across the 4 slots in the asset report is non-trivial — if all four descriptions perform identically, the angle planning was too uniform and step 2 should be re-done with sharper differentiation.
• No "Limited ad serving" or "Few ad combinations" warning on the RSA in the campaign view.
• On the campaign view, total RSA impressions for the edited ads grew over the 14-day window relative to the prior 14 days (typical lift is modest — descriptions are a tie-breaker, not a primary serving driver).

RSA description under-fill rarely appears alone in Whitead audits — it co-occurs with under-filled headlines (see Fix: RSAs with fewer than 8 headlines) and with RSA Ad Strength rated Poor, because the cause is usually the same: ads built once and not refreshed. We sequence remediation headlines-first (15 slots, drives impression eligibility), then descriptions, then pinning hygiene — and fix all three in one editing pass per RSA rather than across three sessions, because Ad Strength recompute and ML re-learning is more efficient when the ad is touched once. Blast radius for description-only under-fill is moderate: Ad Strength tier loss, modest CTR drag on impressions where the second description would otherwise have shown, and a less responsive ML when query mix shifts. The fix is mechanical and high-leverage when bundled with the sibling headline fix.

Related rules + concepts

• Fix: RSAs with fewer than 8 headlines — sibling rule on the headline slot economics; same ML logic, different scarcity math.
• RSA Ad Strength rated Poor — the rating surface that aggregates description count, headline count, pinning, and uniqueness.
• Fix: RSA pinning over-constrained — the pinning side of the combination-math problem.
• Fix: RSA headlines lack uniqueness — the diversity dimension; the same advice applies to descriptions.
• Quality Score — Ad Strength feeds expected CTR and ad relevance, which feed Quality Score.

Sources

1. Google Ads Help — About responsive search ads. https://support.google.com/google-ads/answer/7684791 (accessed 2026-05-27)
2. Google Ads Help — About Ad Strength. https://support.google.com/google-ads/answer/9921843 (accessed 2026-05-27)
3. Optmyzr — We Analyzed ~20,000 Google Ads Accounts. Here's What Actually Drives RSA Performance. https://www.optmyzr.com/blog/google-rsa-performance-study/ (2026-04-06)
4. Adalysis — The Responsive Search Ads Handbook. https://adalysis.com/resources-the-responsive-search-ads-handbook/ (accessed 2026-05-27)