{"id":1038,"date":"2026-09-07T08:44:30","date_gmt":"2026-09-07T12:44:30","guid":{"rendered":"https:\/\/thedoctorads.com\/blog\/google-ads-one-responsive-search-ad-per-group"},"modified":"2026-09-07T08:44:30","modified_gmt":"2026-09-07T12:44:30","slug":"google-ads-one-responsive-search-ad-per-group","status":"publish","type":"post","link":"https:\/\/thedoctorads.com\/blog\/google-ads-one-responsive-search-ad-per-group","title":{"rendered":"Four in Five Ad Groups Run a Single Responsive Search Ad"},"content":{"rendered":"<div class=\"tldr\">\n<p><strong>TL;DR abstract.<\/strong> Google Ads Help asks advertisers to &#8220;implement at least 2 responsive search ads with Good or Excellent Ad Strength per ad group&#8221;, and publishes its own uplift figures for doing so: a second responsive search ad is associated with <strong>6.6%<\/strong> more conversions on average, and a third with a further <strong>3.7%<\/strong>. That is a rare thing in this industry, a specific recommendation with a number attached to it. It is also contested: in December 2025 Optmyzr argued the opposite, that one responsive search ad per ad group concentrates the learning. Nobody appears to have measured which side accounts are actually on.<\/p>\n<p>We counted the enabled responsive search ads in every eligible ad group of a managed portfolio: <strong>100,634<\/strong> enabled ad groups inside Search campaigns across <strong>107<\/strong> accounts, each holding at least one. <strong>78.9% of them hold exactly one.<\/strong> Only <strong>0.9%<\/strong> hold three or more. In the median account the single-ad share is <strong>96.8%<\/strong>, with the middle half of accounts between 78.5% and 100%. The obvious objection is that single-ad groups are dormant leftovers, and it does not survive contact with the spend data: among the 1,966 ad groups that actually spent in the first half of 2026, the single-ad share is <strong>higher<\/strong> at 79.2%, and those ad groups carry <strong>94.5%<\/strong> of the money, <strong>88.3%<\/strong> once the largest account is set aside.<\/p>\n<\/div>\n<aside class=\"key-stat\">\n<p><strong>Among 100,634 enabled Google Ads Search ad groups that hold at least one responsive search ad, 78.9% hold exactly one, against a Google recommendation of at least two. Among ad groups that spent in 2026, the figure is 79.2% and they hold 94.5% of the spend, 88.3% without the largest account.<\/strong><\/p>\n<p class=\"source-line\">Method basis: enabled ads in enabled ad groups inside Search campaigns, 107 accounts under management, snapshot 10 July 2026. Source: Doctor Ads Profit Forensics.<\/p>\n<\/aside>\n<p>Most platform advice is unfalsifiable by design. Add more assets, write compelling copy, test regularly. The responsive search ad recommendation is not like that. Google names the quantity, names the quality bar, and publishes the effect size it claims: <a href=\"https:\/\/support.google.com\/google-ads\/answer\/7684791\" target=\"_blank\" rel=\"noopener\">at least two responsive search ads per ad group<\/a>, each with Good or Excellent Ad Strength and its own final URL, with <strong>6.6%<\/strong> more conversions on average when a second is added and a further <strong>3.7%<\/strong> when a third is added, on Google&#8217;s internal data from August 2025.<\/p>\n<p>A claim like that invites a simple, uncomfortable question. Do accounts do it?<\/p>\n<p>It is also not a settled question, and this is the part that changes how the whole page should be read. In December 2025 Optmyzr published the opposite advice under the title <a href=\"https:\/\/www.optmyzr.com\/blog\/one-rsa-per-ad-group-google-ads-strategy\/\" target=\"_blank\" rel=\"noopener\">Why One RSA Per Ad Group is the Smart Play<\/a>: &#8220;stick to one RSA per ad group&#8221;, because &#8220;when all performance is consolidated in one place, the algorithm learns faster. If you spread it across multiple RSAs, each ad gets too little data to work with.&#8221; So the platform asks for at least two and a major tooling vendor argues for exactly one. Whatever the accounts are doing, they are not simply ignoring advice; they are sitting on one side of a live disagreement.<\/p>\n<p>Neither side has published the denominator. Optmyzr has run two large public studies in the neighbourhood, one on <a href=\"https:\/\/www.optmyzr.com\/blog\/optmyzr-study-responsive-search-ad-performance\/\" target=\"_blank\" rel=\"noopener\">responsive search ad performance across 13,671 accounts<\/a> and a later one on <a href=\"https:\/\/www.optmyzr.com\/blog\/google-ad-strength-study\/\" target=\"_blank\" rel=\"noopener\">Ad Strength across more than a million ads in over 22,000 accounts<\/a>. Both look at how these ads perform and what Ad Strength is worth. Neither publishes the flat count of how many sit in an ad group, which is the one number both recommendations are about.<\/p>\n<h2>Nearly four in five ad groups run a single ad<\/h2>\n<p>We took every ad group inside a Search campaign that was itself enabled and already held at least one enabled responsive search ad. That is <strong>100,634 ad groups across 107 accounts<\/strong>, and the denominator matters: it excludes the 40,785 enabled Search ad groups that hold none at all, and it counts ads by inference rather than by label, because this export has no ad-type field and we recognise a responsive search ad by its headline assets. Counting those ads in each qualifying ad group:<\/p>\n<p><strong>78.9% hold exactly one. 0.9% hold three or more.<\/strong><\/p>\n<p>The account level view is starker than the pool. Across the 76 accounts with at least ten qualifying ad groups, the median account runs a single ad in <strong>96.8%<\/strong> of them, with the middle half of accounts between <strong>78.5% and 100%<\/strong>. Leave one account out and recompute 76 times, once per qualifying account, and the median moves between 96.55% and 97.14%. This is not one messy client.<\/p>\n<p>Loosen the denominator to every ad group in the snapshot holding an enabled responsive search ad, regardless of ad group status or campaign type, and the picture barely moves: <strong>138,862 ad groups in 110 accounts, 77.0% with one, 1.4% with three or more<\/strong>. The finding does not depend on which population you pick, which is usually the first thing worth checking about a finding like this.<\/p>\n<p>One number in that wider cut is worth flagging against ourselves, because a knowledgeable reader will spot it. Google&#8217;s <a href=\"https:\/\/support.google.com\/google-ads\/answer\/6372658\" target=\"_blank\" rel=\"noopener\">account limits<\/a> cap an ad group at three enabled responsive search ads, and yet 63 ad groups here show four or more. That is not the platform breaking its own rule. It is our identification method showing its edge: this export has no ad-type field, so we recognise a responsive search ad by the presence of responsive headline assets, and a handful of ad groups carry other responsive formats that look the same from that angle. 63 out of 138,862 does not move any figure on this page. It is also only the visible class of error: a misclassified ad inside a group counted as holding one, two or three is undetectable by construction, so 0.05% is the rate of impossible counts we can see, not an estimate of the overall misclassification rate. It is the reason the three or more tier should be read as &#8220;at the platform maximum&#8221; rather than as a literal count of responsive search ads.<\/p>\n<figure class=\"chart\"><img decoding=\"async\" src=\"https:\/\/thedoctorads.com\/blog\/wp-content\/uploads\/2026\/09\/tda-w8-rsa-distribution.png\" alt=\"Across 138,862 Google Ads ad groups holding at least one enabled responsive search ad, 106,882 hold exactly one, 30,068 hold two, 1,849 hold three and 63 hold four or more.\" title=\"\"><figcaption><strong>106,882 ad groups run one responsive search ad, 30,068 run two, 1,849 run three.<\/strong> On the clean cut of enabled ad groups inside Search campaigns, 78.9% run exactly one against a Google recommendation of at least two. Source: Doctor Ads Profit Forensics.<\/figcaption><\/figure>\n<p>We are calling this <strong>the Lone RSA<\/strong>, and the number to hold on to is not 78.9%. It is 0.9%: the share of ad groups that reach three, which is both the tier Google attaches its second uplift figure to and the most the platform allows. Fewer than one ad group in a hundred is there.<\/p>\n<h2>Dormancy alone does not explain it<\/h2>\n<p>The honest counter argument is that most of those ad groups are idle. They are enabled at the snapshot, which is what the denominator requires, but an ad group that has not spent in years will still show its one enabled ad, because nobody was ever going to add a second to something that is not running. If the single-ad groups are all like that, the finding is an archive artefact and not worth your time.<\/p>\n<p>So we restricted to ad groups that recorded spend in the first half of 2026, keeping the same population: <strong>1,966 ad groups across 63 accounts, carrying $1.88M<\/strong>. If the objection held, the single-ad share should drop.<\/p>\n<p>It rises. <strong>79.2% of ad groups that actually spent money run exactly one responsive search ad<\/strong>, and in the median of the 26 accounts with at least ten spending ad groups the figure is <strong>99.3%<\/strong>, with a leave-one-out range of 98.6% to 100%. That rules out a purely archival explanation. It does not tell you these ad groups were still active on the day of the snapshot, and it does not explain why they hold one ad.<\/p>\n<p>Weight by money instead of by ad group and it goes further: <strong>94.5% of the $1.88M ran through ad groups holding a single ad<\/strong>. Because one account carries <strong>53.3%<\/strong> of that spend, this figure needs the same whale test we ran on the network settings piece, and it passes: without that account the share is <strong>88.3%<\/strong> of $877,643 across 62 accounts, and the median of the 26 accounts with at least ten spending ad groups is <strong>99.7%<\/strong> (middle half 65.2% to 100%, leave-one-out 99.47% to 100%). The pattern is not concentrated in the dead part of the account, and it is not one client.<\/p>\n<figure class=\"chart\"><img decoding=\"async\" src=\"https:\/\/thedoctorads.com\/blog\/wp-content\/uploads\/2026\/09\/tda-w8-rsa-spend.png\" alt=\"Horizontal bar chart with three bars: all enabled Search ad groups at 78.9%, ad groups that spent in the first half of 2026 at 79.2%, and the share of that half year spend sitting in single-ad groups at 94.5%. The two spend related bars are the longer ones.\" title=\"\"><figcaption><strong>Restricting to ad groups that actually spent raises the single-ad share to 79.2%, and those ad groups hold 94.5% of the spend, 88.3% with the largest account removed.<\/strong> Idleness does not explain the pattern: it is at least as strong where the money is. Source: Doctor Ads Profit Forensics.<\/figcaption><\/figure>\n<p>One more number for scale, and it belongs beside the others rather than inside them. Of the <strong>141,419<\/strong> enabled ad groups sitting in Search campaigns, <strong>40,785, or 28.8%, hold no enabled responsive search ad at all<\/strong>. Some of those are legitimately structured that way, with dynamic or call-only ads doing the work. Others hold no live text ad at all. This count measures the absence of an enabled responsive search ad and nothing about why it is absent.<\/p>\n<h2>What this does and does not tell you<\/h2>\n<p>It does not tell you that adding a second ad would lift these accounts by 6.6%. That figure is Google&#8217;s, measured on Google&#8217;s data, across a population we cannot see, and we are not replicating it. Nor does it say the portfolio is wrong: with a credible vendor arguing for exactly one ad, 78.9% is observed practice inside a disagreement, not a compliance failure.<\/p>\n<p>It also cannot be turned into a within-account test here. Ad groups with a second ad were not assigned one at random, and they may differ systematically from the rest in ways this snapshot does not record, which is the selection effect that makes naive comparisons look flattering.<\/p>\n<h2>When this does not apply<\/h2>\n<p><strong>Your ad groups are single-keyword by design.<\/strong> In a tightly themed structure with one keyword and one landing page, a second ad has less room to differ, and the argument for it is weaker than in a broad ad group.<\/p>\n<p><strong>You have low volume.<\/strong> Two ads split the same impressions, and in an ad group getting a few hundred a month neither accumulates enough data for the system to learn a preference. This is the same argument Optmyzr makes at portfolio scale.<\/p>\n<p><strong>Your second ad would be a near-copy.<\/strong> Two responsive search ads with overlapping headlines do not give the auction more to work with, they give it the same thing twice. If the second ad has no distinct angle or landing page, its absence is a decision rather than an omission.<\/p>\n<h2>How to check yours in about two minutes<\/h2>\n<ol>\n<li>Open the Ads view for a campaign and group by ad group. Look at the count of enabled ads per group rather than at the ads themselves.<\/li>\n<li>Sort by cost, not by name, and read the ad count on the groups at the top. In this portfolio 94.5% of the half year spend sat in single-ad groups, so that is where the audit pays.<\/li>\n<li>Where you add a second, give it a genuinely different angle or landing page. Google&#8217;s own wording asks for a unique final URL for each responsive search ad.<\/li>\n<li>Note the date you added it. Without that you cannot read the before and after later, and you will be back to trusting a platform average.<\/li>\n<\/ol>\n<h2>Key takeaways<\/h2>\n<ul>\n<li>Google asks for <strong>at least two<\/strong> responsive search ads per ad group and publishes its own uplift figures, <strong>6.6%<\/strong> and a further <strong>3.7%<\/strong>. Optmyzr publicly argues for exactly one.<\/li>\n<li><strong>78.9%<\/strong> of the 100,634 enabled Search ad groups that hold at least one such ad run exactly one. Only <strong>0.9%<\/strong> run three or more.<\/li>\n<li>Median account: a single ad in <strong>96.8%<\/strong> of its ad groups, leave one account out <strong>96.55% to 97.14%<\/strong>.<\/li>\n<li>Dormancy does not explain it: among ad groups that spent in 2026 the share is <strong>79.2%<\/strong>, and they carry <strong>94.5%<\/strong> of the spend, <strong>88.3%<\/strong> with the largest account removed.<\/li>\n<li>A further <strong>28.8%<\/strong> of enabled Search ad groups hold no enabled responsive search ad at all.<\/li>\n<\/ul>\n<p><!-- cta-anchor --><\/p>\n<h2>Researcher&#8217;s take<\/h2>\n<p>The gap between 78.9% and Google&#8217;s recommendation is not the interesting part, because the recommendation is contested and the accounts may simply be on Optmyzr&#8217;s side of the argument. The interesting part is 0.9%. The third ad, the tier carrying Google&#8217;s smaller uplift claim and the platform&#8217;s own ceiling, is almost absent from a portfolio of a hundred-plus professionally managed accounts. Two stories fit that and this data separates neither: a deliberate consolidation strategy, or a second ad that never got written because it is real work with an uncertain payoff. My own bet, and it is a bet rather than a finding, is that both are present and the second is more common. What I would actually do is narrower than either camp: pick the handful of ad groups carrying most of your spend, add one genuinely different ad there, and write down the date so the comparison is readable later.<\/p>\n<p><em>Igor Ivitskiy, Doctor Ads<\/em><\/p>\n<h2>Method<\/h2>\n<p>Data statement v.2026.09. Source: ads, ad groups, campaigns and ad group performance for a managed portfolio, one snapshot dated 10 July 2026. Unit of analysis is one ad group. Lead population is enabled ad groups inside campaigns of type Search holding at least one enabled responsive search ad: 100,634 ad groups in 107 accounts. Accounts are pseudonymised, and no ad text, headline, final URL, ad group, campaign or vertical name is published.<\/p>\n<p>Definition. Responsive search ads are identified by the presence of responsive headline assets on the ad, because the ad-type field is empty across all 911,619 ad rows in this export. The one visible symptom of that inference is the 63 ad groups showing four or more, above Google&#8217;s cap of three: an observed impossible-count rate of 0.05% of ad groups. Misclassification inside groups counted as one to three cannot be detected this way, so that figure is a floor on the error, not a measure of it. An ad counts when its status is enabled at the snapshot. Spend weighting uses ad group cost for the first half of 2026, normalised to USD, on the same population as every other figure here: enabled ad groups inside Search campaigns holding at least one enabled responsive search ad.<\/p>\n<p>Weighting. Figures are reported both pooled across ad groups and as the median across accounts with at least ten qualifying ad groups, with leave-one-account-out ranges on both the full population and the spending subset. The spending subset has only 26 accounts meeting the ten ad group threshold, and that thinness is why its median of 99.3% is quoted with its account count rather than on its own.<\/p>\n<p>Counter hypothesis. The proposed alternative was that single-ad groups are dormant. It was tested by restricting to ad groups with spend in the first half of 2026, holding the population constant, and did not hold: the share rises to 79.2% and those ad groups carry 94.5% of the spend, 88.3% without the largest account.<\/p>\n<p>Selection note. Agency managed portfolio, not a random sample of advertisers, and skewed toward accounts large enough to be managed by an agency at all.<\/p>\n<p>Limitations. Four. Status is a snapshot, so an ad paused the day before the pull counts as not enabled, and the wide cut is reported alongside the clean one for that reason. Ad type is inferred from assets rather than read from a field. No performance comparison between single and multi-ad groups is made or implied. And Google&#8217;s 6.6% and 3.7% figures are its own, reported here as the benchmark being measured against, not as a result of ours.<\/p>\n<p>Changelog. First published September 2026. Recomputed quarterly against the live claim registry.<\/p>\n<h2>What to read next<\/h2>\n<ul>\n<li><a href=\"https:\/\/thedoctorads.com\/blog\/google-ads-ad-strength-same-ad-group\">Ad Strength inside the ad group<\/a>, the paired test on what the label is worth.<\/li>\n<li><a href=\"https:\/\/thedoctorads.com\/blog\/google-ads-search-partners-who-opted-out\">Who actually took the search partners opt out<\/a>, the other default nobody revisits.<\/li>\n<li><a href=\"https:\/\/thedoctorads.com\/blog\/google-ads-quality-scores-tested\">Which Google Ads metric actually matters<\/a>, on the platform&#8217;s own scores against their own promises.<\/li>\n<li>The opposing case, stated by a tool vendor rather than by us: <a href=\"https:\/\/www.optmyzr.com\/blog\/one-rsa-per-ad-group-google-ads-strategy\/\" target=\"_blank\" rel=\"noopener\">why one responsive search ad per ad group may be the smart play<\/a>.<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Google recommends at least two responsive search ads per ad group; Optmyzr argues for exactly one. Across 100,634 enabled ad groups in 107 accounts, 78.9% run one and 0.9% run three or more.<\/p>\n","protected":false},"author":0,"featured_media":1037,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[18],"tags":[],"class_list":["post-1038","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-account-management","generate-columns","tablet-grid-50","mobile-grid-100","grid-parent","grid-33"],"_links":{"self":[{"href":"https:\/\/thedoctorads.com\/blog\/wp-json\/wp\/v2\/posts\/1038","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/thedoctorads.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/thedoctorads.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/thedoctorads.com\/blog\/wp-json\/wp\/v2\/comments?post=1038"}],"version-history":[{"count":0,"href":"https:\/\/thedoctorads.com\/blog\/wp-json\/wp\/v2\/posts\/1038\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/thedoctorads.com\/blog\/wp-json\/wp\/v2\/media\/1037"}],"wp:attachment":[{"href":"https:\/\/thedoctorads.com\/blog\/wp-json\/wp\/v2\/media?parent=1038"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/thedoctorads.com\/blog\/wp-json\/wp\/v2\/categories?post=1038"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/thedoctorads.com\/blog\/wp-json\/wp\/v2\/tags?post=1038"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}