5 Things Cogny Fixed in My Google Ads Account That No Dashboard Would Have Caught

Google Ads gives you a lot of data.

Impressions, clicks, conversions, ROAS, Quality Score, Search Impression Share, Top-of-page rate, Absolute top-of-page rate, Conversion lag, Attribution windows, Auction insights — the dashboard surface area is enormous.

And yet, the most damaging issues in Google Ads accounts are almost never the ones the dashboard surfaces on its own.

They're the ones that require correlating data across multiple views, noticing what's missing, or running analyses that require querying your ad data programmatically rather than clicking through the interface.

These are the findings Cogny's AI agent catches. Here are five specific examples — anonymized, but real — from accounts it audited.

TL;DR

Google Ads dashboards show aggregates. AI agents go looking for specific diagnoses.
The 5 most common high-impact issues Cogny finds: keyword cannibalization, wrong-landing-page routing, Quality Score blind spots, bid strategy conflicts, and budget timing mismatches.
Each finding creates a specific action ticket — not a suggestion, but a "do this, expect this result" instruction.
Connect your Google Ads account to Cogny to run the first audit.

Why Google Ads Dashboards Have a Structural Blind Spot

The Google Ads interface is designed to optimize individual campaign performance. It shows you metrics within a campaign or ad group, and it surfaces recommendations based on signals within its own ecosystem.

What it doesn't do:

Cross-reference your ad performance with your organic search data — so it can't tell you when a keyword you're paying for is also ranking organically, and you're competing with yourself
Correlate your ad traffic with post-click behavior in GA4 — the Google Ads conversion pixel can lie; GA4 shows you what actually happened after the click
Identify what's missing — dashboards show existing data; they don't flag that a converting keyword has no dedicated landing page, or that a budget caps out at 11am every day

An AI agent that connects to both Google Ads and Google Analytics 4 simultaneously — and can query both programmatically — catches things the dashboard structurally cannot.

Here's what that looks like in practice.

Finding 1: The Keyword Cannibalizing Itself

What it was: An account running search campaigns for a SaaS product had a keyword — "[product category] software" — appearing in two separate ad groups with different match types and different bids. The exact match version was bidding €2.80 CPC. The phrase match version in another campaign was bidding €1.40 CPC.

Both were targeting the same queries. Google was entering both into the same auction — and the higher-bid version was winning the impression while the lower-bid version was logging the conversion (because the click happened on a different session).

What the dashboard showed: Both ad groups appeared to be performing. One had better CPA than the other, but neither was flagged as anomalous.

What Cogny found: By querying the search term report at the query level — not the keyword level — the agent found that 34% of the clicks on the phrase match ad group were exact match queries that should have been captured by the exact match ad group. The bid competition between the two was inflating cost on the high-bid version while diluting credit attribution for the low-bid version.

The ticket: "Consolidate [keyword] targeting. Remove phrase match from Campaign B. Add negative phrase match [keyword] to Campaign B to prevent overlap. Estimated CPC reduction: 18–22%. No impression share impact expected."

Finding 2: The Landing Page That Matched the Ad But Not the User

What it was: A campaign targeting "B2B project management tool" was sending clicks to the homepage. The homepage was general — it covered all use cases, individual and team. The ad copy specifically mentioned team features and collaboration.

The gap: users who searched for "B2B project management tool" were clicking an ad that promised team features, arriving at a homepage that led with individual productivity copy, and leaving.

What the dashboard showed: The campaign had a 5.8% CTR (strong) and a 78% bounce rate (bad). Conversion rate was 0.9% against an account average of 2.4%.

What the dashboard didn't show: The bounce rate was attributed to the campaign in aggregate — not linked to the specific mismatch between the ad's message and the landing page's message. Google's own recommendations suggested "improving ad relevance," which is correct but not specific enough to act on.

What Cogny found: By pulling the ad copy, the keyword intent data, and the GA4 landing page behavior in a single analysis, the agent identified the message mismatch. It also found that an existing page — /teams — had a 2.1% conversion rate on direct traffic and directly addressed team use cases.

The ticket: "Redirect B2B project management campaign traffic from / to /teams. No copy changes needed — the landing page already exists and converts at 2.1%. Estimated conversion rate improvement: from 0.9% to 1.8–2.1%. At current traffic volume, that's approximately 12 additional trial sign-ups per month from this campaign alone."

Finding 3: Quality Score Killing Cost Efficiency in a Hidden Ad Group

What it was: Quality Score affects how much you pay per click. A Quality Score of 7 vs. 4 on the same keyword can mean paying 30–50% more per click for the same position.

Most advertisers check Quality Score at the keyword level on the main keywords they're watching. They don't systematically audit Quality Score across every ad group, especially in large accounts where long-tail ad groups accumulate over years.

What the dashboard showed: The account had a blended Quality Score that looked acceptable — around 6.2 account-wide.

What the dashboard didn't show: One ad group — added 18 months ago for a product category that was later deprioritized — had Quality Scores of 2–3 across its keywords. It was spending €340/month. Every click it was generating cost 40–60% more than the same clicks would cost in a well-managed ad group.

What Cogny found: The agent pulled Quality Score data across every active keyword in the account and flagged the outlier ad group. It then correlated the Quality Score with the landing page experience and found that the destination URL was returning a 301 redirect — a page that had been moved but whose ad destination hadn't been updated.

The ticket: "Ad group [name] — 3 keywords with Quality Score 2–3, spending €340/month. Root cause: destination URL redirects to /new-category-page (301). Update destination URLs to /new-category-page directly. Estimated Quality Score recovery: 5–7 within 30 days. Estimated CPC reduction: 35–45% on affected keywords. Estimated monthly savings at current traffic: €120–€150."

Finding 4: The Budget That Ran Out Before Lunch

What it was: A campaign with a €80/day budget was exhausting its budget by 11:30am on weekdays. After that, the account went dark for the rest of the day — including the afternoon and evening hours when the account's own GA4 data showed the highest conversion rates.

What the dashboard showed: "Budget limited" status on the campaign. Google's recommendation: increase budget to €140/day.

What the dashboard didn't show: The hourly breakdown of when conversions were actually occurring (which requires pulling data from GA4, not just Google Ads), and whether the morning impression share was actually converting or just generating clicks.

What Cogny found: By pulling hourly conversion data from GA4 and cross-referencing it with the Google Ads impression share data, the agent found:

68% of conversions occurred between 1pm and 8pm
The campaign was generating 0 impressions between 11:30am and midnight (budget exhausted)
The morning traffic (8am–11:30am) had a conversion rate of 0.8% — half the account average
The afternoon/evening traffic that the campaign was missing had an estimated conversion rate of 2.4%

The recommendation wasn't "increase budget." It was to use ad scheduling to suppress morning bidding and concentrate the same €80/day budget in the high-conversion window.

The ticket: "Add ad schedule: reduce bids by 40% from 6am–12pm, maintain bids 12pm–9pm, pause 9pm–6am. Estimated effect: campaign runs until 6–7pm at current budget. Conversion volume estimate: +35–45% with no budget increase. Run for 14 days before evaluating."

Finding 5: The Competitor Campaign Targeting the Wrong Competitor

What it was: The account was running a competitor keyword campaign targeting the brand names of three competitors. The goal was to capture comparison-shopping traffic from users evaluating alternatives.

One of the three competitors — a platform that had been acquired and rebranded 8 months ago — was still being targeted under its old name. Search volume for the old name had dropped 85% since the rebrand. The campaign was generating 2–3 clicks per month on those keywords.

Meanwhile, the new brand name — which had accumulated significant search volume as the acquirer promoted the rebrand — was not being targeted at all.

What the dashboard showed: Low impression share on the old-brand keywords. Low click volume. The campaign looked like it just wasn't competitive.

What the dashboard didn't show: That the search demand had migrated to a different keyword that wasn't in the account at all. The dashboard can't show you what's missing from your account.

What Cogny found: The agent cross-referenced the targeted competitor keywords against search volume trends from Search Console and external keyword data. It identified the brand migration and flagged the missed opportunity.

The ticket: "Competitor targeting gap: [Old Brand Name] has dropped from ~4,200 monthly searches to ~630 (85% decline, due to rebrand to [New Brand Name]). [New Brand Name] currently has ~3,800 monthly searches — not targeted in your account. Add [New Brand Name] as competitor keyword in Campaign [X]. Estimated incremental monthly clicks at 3% CTR: 114. At account average conversion rate of 2.2%: 2–3 additional trials/month from competitor traffic."

What Makes These Findings Different From Dashboard Recommendations

Google's automated recommendations are useful. "Add responsive search ad variations" and "expand to additional keywords" are real suggestions that improve accounts.

But notice the pattern of the five findings above:

Finding 1 required comparing two ad groups simultaneously and querying the search term report at query level — not keyword level
Finding 2 required correlating ad copy, keyword intent, and GA4 landing page behavior in one analysis
Finding 3 required auditing every keyword's Quality Score and cross-referencing destination URL status
Finding 4 required pulling hourly conversion data from GA4 and hourly impression data from Google Ads and comparing them
Finding 5 required checking whether targeted keywords still had search volume, which required external search data

None of these are things you can find by clicking around the Google Ads interface. They all require:

Access to multiple data sources simultaneously
The ability to query at the right level of granularity
The ability to notice what's absent, not just what's present

This is what a connected AI agent does that dashboards can't.

Frequently Asked Questions

How long does a Cogny Google Ads audit take?

The agent runs automatically. The first full audit — covering all campaigns, ad groups, and keywords in your account, cross-referenced against GA4 — typically completes within 15–20 minutes. You get a set of prioritized tickets, not a report to read.

Does Cogny need access to Google Ads and GA4 separately?

Yes. Cogny connects to Google Ads via the Google Ads API and to Google Analytics via the GA4 MCP connector. The cross-platform findings (like the budget timing analysis) require both. You can connect one without the other, but the highest-value findings come from the cross-platform correlation.

Will Cogny make changes to my account automatically?

No. Cogny creates tickets and recommendations. It doesn't execute changes in your ad account without your approval. Each ticket includes the specific change to make — you execute it (or your team does). This is by design: ad spend decisions should have a human sign-off.

What size accounts does Cogny work with?

Accounts spending from €500/month to €200,000+/month. The value compounds at higher spend levels because the cost of undetected inefficiency is proportionally larger, but the findings pattern above appears at every spend level.

How often does the agent run?

By default, weekly. You can also trigger an on-demand audit at any time. Weekly audits catch issues before they compound; the budget timing issue (Finding 4) was burning approximately €1,200/month unnecessarily before the agent flagged it — a weekly audit catches that within 7 days maximum.

Getting Started

Connect your Google Ads account to Cogny and run the first audit. The findings above are representative of what comes out of the first analysis for most accounts.

The audit is automated. The tickets are specific. The expected outcomes are quantified.

Connect your Google Ads account at cogny.com/integrations/google-ads

Related reading: