Notice Me Senpai
  • Home
  • About
pillar-content

The 30-Minute Google Ads Audit That Beats Most Agency Reports

NMS Editorial Team

NMS Editorial Team

27 Apr 2026, 10:06 AM 6 min
The 30-Minute Google Ads Audit That Beats Most Agency Reports
Six checks. Thirty minutes. Most of what is quietly draining a Google Ads account fits in one comic panel.

A Google Ads audit checklist with 100 items is mostly a stalling tactic. The fastest 80% of an account audit takes 30 minutes and looks at six things: conversion source, network opt-ins, the search terms report, Smart Bidding signals, Performance Max channel split, and asset group thinness. Everything else can wait until those six are clean.

I've watched agency audits run six weeks and produce a 47-slide deck. Most of those slides are filler. The first six checks below are the ones that actually move accounts. A monthly audit cadence improves performance roughly 23% over quarterly, according to data cited in Digital Applied's 2026 audit guide, but only if the audit is the version that fixes things, not the version that documents them.

Pull up the account. Set a timer.

Minute 0–5: Confirm conversions aren't being counted twice

The single most damaging mistake in a Google Ads account is double-counting conversions. It happens when the same purchase fires once from a GA4 import and once from a Google Ads tag, so every conversion looks like two. Smart Bidding then optimizes toward inflated numbers, and the account "performs" beautifully on a dashboard that's lying to you. GROAS's 2026 setup guide calls this one of the five most common tracking failures.

Tools and Settings, then Conversions. Sort by "All conversions." Look for the same event imported from two different sources. If you see "Purchase (GA4)" and "Purchase (Google Ads)" both marked Primary, one of them is shadow-doubling every Smart Bidding signal in the account.

While you're there, check that Enhanced Conversions is on. From what I've seen, this is the single highest-leverage tracking change in 2026. Most accounts that turn it on see somewhere in the 5 to 15% range of recovered conversions, depending on how leaky their first-party setup was to start.

If both fixes are needed, do them now. The rest of the audit assumes Smart Bidding is reading clean signal. We've covered the broader version of this problem in our piece on Smart Bidding's signal problem.

Minute 5–10: Turn off whatever Google turned on for you

Google's defaults aren't malicious. They're just expensive. Two settings come pre-flipped in most accounts and quietly burn 15 to 30% of budget on traffic that converts at half the rate of regular Search.

The first is Search Partners. New Search campaigns ship with Search Partners on. WordStream's analysis put Search Partners as one of the most common sources of silent waste in audited accounts. Segment your Search campaigns by Network. If Search Partners CPA is more than 1.3x your Google.com CPA, turn it off. If you can't tell because volume is too low to be statistically meaningful, also turn it off. Google will not penalize you for it.

The second is the Display Network opt-in on Search campaigns. This was the legacy default for years and still appears in older campaigns. Search Engine Land confirmed Google finally stopped auto-opting new Search campaigns into Display, but anything older than a couple of years probably still has it on. The performance is, in roughly 99 out of 100 audits, terrible. Turn it off.

The fix is two clicks per campaign. The cumulative budget impact is usually larger than anything else on this list.

Minute 10–15: Pull a 90-day search terms report and look at one column

Open a Search campaign. Insights and reports, then Search terms. Date range: last 90 days. Sort by Cost descending. Filter for Conversions = 0.

This is the single most useful five minutes you can spend in a Google Ads account. Optmyzr's data suggests most accounts have under 200 negative keywords when they probably need closer to 1,000+ for real filtering. The search terms report is where you find the missing 800.

Anything in that filtered list with more than $50 in spend and zero conversions is a candidate for a negative. Add the obvious ones immediately. The rest, copy into a sheet and sort by intent: Tier 1 is high-intent and converting (keep), Tier 4 is irrelevant or junk traffic (negative). The Tier 4 column is your homework. ALM Corp's 2026 search audit framework uses a similar four-tier classification and it's the cleanest version I've seen.

One note: don't dump unstructured negatives into one giant list. Document the rationale next to each. The "why" is what stops a future you (or a freelancer) from undoing the cleanup six months later.

Minute 15–20: Audit the Smart Bidding signal

Smart Bidding only works if it's optimizing toward something that matters. If your account is feeding it "Form submission" but half those forms are bots or low-quality leads, the algorithm will faithfully bring you more bots.

Tools and Settings, then Conversions. Click into your Primary conversion. Check three things: the count window, the action category, and the value. If the value is $0 or "Don't use a value" on a sales account, Smart Bidding is treating every conversion as equal even though some are worth $5 and some are worth $5,000. From what I've seen, that's the biggest lever in most accounts.

If you're running tROAS or Maximize Conversion Value, check that the conversion value is dynamic (passed from the cart) and not a flat estimate. A flat value will train Smart Bidding to chase low-AOV customers because they look identical to high-AOV ones in the model.

Google rolled call-recording-as-default into the lead gen flow this year, with AI deciding what counts as a qualified call. (We covered why that's a problem when it shipped.) If you import qualified calls as conversions, audit the qualification logic. The bar between "qualified" and "rang for 60 seconds" is doing a lot of work in your bidding right now.

Minute 20–25: Open one Performance Max campaign and check three reports

Performance Max stopped being a black box in early 2026, when Google rolled out asset-group-level reporting and a channel performance timeline. Now you can actually see which channel (Search, YouTube, Display, Discover, Gmail, Maps) is driving each result. Yellowjack's writeup has the clearest screenshots if you've never opened the new view.

Three reports, two minutes each:

Asset group performance. PMax now shows ROAS, CTR, and CPC per asset group. Stormy AI's 2026 audit framework recommends using these breakdowns to find the asset groups doing real work. If one group is producing 80% of the volume, your other groups are theater. Either fix them or fold them in.

Channel timeline. If 70% of your PMax spend is going to Display or Gmail and your CPA there is 4x your Search CPA, you're effectively running an underpriced display campaign with extra steps. Google's own documentation confirms PMax now exposes this split per asset group.

Search Partner placements in PMax. Q1 2026 added full placement transparency for the Search Partner network inside PMax. ALM Corp's analysis of the February rollout walks through how to pull the placements report. Add anything that looks like a parked domain or a clearly low-quality publisher to your account-level placement exclusion list.

Minute 25–30: Check asset group thinness

Last five minutes. Open the asset groups in your top PMax campaign. The 2026 quality bar, per Stormy AI's audit standards, is roughly 15 high-quality images, 5 diverse videos, and the full slate of 5 headlines and 4 descriptions per group. Anything below that is what Google quietly downranks as "thin assets."

Most accounts I've seen are missing on video. Asset groups with one stitched-together stock video aren't going to survive against competitors who shipped three real ones. The fix isn't a $20K production. It's three 6-second cuts of an existing creator-style video, repurposed across vertical, square, and horizontal. From what I've seen, that alone moves PMax serve rate noticeably within a week or two.

What this audit is and isn't

A 30-minute audit can't fix structural problems. If your account has 40 campaigns where 10 would do, or your Smart Bidding strategies are mismatched to your conversion volume, no fast pass will catch it. Those are the real audit conversations, and they take a day, not half an hour.

But the six checks above will catch most of what's actively bleeding budget right now. The conversion-double-count alone has wrecked more accounts than any algorithm change. Search Partners and the Display opt-in are still on in a surprising number of mature accounts. The negatives gap is real. The PMax channel split was, until February, literally invisible.

Most agency audits I've seen open with a 47-slide deck and close with a vague "we recommend a strategic refresh," which is a sales motion in audit drag. The version that helps is the one that ends with six specific things that got fixed by the time the timer ran out.

If you want the longer playbook on running these accounts without letting Google quietly take the wheel, the pillar piece is here.

The teams I've watched do this monthly drift less. The ones who treat it as an annual deep-clean spend the rest of the year explaining performance dips that started 11 months ago.

Notice Me Senpai Editorial

Read next

Propaganda-poster illustration of a brand voice framework handed from founder to writing team, with exclusion-list clipboard, voice waveform, and grading rubric.

Brand Voice Development: A Framework for Teams, Not Just Founders

A brand voice framework only works if it survives the handoff from founder to team. Most fail because they describe how the founder sounds rather than giving writers a repeatable way to produce that sound. The fix is operational, not philosophical: rules, examples, and exclusion lists, graded on every piece
NMS Editorial Team 20 May 2026, 10:05 AM
Pop-art illustration of a locked vault door stamped June 15 with a developer token key dangling beside it and a torn calendar strip reading Dec 2025 to May 2026.

Google's Conversion API Cutoff Reads Six Months of Your Token Logs

Google's Ads Developer Blog confirmed on May 15 that the Google Ads API will block new offline conversion imports starting June 15, 2026. The cutoff applies to any developer token that did not call ConversionUploadService.UploadClickConversions between December 2025 and May 2026. Tokens outside that six-month usage window
NMS Editorial Team 19 May 2026, 9:25 AM
Propaganda-poster collage: a Lando Lane street sign, stacked newspaper front pages, a 30-day calendar with four red pitch-wave markers, and a Reddit thread on a CRT screen.

Earned Media Arbitrage: How to Turn One Stunt Into a Month of Coverage

Earned media arbitrage is the practice of treating one PR stunt as the raw material for 30 days of follow-on coverage, not a 72-hour spike. Most brands capture the initial wave of articles, then watch the story die by Friday. Teams that plan the second and third waves on day
NMS Editorial Team 18 May 2026, 10:05 AM
  • Sign up
Notice Me Senpai © 2026.