Members only · Google Ads audit

Read the accountbefore anythingin it is changed

One report, read from your own account: where it stands, what is wrong and how it is fixed, the first three months, and the evidence behind every line. Every figure comes from the data and carries its source.

$49/mo & up · members only · or $250 each

What happens next
01

What you get

One report, four parts

Everything the audit needs from you is asked before it starts: your target cost per customer or return on ad spend, what counts as a sale, what a customer is worth and over what payback period, the funnel after the click, who else changes the account, seasonality, and any compliance limits for your category. Nothing is asked mid-audit and the report never asks you a question. What the account cannot show is handled with common sense and written in as a plain assumption you can correct.

  1. 1

    Where the account stands

    The account read once, in plain English, before a single recommendation. Cost per customer against your target, the spend trend, whether the season is or is not the cause, where the money actually goes, and what is already working. The headline numbers come first, then the detail behind them.

    • Cost per customer against your target, month by month
    • Whether the trend is seasonality, spend, or the account itself
    • Where the money goes, and which campaigns are already earning
  2. 2

    What is wrong and how we fix it

    Every major flaw, in the order we would fix it. Each one carries the problem with its numbers, what we will do about it, why that works, and how we will know it did. Week-one quick wins come first, then the biggest changes from most to least impact, then the lower-impact and longer-term work. It closes with the strategy for scaling, the options for new audiences, and what the plan cannot promise.

    • Quick wins for week one, then changes ordered by impact
    • Each fix with its numbers, its risks and the test that proves it
    • Scaling rules, new audience options, and what we cannot promise
  3. 3

    The first three months

    Week by week and month by month. Every necessary change is made in month one, as many as possible in the first week, with fixed rules for when we hold, scale, cut or pivot agreed before any change is made. It also sets out how often the account is watched and what each month should look like.

    • Everything that must change, in month one and mostly in week one
    • Fixed rules for holding, scaling, cutting or pivoting
    • How often the account is watched, and what to expect each month
  4. 4

    The account breakdown

    All the data and evidence behind every conclusion, for whoever wants to check the working. The campaign map and its cards, three-month trends, seasonality, creative, settings, landing pages, tracking, change history and the customer-cost data, with each finding stated alongside the comparison it rests on.

    • The campaign map: a card for every live or recently spending campaign
    • Trends, seasonality, creative, settings, landing pages and tracking
    • Evidence tables with an interval on every comparison
Sample · Report page one
SEO KarbonGoogle Ads Audit · Report

Google Ads audit · Report

Audit Report

Sample account · Google AdsAll figures in GBP, the account's own currency

Cost per customer is running 23% above target, and almost all of the gap sits in two campaign groups that have never been separated.

Contents

What This Report Holds

1Where the account stands
2What is wrong and how we fix it
3The first three months
4The account breakdown

Headline numbers

Cost per customer

£184

Against a target of £150

Return on ad spend

3.4x

Against a target of 4.0x

Spend, 90 days

£86,400

Google Ads, read from the account

Customers

640

From the store's own order data

Cost per customer by month

Goal £150
£158
£163
£171
£196
£203
£184
AprMayJunJulAugSep

Spend by month, split by campaign type

AprMayJunJulAugSep
  • Brand
  • Non-brand search
  • Performance Max

Where these numbers come from

Every figure on this page is read from the account, in its own currency and time zone, and carries its source. What was not measured is said so rather than shown as a zero.

The four parts that follow explain the page in this order, ending with the evidence behind every line.

Page one of the report, drawn the way it prints.

Three things the report never does

The account is read once, carefully, and the report has to survive being checked by whoever owns the numbers.

  • It never types a number by hand

    Every figure is read from the account, carries its source, and is shown in the account's own currency and time zone, with any exchange rate and its date beside it.

  • It never scores what was not measured

    If something could not be measured, the report says so plainly rather than scoring it zero or quietly leaving it out.

  • It never asks you a question

    Whatever the data cannot show was asked before the audit started, then written in as a plain assumption, so the report is read rather than answered.

02

What we need

Four read-only accesses

All read-only, all approved by you in the platform itself. The audit never changes the account, and any of it can be revoked afterwards.

  • Google Ads

    The account being audited: campaigns, budgets, conversions, change history and spend.

  • Google Analytics 4

    The property linked to the account, for on-site behaviour and to check the tracking.

  • Google Tag Manager

    The container behind the conversions: wrong IDs, duplicate tags, dead triggers, consent.

  • Google Merchant Center

    Ecommerce accounts only. Feed health and disapprovals. Not asked for otherwise.

03

How it stays honest

Six rules that keep it defensible

These are applied to every number and every recommendation in the report.

  1. 01

    Platform questions get that platform's data

    A question about Google Ads is answered from Google Ads. GA4 is used for on-site behaviour and to check the tracking, never to judge the channel, and the attribution basis is stated next to every number.

  2. 02

    Costs are judged at the deepest stage with enough volume

    Customers are usually too few to compare campaigns, so the judgement sits at the steady, plentiful stage just before the sale, and the report says which stage it is.

  3. 03

    Every comparison gets a 95% interval

    Two groups are only called different if their intervals do not overlap. Otherwise we say they cannot be ranked, and we do not rank them.

  4. 04

    The real outcome gets the last word

    A cheap lead can hide a rising real cost per customer. As soon as order or CRM data exists it is loaded, every earlier claim is rechecked against it, and if something changes the report says what changed.

  5. 05

    Marginal cost, not just average

    Outcomes are regressed on spend to estimate what the next unit of outcome costs, with a range, because an average hides the point where growth stops paying.

  6. 06

    Nothing is recommended that the data cannot carry

    The prize and the risk are both sized, and what the plan cannot promise is written down rather than left out.

04

What it checks

Extra checks on an ecommerce account

A store gets three more passes, because its conversions are only as good as the tags and the feed behind them.

Google Tag Manager

Wrong IDs, duplicate pixels, dead triggers and missing conversion tags.

Google Analytics 4

Whether its conversions reconcile with what Google Ads reports, and where they do not.

Google Merchant Center

Feed health and disapprovals, with each finding split by where the fix lives: feed configuration or product data.

05

Why it matters

Most accounts do not need a bigger budget. They need to be read correctly first, and then changed one tested step at a time.

01

A misread account poisons every recommendation

Advice built on a wrong reading is confident and wrong. Every figure in the report is read from the account and carries its source, so the reading can be checked line by line before anything in the account is changed.

02

Lead cost is a poor guide to what converts

The cheapest leads are often not the ones that go on to buy. Costs are judged at the deepest stage with enough volume to tell groups apart.

03

Small counts move around by chance

A difference only counts if it survives a test. Where two results cannot be separated, the audit says so instead of ranking them.

06

Questions

Questions