QuantumAdsLab Logo Quantum Ads Lab
Google Ads reporting: how to build reports that lead to decisions
The difference between a report someone reads and one that changes what they do next

Google Ads Reporting: Building Reports That Drive Decisions, Not Just Numbers

In brief

What you will find in this article

  • Why most Google Ads reports are built around data availability rather than around decisions, and how to fix that
  • The decision-first framework: the single question to ask before building any report
  • Five distinct report types, each mapped to a specific decision context
  • The metrics that actually connect to decisions vs the vanity metrics that fill space
  • Direct observations from real accounts: what clients actually read, what gets ignored, and what changed behaviour
  • A concrete report structure template and the reasoning behind each section

Every week, across hundreds of Google Ads accounts, reports get sent that nobody acts on. Not because the data is wrong. Not because the account manager is bad at their job. But because the report was designed to show what happened, not to answer the question that actually matters: what do we do next?

I have been building Google Ads reporting systems since 2021, across accounts ranging from €500/month to well over €50,000/month in managed spend. The pattern is consistent: the reports that change behaviour are not the most detailed ones. They are the ones structured around a specific decision the reader needs to make.

This article is the distillation of that experience. It is separated clearly between what is principle (the framework) and what is observation (real account patterns). Both matter, but they are different kinds of knowledge.

Why most Google Ads reports fail to drive action

The failure mode is structural, not content-related. A report built around data availability looks like this: impressions, clicks, CTR, average CPC, conversions, conversion rate, cost per conversion, ROAS. All correct metrics. All present in the Google Ads interface, or in whatever google ads reporting tool pulls from the same data sources. All perfectly useless for making a decision without a framework connecting them to a question.

The person receiving the report has to do two things before they can act: first, figure out which of these metrics actually matter for their situation; second, interpret the numbers in a direction. Neither of these steps is supported by the report. So they do the easiest thing: they acknowledge the numbers, note that performance is "up" or "down", and wait for next week.

❌ Data-availability report

Built from what the platform exports. Covers everything. Answers nothing specific. The reader has to do the interpretive work. Usually ends with "overall performance was stable this period." Nothing changes.

✅ Decision-driven report

Starts with the question that needs answering this period. Shows only the data required to answer it. Ends with a specific recommendation and a predicted outcome. The reader knows what to approve or reject.

❌ Vanity metrics dominant

Impressions as the headline. CTR presented without benchmark context. "Clicks increased 12%" without connecting to revenue or conversion trend. Looks impressive. Means nothing actionable.

✅ Decision metrics dominant

ROAS vs target ROAS, with trend. CPA vs historical average, with attribution context. Impression Share Lost to Budget when budget is the constraint. Each number is tied to a lever the reader can pull.

⚠️ From direct observation, not academic research: In five years of account management I have never had a client act on a report that led with impressions as the primary KPI. I have seen clients immediately approve budget increases when the report showed, specifically, that Impression Share Lost to Budget exceeded 30% during peak conversion hours. The difference is not the metric, it is the connection between the metric and the decision.

There is a second, less obvious failure mode: the report that is too right. Too comprehensive, too detailed, with too many segments, too many charts. I have built these. They take three hours to produce and are read for 90 seconds. The client scrolls to the number they already know, total spend vs budget, confirms it, and moves on. Everything else is decoration.

The constraint to internalise: a report is not an archive. It is a communication designed for a specific decision at a specific moment. Anything in the report that does not serve that decision is friction, not value.

The decision-first framework: one question before every report

Before building any report, ask one question: what decision does the reader need to make in the next 7 days based on this data?

This is not a rhetorical question. It has a specific, concrete answer for every reporting cycle if you are doing the job properly. Some examples of real answers from real accounts:

  • "Should we increase the Performance Max budget by 30% next month?", This requires: current ROAS vs target, budget utilisation rate, Impression Share Lost to Budget, conversion volume trend.
  • "Should we pause the Display retargeting campaign?", This requires: ROAS vs Search baseline, frequency per user, CTR trend (creative fatigue signal), CPA delta vs target.
  • "Is the new target ROAS setting working?", This requires: actual ROAS vs new target over the past 4 weeks, conversion volume before and after the change, whether the campaign is spending its full budget.
  • "Which campaign gets the extra €2,000 this month?", This requires: marginal ROAS comparison across campaigns, current budget utilisation, conversion volume headroom.

Notice that each decision maps to a small, specific set of metrics. None of them are "CTR" or "impressions". They are all tied to levers: budget, bid strategy, creative, targeting. A good report is essentially a pre-made answer to the decision question, with the data arranged so the reader can validate or challenge the recommendation. It does not matter whether you connect your google ads account to a live google ads dashboard or export to a static document, the structure is what does the work.

✅ Principle grounded in communication research: The concept of "decision-oriented communication" is well established in management literature. The relevant insight for reporting is that recipients process information better when it is presented in the context of a decision they need to make, rather than as an abstract data set. This is why the same data presented as "ROAS was 4.2x this month" versus "ROAS is currently above target, we have room to scale spend by 20–30% without compromising efficiency" produces different responses in the reader.

A practical test: write the decision question at the top of your report draft before you add a single number. If you cannot articulate the question clearly, you are not ready to build the report, you need to have a conversation with the client first to understand what decision is actually on the table.

Five report types, each mapped to a decision context

Not all reporting cycles serve the same decision. After working across account types and stakeholder levels, I have converged on five report types that cover the decision landscape of a Google Ads account. Each has a different structure, different metrics, and a different cadence.

Report type Decision it serves Cadence Length Primary audience
Weekly optimisation pulse Bid adjustments, budget shifts, creative interventions this week Weekly 1 page / 5 min read Account manager (self) or hands-on client
Monthly performance review Budget allocation for next month, strategy continuation vs change Monthly 3–5 pages / 15 min read Client stakeholder, marketing manager
Quarterly strategy brief Channel mix, campaign architecture, annual budget planning Quarterly 8–12 pages / 30 min read Director, CMO, business owner
Intervention report Approve a specific change: budget increase, new campaign, bid strategy shift Ad hoc 1–2 pages / 5 min read Decision-maker who needs to sign off
Post-mortem / retrospective Learn what worked during a peak period (sale, launch, seasonal event) Post-event 4–6 pages / 20 min read Marketing team + account manager

Most reporting frameworks collapse these into a single pre built template sent on a fixed schedule, and automating reports on that single template only scales the underlying error. A monthly report sent to a CMO who needs to approve next quarter's budget looks nothing like a weekly optimisation note for an account manager deciding whether to adjust bid strategy on a campaign that just exited its learning phase.

💡 The intervention report is the most underused type. When I started doing explicit intervention reports, a single-page document structured as "here is what I want to change, here is the data supporting it, here is what I expect to happen", the approval rate on recommended changes increased significantly. The client had a concrete thing to read and respond to, rather than an implicit request buried in a monthly summary. For budget increases especially, an intervention report that shows Impression Share Lost to Budget and projects incremental conversion volume at the current ROAS almost always gets approved on the first read.

The post-mortem is the other underused type. After every significant campaign period, Black Friday, a product launch, a summer sale, a structured retrospective is the highest-value activity you can do. It answers: did our predictions match reality? What would we do differently? What data patterns should we watch for next time? Most agencies skip this because it is not billable in a traditional model. It is, however, the document that compounds knowledge across campaigns and makes the next period measurably better, even when you manage multiple clients and rely on white labeling to deliver under your own brand.

Metrics that matter vs metrics that fill space

The fastest way to evaluate a report is to go through each metric and ask: what decision does this number inform? If the answer is "none specifically" or "it gives context", that metric is a candidate for removal or demotion to an appendix. A number that describes ad performance but never informs decisions is decoration, not reporting.

Metric Decision it directly informs Role in report
ROAS vs target ROAS (4-week trend) Is the bid strategy calibrated correctly? Should we adjust the target? Primary, always in headline
CPA vs historical average Is performance improving, degrading, or stable? Primary, always in headline
Budget utilisation rate Is budget the constraint, or is the bid target? Primary for budget decisions
Impression Share Lost to Budget Should we increase budget? Primary for scaling decisions
Impression Share Lost to Rank Is bid/quality score the constraint? Supporting for bid decisions
CTR trend by ad group (4-week) Is creative fatigue occurring? Time to refresh assets? Supporting for creative decisions
Asset rating distribution in PMax Which assets to replace or supplement? Primary for PMax creative decisions
Search Terms report: new intent signals Are there new queries worth targeting or excluding? Primary for keyword/negative strategy
Conversion volume stability (weekly variance) Is the algorithm in a stable learning state? Supporting for timing of changes
Impressions (raw) Almost never a direct decision input Appendix only, context metric
Average CPC Rarely actionable without Impression Share context Appendix, remove from headline
CTR (absolute, single period) Meaningless without trend and benchmark context Only as trend, not as snapshot
⚠️ The attribution trap: ROAS and CPA numbers in Google Ads reports almost always overstate performance due to attribution window overlap, view-through attribution counting, and last-click vs data-driven model differences. Including a note on attribution model in every report is not optional, it is the context that prevents clients from making decisions based on numbers that are 20–40% inflated relative to reality. I present Google Ads ROAS alongside the same period's GA4 revenue attributed to paid search, and explain the delta explicitly. This builds trust and prevents the conversation about "why did ROAS drop 15% last month?" when the answer is actually "the attribution window changed."

Real-account observations: what I have seen work and fail

⚠️ This section describes patterns observed across managed accounts, not statistically validated research. These are operational patterns, not universal laws.

The 90-second scroll test. I started asking clients to screen-record themselves opening and reading my monthly reports. What I found: in every case, the first thing they did was scroll to the number they already knew, total spend vs budget. Then they scanned for a number that was red or in bold. If neither was alarming, they scrolled to the recommendations section. Everything in between was read only when they were looking for something specific. This observation changed how I structure reports: the summary section now contains the number they will look for, plus a single sentence framing it in business terms ("We spent 98% of the April budget and exceeded ROAS target by 12%, this creates a case for increasing May budget by €3,000").

Recommendation format matters more than recommendation quality. A brilliant recommendation buried in a paragraph gets ignored. The same recommendation formatted as "Action: increase Search campaign budget from €4,000 to €5,500 | Reason: ROAS 4.8x vs 4.0x target, budget utilisation 100% for 3 consecutive weeks | Expected outcome: +280–350 conversions/month at current ROAS | Risk: minor instability during learning adjustment" gets read, discussed and approved in the same meeting. I switched to this format in 2023 and the average time from recommendation to approval dropped from 11 days to 4 days across accounts.

The decision log compounds over time. I maintain a running decision log in each client report: every significant change made to the account in the past 12 months, the reason for the change, and the measured outcome 4 weeks later. After 6 months of consistent tracking, this becomes the most valuable section of the report, not for the current month, but because it shows patterns. It is how I identified, for one fashion e-commerce client, that budget increases made in the first week of a month consistently outperformed increases made in the last two weeks by 18–23% in incremental ROAS. That insight was only visible because the log existed.

Performance Max reporting requires explicit acknowledgement of opacity. A PMax section in a monthly report that does not acknowledge the absence of channel-level data is misleading. I write explicitly: "Note: channel-level breakdowns are not available for Performance Max campaigns. The metrics below represent aggregate campaign performance. Asset group ratings and search terms insights are the most reliable proxies for creative and intent-signal decisions." Clients who understand the opacity manage their expectations appropriately. Clients who do not understand it will ask for channel data you cannot provide, every single time.

💡 The report that changed a client's budget planning process. One e-commerce client was allocating budget on a fixed monthly basis regardless of performance. I built a single-page "budget efficiency report" that showed, for each campaign: current budget, budget utilisation rate, Impression Share Lost to Budget, estimated incremental conversions if budget were increased by 20%, and estimated lost conversions during budget-exhausted hours. The visual of "you are leaving an estimated 180 conversions per month on the table because this campaign runs out of budget at 3pm daily" was more persuasive than any recommendation I had ever written. Within two months they restructured their entire budget allocation process around this data. The report was one page, built in 45 minutes, and had more business impact than any other document I produced that year.

Frequency mismatch is the most common reporting problem. Sending weekly reports to a client who makes decisions monthly produces decision fatigue and teaches them to ignore the reports. Sending monthly reports to a client actively managing a peak period produces dangerous information gaps. I ask explicitly at account onboarding: "How often do you make decisions about this account, and who makes them?" The answer determines cadence more reliably than any industry best practice.

Concrete report structure: what I use and why

Below is the structure I use for the monthly performance review, the most common report type. Every section has a specific function; sections without a function get removed.

  • Section 1: The one thing (3 sentences max), The single most important thing that happened this period, stated in business terms. "In April, ROAS exceeded target for the third consecutive month while conversion volume grew 22% YoY. Budget is now the primary constraint, not efficiency. The recommendation for May is a structured budget increase." This section gets read by every stakeholder, including those who will not read anything else.
  • Section 2: Decision log update (table format), Each change made in the past 30 days: date, what changed, rationale, 4-week outcome. This section builds trust over time by showing that recommendations have documented outcomes, not just intentions.
  • Section 3: Current account health (5 metrics max), ROAS vs target, CPA vs target, budget utilisation, conversion volume trend (4-week), and one campaign-specific metric relevant to the current period. No impressions, no average CPC, no CTR as standalone numbers.
  • Section 4: Recommendations (1–3 max, structured format), Each recommendation in the format: Action | Reason (data reference) | Expected outcome | Risk. If you have more than 3 recommendations, prioritise ruthlessly. A client who receives 8 recommendations every month learns to approve none of them.
  • Section 5: Appendix, full data set, Complete campaign-level metrics, ad group breakdown, search terms highlights, asset performance. This is for the reader who wants to go deeper. Most will not read it every month. It serves the quarterly review and any audit conversation.
✅ Principle aligned with communication research: The "pyramid structure" in reporting, conclusion first, then supporting evidence, then detail, is consistent with how decision-makers process information under time constraints. The pyramid principle, attributed to Barbara Minto's work at McKinsey, is the structural basis for effective executive communication. In Google Ads reporting, this means the recommendation comes before the data, not after. The data is the supporting evidence for a conclusion already drawn, not a journey the reader takes to arrive at the same conclusion themselves.

FAQ on Google Ads reporting and decision-making

What is the most common mistake in Google Ads reporting?
Building reports around data availability rather than around the decisions that need to be made. The result is a document full of metrics, impressions, clicks, average CPC, that describe what happened but do not tell you what to do next. A decision-first report starts with the question "what do I need to decide this week?" and works backwards to the minimum data required to answer that question reliably.
How often should you send a Google Ads report?
Reporting frequency should match decision frequency, not the calendar. Weekly reports make sense for campaigns with sufficient data volume (50+ conversions/week) and active optimisation phases. Monthly reports work for stable campaigns where strategic adjustments are the only lever. Daily reporting almost always produces noise, not signal: with most Smart Bidding campaigns the data variance within a single day is too high to draw actionable conclusions.
Which Google Ads metrics actually matter for decision-making?
The metrics that matter are those connected to a specific decision. For budget allocation: ROAS by campaign, Impression Share Lost to Budget. For bidding: target vs actual ROAS/CPA trend over 4+ weeks, conversion volume stability. For creative: asset ratings in Performance Max, CTR trend by ad group. For audience: conversion rate by segment, Search Terms report for intent signals. Impressions, average CPC and average position rarely drive decisions directly, they belong in the appendix, not the headline.
What is the right structure for a Google Ads client report?
The most effective structure: (1) The one thing, the single most important development this period, in business terms. (2) Decision log, changes made and their 4-week outcomes. (3) Current health, 3–5 metrics reflecting account status. (4) Recommendations, 1–3 actions with structured rationale and expected outcome. (5) Appendix, full data for those who want detail. Most clients only read sections 1 and 4. Structure accordingly.
How do you report on Performance Max where data is limited?
Performance Max reporting requires explicit acknowledgement of what is not observable. The most useful proxy metrics are: asset group performance ratings (Low/Good/Best), search terms report for intent signals, conversion value trend relative to spend, and audience insights for demographic shifts. State clearly in the report that channel-level breakdowns are not available and explain why. Clients who understand the opacity manage expectations correctly; those who do not will ask for data that does not exist.
How do you handle attribution discrepancies in Google Ads reports?
Present Google Ads ROAS/CPA alongside GA4 revenue attributed to paid search and explain the delta explicitly. Google Ads numbers typically overstate performance by 20–40% due to attribution window overlap, view-through counting, and model differences. Including a consistent attribution note in every report builds trust and prevents the inevitable conversation when numbers suddenly "drop" because a setting changed. Make the attribution model part of the report infrastructure, not a footnote.

All articles

See all →