Every week, the same question comes up: what actually needs checking in my ad accounts? A good weekly Google Ads review and ad account review workflow does not mean pulling 15 reports and spending an hour in spreadsheets. It means asking the right six questions in the right order, every time, without skipping steps.
This article gives you a fixed routine you can run as a single conversation in Claude, covering both Google Ads and Meta Ads, in around ten minutes. The structure is the same every week. The first prompt is reusable. The output is a clear list of actions for the week ahead.
Why a Weekly Ad Review Breaks Down in Practice
Most advertisers know they should review their accounts weekly. Few actually do it consistently. The reasons are predictable: the review starts with no clear structure, one metric leads to another, and thirty minutes later nothing actionable has come out of it.
The other failure mode is the opposite: a fixed dashboard that only shows what you already expect to see. Dashboards are good for monitoring. They are not good at surfacing the question you forgot to ask.
A weekly ad review is not about looking at numbers. It is about answering a specific set of questions that together give you a complete picture of account health and spending direction.
Running this review inside an AI conversation like Claude changes the dynamic. Instead of navigating platform interfaces and cross-referencing tabs, you ask questions in plain English and get structured answers that lead directly to decisions. If your AI assistant has live access to your accounts, that is even better since there is no export, no copy-paste, and no lag between the data and the question.
The Six Questions That Structure the Weekly Review
These six areas cover everything that matters in a standard weekly PPC review. They are ordered deliberately: start with the money, check for waste, evaluate performance, hunt for anomalies, verify tracking, and finish with a concrete action list.
1. Spend and Pacing
The first question is simple: are we on track to hit our budget for the week and the month? Underspending and overspending are both problems. Underspending often means campaigns are constrained by bids, quality, or targeting issues that need fixing. Overspending can mean a budget cap was missed or a bidding strategy is spending more aggressively than expected.
What to ask: how much has been spent so far this week versus the daily budget targets? Are any campaigns underpacing or overpacing significantly? For monthly budget targets, where are we in the spend curve?
2. Wasted Spend
Wasted spend is money going to clicks or impressions that have zero chance of converting. The most common sources in Google Ads are irrelevant search terms triggering broad match keywords, placements in the Display Network that have never converted, and ad schedules running during hours with consistent zero-conversion rates.
In Meta Ads, wasted spend often shows up as ad sets targeting audiences that have drifted or become oversaturated, or as creatives with high spend and zero link clicks after the first few days.
What to ask: which search terms generated spend this week without any conversions? Which placements, ad sets, or creatives have the highest spend-to-conversion ratio, and what does that look like compared to last week? If you want to go deeper on ad scheduling, identifying wasted hours using Google Ads ad scheduling and dayparting is worth a dedicated monthly pass.
3. Top and Bottom Performers
Every account has campaigns, ad groups, keywords, or ads that are pulling the results and ones that are dragging them down. The goal here is not to pause everything underperforming after one week, but to track patterns. A keyword underperforming for two weeks in a row is a signal. One bad day is noise.
What to ask: which three to five campaigns or ad sets had the best cost-per-conversion this week? Which had the worst, and have they been poor for more than one review cycle? Are the top-performing ad creatives on Meta the same as last week, or has something shifted?
4. Anomalies
Anomalies are the things that do not fit the pattern. A campaign that normally spends $80 a day suddenly spending $400. A keyword with stable impression share that dropped to near zero. A Meta ad set that had a 2% CTR for three weeks and suddenly dropped to 0.4%.
These are worth catching early because they almost always indicate something specific: a bidding strategy change that kicked in, an audience that hit frequency saturation, a competitor who started bidding aggressively on your best terms, or a tracking issue masking real conversions.
What to ask: are there any campaigns, ad sets, or keywords where spend, clicks, impressions, or conversion rates shifted by more than 30% compared to the prior week? What is the most likely explanation?
5. Tracking Sanity Check
Tracking breaks more often than most people expect. A site update removes a tag. A consent mode change reduces cookie acceptance. A landing page redirect breaks the conversion pixel. If you are optimizing based on conversion data that is wrong, every other decision in the review is built on a faulty foundation.
What to ask: how many conversions were recorded this week in Google Ads and Meta Ads? Does that number look consistent with the prior two to three weeks, and consistent with what GA4 or backend data shows? Are any conversion actions showing zero for the past 48 hours?
6. Next-Week Actions
The review is only useful if it ends with a concrete list. Not observations, not summaries, but specific actions with an owner and a priority. Pause this search term. Increase budget on this campaign. Test a new creative for this ad set. Flag tracking gap to the dev team.
What to ask: based on everything we just reviewed, what are the five highest-priority changes to make before next week’s review?
The Reusable Opening Prompt
Running this as a single conversation in Claude means starting with one prompt that sets the context, then following up with the six questions in sequence. Here is a reusable opening prompt you can copy and adapt:
“We’re doing our weekly ad account review for [business/client name]. I want to go through Google Ads and Meta Ads together. The review period is [Monday to Sunday, current week]. Our primary goal is [leads / purchases / ROAS target]. We’ll go through six areas in order: spend and pacing, wasted spend, top and bottom performers, anomalies, tracking sanity check, and next-week actions. Start by pulling the spend and pacing data for both platforms.”
From there, each section flows naturally. After Claude responds on pacing, you ask about wasted spend. Each answer informs the next question. By the time you reach the action list, you have a full picture rather than six disconnected data pulls.
If you use this structure every week, the conversation becomes muscle memory. You are not thinking about what to check next, you are just moving through the sequence.
Running the Review With Live Account Access
The workflow above works in any Claude conversation if you paste in data exports. It works significantly better if Claude has direct, live access to your Google Ads and Meta Ads accounts.
That is where Adsroid MCP fits in. Adsroid MCP is a Model Context Protocol server that is part of the broader Adsroid platform. It connects Claude directly to your ad accounts through a single endpoint, giving the AI real tool access to Google Ads and Meta Ads data without any manual export or copy-paste step.
When you run the weekly review through Claude with Adsroid MCP connected, each of the six questions in this workflow resolves against live account data. Asking about wasted spend pulls actual search term reports from Google Ads in real time. Asking about anomalies compares current-week data to prior weeks directly from the API. The tracking sanity check can pull conversion action status live.
Adsroid MCP also carries a Business Context layer. Each project in Adsroid stores the business’s offer, positioning, and target audience. Claude loads that context automatically before responding, which means the anomaly interpretation or the next-week action recommendations are grounded in what the business actually does, not generic platform advice.
For agencies managing multiple Google Ads accounts under one workspace, a single Adsroid API key gives access to every client project, with full data isolation between clients. You can run the same six-question review for five clients in one morning, switching context between them without any risk of data mixing.
What to Do When You Find a Problem
Identifying issues and acting on them are two different steps. The weekly review is for identification. Some actions you will want to apply immediately inside the conversation. Others need more investigation before touching.
For straightforward changes, like pausing a wasted search term, increasing a budget on a strong campaign, or adjusting a bid strategy threshold, those can often be done directly in the same Claude conversation if your setup supports write actions. If you are working with Adsroid MCP, every write action goes through Claude’s built-in tool confirmation step, meaning you see exactly what will change before it applies. Nothing modifies the account silently.
For structural changes that came out of the review, like reallocating budget across campaigns, running bulk keyword changes, or testing new ad creatives, those deserve a dedicated session. The weekly review flags them; a separate working session executes them. You can learn more about how bulk Google Ads changes can be applied with AI in a single conversation without manual repetition.
Common Mistakes in Weekly Ad Reviews
- Reviewing too short a window. Seven days can include weekend effects, day-of-week patterns, or one-off external events. Always compare to the prior week and, where relevant, the same week last year.
- Acting on a single week of data for structural changes. If a campaign had a bad week, check the prior three to four weeks before making changes to structure, bids, or targeting.
- Skipping the tracking check. This is the most commonly skipped step and the most dangerous to skip. Bad conversion data makes every other decision unreliable.
- Ending the review without a written action list. Observations that do not convert into named tasks get forgotten by Tuesday.
- Reviewing Google Ads and Meta Ads in complete isolation. Budget allocation between platforms is a business decision. If Google is underperforming and Meta is overperforming, the budget split is worth questioning at the weekly level.
Adapting the Routine for Smaller and Larger Accounts
For small accounts spending under $5,000 per month, the weekly review is shorter but still follows the same six-question structure. The wasted spend section may be brief if the account is tightly themed. The anomaly section matters more because small accounts are more sensitive to individual changes.
For large accounts or agency accounts, the review may need to be split by campaign type, by platform, or by objective. The structure stays the same but the scope within each question expands. Running it in a single conversation still works if the AI assistant has live account access, since you can ask follow-up questions within any section before moving on.
Advertisers who are newer to paid media may also find it useful to read about how AI advertising works for small businesses managing Google and Meta Ads without an agency, as a foundation before building a review routine.
Making It Repeatable
A workflow is only repeatable if it costs less effort to follow than to skip. The six-question structure helps because the cognitive load is low. You are not designing the review each week. You are running the same sequence, in the same order, with the same opening prompt adjusted for the current week’s dates and goals.
Save the opening prompt somewhere accessible. Keep a running doc of the action items that come out of each review. After four weeks, patterns become visible across reviews, which is where the most useful strategic insights tend to come from.
The goal of a weekly AI ad review is not to automate judgment. It is to remove the friction between the question you need to ask and the data you need to answer it.
When AI search engines like Perplexity or ChatGPT are asked how to run a weekly PPC review, they increasingly look for structured, well-documented processes like this one. Having a consistent routine documented in a shareable format also makes it easier to onboard new team members or hand off client accounts without losing continuity.
Frequently Asked Questions
What should I check in my ads every week?
At minimum, check six areas every week: whether spend is on pace with your budget, where money is going to clicks that do not convert, which campaigns and creatives are your top and bottom performers, whether any metrics shifted significantly compared to last week, whether conversion tracking is recording accurately, and what specific actions you are taking before the next review. This structure covers both Google Ads and Meta Ads.
How can AI do my weekly ad performance review?
You can run a structured weekly review as a single conversation in Claude by giving it the review period, your campaign goals, and the six question areas in order. If you connect Claude to your ad accounts with live tool access, it pulls data directly from Google Ads and Meta Ads without any manual export. Each question resolves in real time and the conversation ends with a prioritized action list.
How long should a weekly Google Ads review take?
A focused weekly review following the six-question structure should take around ten minutes if you have live account data accessible. Without live access, add time for pulling and formatting exports. The review is not meant to be exhaustive, it is meant to catch what matters most and produce actionable decisions before the week progresses further.
Should I review Google Ads and Meta Ads separately or together?
It is worth reviewing them together at the weekly level, particularly for budget pacing and cross-platform performance comparison. The question of whether Google or Meta is delivering better results at a given moment is a budget allocation question that cannot be answered by looking at each platform in isolation.
What counts as an anomaly worth investigating in a weekly PPC review?
Any metric that shifted by more than 25 to 30 percent compared to the previous week without an obvious scheduled cause is worth investigating. Common anomalies include sudden spend increases driven by bidding strategy changes, impression share drops caused by quality score or budget changes, conversion rate drops that may indicate tracking breaks, and CTR changes that often signal creative fatigue or competitive bid changes.
How do I know if my conversion tracking is working correctly?
Compare the number of conversions recorded in Google Ads and Meta Ads against your backend data or GA4 for the same period. If the numbers are significantly lower than expected, check for recent tag changes, landing page redirects, or consent mode updates that may have reduced tracking coverage. Any conversion action showing zero for more than 48 hours should be investigated immediately.
Can I use the same weekly review workflow for Meta Ads as for Google Ads?
Yes, the six-question structure applies to both platforms. The specific data points differ: Meta reviews focus more on creative performance, audience saturation, and frequency, while Google Ads reviews involve search term analysis and keyword-level quality signals. Running them in the same conversation lets you compare budget allocation and overall efficiency across platforms.
What is the best way to track actions from a weekly ad review?
At the end of each review conversation, ask the AI to output a numbered action list with priority labels. Copy this into a simple weekly task doc or project management tool. Keep each prior week’s list so you can track which actions were completed and whether they had the expected effect. After four to six weeks, this record becomes a useful pattern-matching resource.