What tool tells me what to change in Google Ads every week?

Updated November 12, 2026

The weekly Google Ads change list

The weekly change list in Google Ads has five fixed categories: negative keywords, budget moves, pauses, bid target adjustments, and ad maintenance. Everything else is either an urgent same-day item (tracking breaks, disapprovals, runaway spend) or a monthly strategy item (structure, offers, landing pages).

Working from fixed categories matters more than it sounds. The Google Ads interface offers you several hundred things you could touch on any given day, and most of them should not be touched. A fixed weekly list is how experienced account managers avoid thrash, and it is the format AdsGuru's weekly proposals follow.

• Negatives: junk queries from the week's search terms, clustered by intent and placed at the right level.
• Budgets: pace month-to-date spend against plan; move budget toward campaigns clearing the CPA or ROAS target.
• Pauses: keywords and ads that have spent past the threshold with nothing to show for it.
• Bid targets: nudge tCPA or tROAS where conversion data moved meaningfully, in small steps.
• Ads and assets: fix disapprovals, keep two-plus active RSAs per ad group, refresh what fatigued.
• Flags to read, not act on: auction shifts, PMax drift, conversion lag artifacts.

Why weekly is the right cadence

Weekly is the right cadence because it is slow enough for the data to mean something and fast enough that mistakes do not compound. Daily tinkering reads noise: on most accounts a single day holds too few conversions to judge anything, and every meaningful edit can nudge Smart Bidding back toward relearning. Monthly reviews have the opposite problem: a broken search term or a mispaced budget runs for weeks before anyone looks.

Smart Bidding is the technical reason to batch changes. Target CPA and target ROAS strategies build a model of your account's conversion behavior; significant changes to targets, budgets or structure send parts of that model back into adjustment. One deliberate change-day per week gives the algorithm six quiet days to settle, which is the pattern Google's own documentation points toward when it warns against frequent edits.

The exception list is short and genuinely urgent: conversion tracking failures, policy disapprovals on core ads, and spend running far past plan. Those get fixed the day they appear. Everything else waits for the batch, and waiting is a feature.

The Monday routine, step by step

Here is the full weekly routine as a checklist. By hand it takes 30 to 60 minutes on a typical single account; with a tool proposing the changes it compresses to the approval pass.

1. Check the vitals first: did conversions record every day last week, are there new disapprovals, is month-to-date spend within plan? Any failure here becomes the whole session.
2. Pull 7 days of search terms. Add negatives for junk queries, at ad group level for specific mismatches and shared lists for account-wide junk. Never negate a query with conversions.
3. Review pacing: divide the monthly budget by days elapsed, compare to actual. Throttle or release campaign budgets in steps of 20% or less.
4. Scan for pause candidates: keywords and ads past one target CPA of spend since the last conversion. Check sibling context before pausing; a weak ad in a strong ad group is different from a weak ad group.
5. Review bid targets: if a campaign beat its target CPA consistently for two weeks, tighten or scale; if it missed, loosen the target 10-15% or fix the underlying issue instead.
6. Check ads: every active ad group has at least two RSAs, no disapprovals, assets serving. Note which creative is fatiguing for the monthly refresh.
7. Read the flags you will not act on: Auction Insights shifts, PMax search themes, conversion lag. These inform next week's judgment.
8. Write down what you changed and why. The log is what lets you learn from the account instead of repeating experiments.

Search terms and negatives: the weekly pass

The search terms pass is the highest-yield ten minutes in Google Ads. Broad match and Performance Max both expand your reach into queries you never chose, and the search terms report is the only place you see what you actually paid for. Left unreviewed for a month, most accounts accumulate a visible layer of junk spend: job seekers, DIY researchers, wrong products, wrong cities.

The discipline is to cluster rather than whack individual queries. One 'plumber salary' query is a data point; the cluster of salary, jobs, apprenticeship and hiring queries is an intent family that one phrase negative kills forever. Clustering is also what keeps the negative list maintainable instead of becoming a 4,000-row pile nobody understands. The full method is on the search terms cleanup page.

Budget pacing: the weekly pass

Budget pacing answers one question: will this month's spend land on plan, and is it flowing to the campaigns that earn it? Google spends up to twice a campaign's daily budget on strong days and compensates on weak ones, capping the month near daily budget times 30.4, so daily wobble is normal and monthly drift is the thing to manage.

A worked example, with illustrative numbers: say the plan is $9,000 for the month and on day 15 you have spent $5,200. You are pacing to roughly $10,400, about 15% hot. The wrong fix is slashing budgets in half, which starves the strong campaigns at mid-month. The right fix is trimming the campaigns running past plan that are also missing their CPA target, in 20% steps, and rechecking in a week. Pacing corrections should be boring; drama in budget management is a sign the cadence was too slow. When you want this pass run daily instead of weekly, that is the budget automation loop.

Bids and targets: change them less than you think

Bid targets reward patience more than any other lever in the account. With target CPA or target ROAS bidding, the weekly question is whether the target still matches reality, and the answer is usually yes, leave it alone.

When a change is justified, small and infrequent wins. Moving a target CPA more than about 15-20% in one step can push the strategy into a relearning period where delivery goes volatile for days. The compounding pattern that works: a campaign beats its target for two straight weeks, you tighten 10%, wait two weeks, repeat. Over a quarter this walks the account toward efficiency without ever destabilizing it.

Also respect conversion lag. If your typical buyer converts three days after the click, the last three days of any report are systematically undercounted, and judging them as failures leads to Friday panic edits that Monday data would have reversed.

What not to change every week

Half the value of a weekly list is what it keeps you from touching. These levers move slower than a week and churn the account when handled weekly:

• Bidding strategy type: switching between manual, tCPA and tROAS resets learning every time; change strategy at most quarterly and with a reason.
• Campaign structure: merging and splitting campaigns weekly guarantees you are always in a learning phase.
• Attribution settings: changing the model mid-flight rewrites your history and makes week-over-week comparison meaningless.
• Match types wholesale: converting the account to broad match (or back) is an experiment to run deliberately on one campaign, with a holdout.
• Geo and audience targets on thin data: a location with 40 clicks and no conversions is noise, a quarter of data is signal.
• Creative wholesale swaps: replace fatigued ads one at a time so you can attribute the result.

The cadence table

Every change type has a natural cadence, a trigger that justifies it, and a cost when overdone. This table is the whole discipline in one view:

Reading the week honestly: sometimes the list is empty

An honest weekly review sometimes produces no changes, and that outcome should be reported as proudly as a long list. Accounts in a stable groove are earning the compound interest of past discipline; stirring them for the sake of visible activity is how agencies justify retainers and how tools justify dashboards.

The practical test for any change on the list: can you say what data triggered it, what you expect it to do, and when you will check whether it worked? A change that cannot answer all three is activity, and activity is a cost.

Where AdsGuru automates the weekly list

AdsGuru runs the daily checks and batches the week's proposed changes into exactly this list: each negative, budget move, pause and target nudge arrives with the reason and the underlying data, and applies only after you approve it. Urgent items, like tracking breaks and disapprovals, alert immediately instead of waiting for the batch. Every applied change is logged, so the account history reads like a decision journal.

Honest limits: AdsGuru needs 7 to 14 days of account data before its weekly proposals carry real weight, and it will propose an empty list when the account does not need changes. The approval pass takes most users a few minutes; see how AdsGuru works and pricing, which is $89/month flat with a 7-day free trial.

Common questions

Related