Most merchants evaluate their paid search agency through the weekly or monthly report. Spend, revenue, ROAS, conversions, week-over-week movement. The report either reassures the merchant or starts a conversation about why the numbers look off.
That’s a sensible instinct. It also misses the most consequential work a paid search agency does on a Google Ads account.
Weekly reports are reactive. They surface what happened this week. They don’t reveal whether the negative keyword list has been audited in the last eight months, whether conversion tracking is still firing correctly after the last website rebuild, whether the campaign structure still reflects what the business actually sells, or whether a branded shopping campaign that should exist simply doesn’t. All of those questions shape what the dashboard says next week, even though none of them appear on the dashboard itself.
We’ve taken to calling this dry rot. From the outside, an account looks structurally sound, the reports arrive on time, and the numbers move in a normal range. Underneath the floorboards, a tag manager change three months ago is counting an unfired event as a conversion, the bidding system is optimizing toward something meaningless, and nobody knows because nobody has looked.
What Weekly Reporting Can’t Catch
The work that finds dry rot runs on a longer cycle, and the longer cycle is the easiest one to skip.
When we audit an account that’s new to us, the pattern is consistent. Negative keyword lists are stale or never built out. Conversion tracking is wrong in ways that never got caught, because the conversion volume looked stable. Campaign structures carry layers of legacy decisions still running, because nobody went back to ask whether they should. Ad copy from two years ago runs alongside a Performance Max campaign competing for the same queries. Sitelinks, structured snippets, and callouts are outdated or absent. Half the campaign types Google now offers aren’t in use at all.
Each issue looks small in isolation, and each one stays below the weekly report’s alarm threshold. Put it all together on a single account and you have a business losing money in ways the reporting system isn’t built to show.
This is longer-cycle work, on a different cadence from weekly optimization. Roughly quarterly on most accounts, faster on accounts that change a lot, slower on small or stable ones, and triggered by specific events: site migrations, big catalog changes, tag manager reworks, unexplained shifts in conversion volume. The point is that it’s scheduled, recurring, and someone is responsible for it. If nobody owns it, it doesn’t happen.
What the Longer Cycle Is Actually For
A periodic deep review involves several distinct kinds of work, with some categories worth running every pass and others less often. Different parts matter more on different accounts.
Negation Audits
The most consistent finder of waste in a paid search account. When we run a real negation pass on an account we’ve just taken over, we routinely add hundreds of negative keywords on the first pass. On mature accounts we’ve managed for years, the count drops to a handful per pass. The work stays valuable as the account matures, even as the volume falls.
We’ve also been running the reverse operation lately. We call it un-negation: pulling negatives that no longer earn their place, because the business has changed, because match types have evolved, or because something we negated in 2023 is now a term worth letting back in. A mature account un-negation pass typically surfaces a few dozen terms worth restoring to broad matching or Performance Max eligibility.
Conversion Tracking Audits
Closer to annual than quarterly in normal cadence, but triggered any time the site, the tag manager, or the analytics setup changes. This is the dry-rot category that worries us most. The account keeps reporting numbers, the numbers look credible, and the entire optimization layer is now aimed at the wrong target.
We’ve seen a cart plugin update break tracking for users on one particular browser-and-device combination, in a way nobody had spotted, because the affected users were a fraction of the total and the conversion volume only dipped a few percent. We’ve inherited accounts running a conversion value multiplier rule set up years earlier, forgotten, and inflating reported conversion value above the real number. Neither showed up in a weekly report. Both would have shown up in an audit.
Account Structure Reviews
Step back from individual campaign tweaks and ask whether the account is built the way it ought to be built. This is a different question from “are the campaigns performing.” It includes whether campaign types that should exist actually do: brand search, branded shopping, DSAs, AI Max where appropriate, Performance Max with proper text and URL expansion, demand gen where it fits. Accounts drift away from a clean structure for understandable reasons. A periodic pass is the mechanism for deciding which drift is intentional and which is just neglect.
What the business sells changes faster than the campaign structure usually catches up, and the deeper drift is often in the product data underneath: variants that no longer exist, categories that have shifted, suppliers that have come and gone. A structure review that ignores the catalog can mistake the symptom for the cause.
Search Harvesting
Mine the broad-match, DSA, Performance Max, and AI Max search terms reports for queries that deserve dedicated targeting. Cross-reference against what Google Search Console shows you’re already ranking on organically. Anything you’re getting traffic on but not bidding on directly is a candidate.
Account-Level Extensions
Sitelinks, callouts, structured snippets, promotions, business info. Account-level material that doesn’t belong to any one campaign and ages without anyone noticing. Twice a year is usually enough.
Competitive Positioning
The longer-cycle version of competitive review asks whether the competitive set has shifted over the last few months: new competitors gaining share, competitor landing pages improving enough to warrant a strategy response, trademark behavior worth responding to.
Google Beta and New-Release Scan
Google updates the platform constantly, and most updates don’t matter on any given account. The point of a periodic scan is to make sure nothing meaningful has gone three months unnoticed. The goal is awareness rather than feature-chasing.
Business Health Review
This sits outside SEM proper, and we suspect it gets skipped most often across the industry. The work is to step back from channel metrics and look at the underlying business: total site revenue trend, new versus returning customer mix, what other channels are doing, and the supply, pricing, retail, or competitive dynamics that have shifted since the last time anyone reviewed them. An operator can run this review on their own business, or work through it with whoever has visibility across the channels. What matters is that someone owns it.
When we run a version of this with clients, we’re often surfacing shifts the client hasn’t fully named yet themselves. The worst version of that conversation is the one where SEM looks fine right up until the client tells us they’re cutting the budget because the company is in trouble.
We’re planning a separate piece on what this review actually looks like in practice. The flag here is just that it belongs on the longer cycle.
If you’re reading this list and wondering which categories matter most: negation is the work that pays for itself most reliably, conversion tracking is the work that costs the most when it’s wrong, and the business health review is the one most often skipped. The first two are the ones to confirm are scheduled on your account. The third is the one most worth asking about, because most merchants don’t know it exists as a category of work.
What This Looks Like From the Merchant Side
These audits require depth: account access, working knowledge of how a paid search account is supposed to be put together, and a few hours of focused work per pass. That’s outside the merchant’s lane, but knowing whether they’re happening is well inside it.
A few questions worth being able to get clear answers to from your agency:
When was the last full negation pass on this account, and when is the next one scheduled? A real negation pass is a half-day of structured work, separate from the weekly glance at the search terms report. If those two pieces of work blur together in the answer, the deeper version probably isn’t happening at the depth it needs to.
When was the last time someone audited conversion tracking end to end? If the site has changed at all in the last year and nobody has done a tracking audit since, schedule one now. Don’t wait for the next cycle.
When was the last time someone looked at the account structure and asked whether it still matches what the business is trying to sell? This is a different question from “are the campaigns optimized.” It’s the question of whether the campaigns that exist are the right ones, whether ones that should exist are missing, and whether the catalog underneath has drifted in a way the structure hasn’t caught up with.
What are you noticing about our business overall, beyond paid media? If the answer is some version of “we focus on SEM,” that’s useful information. It tells you who’s responsible for watching the business-level dynamics that affect everything paid media does, and whether that role needs to be filled somewhere.
The Pattern Worth Caring About
A weekly report works like a smoke detector. It catches the obvious fires, but the slow accumulation of small structural problems that erodes an account over time runs below its threshold.
The work that finds dry rot is slower, less visible, harder to demonstrate in a meeting, and easier to skip when everyone is busy. It’s also the work that separates an account growing in real terms from one running on borrowed time. If your agency can’t tell you when these passes happen, who owns them, and what the last few turned up, that’s the answer.
Find out what your weekly reports aren’t showing you.
Andrew Flicker is the VP of Operations at StatBid. With 18 years in ecommerce, his work has focused on marketing, pricing, merchandising, product content, and using large, imperfect datasets to solve practical problems - from organizing catalogs and positioning inventory to optimizing paid channels for maximum profit and efficiency. Andrew brings an operator’s mindset to StatBid, grounded in disciplined measurement, durable systems, and turning complex problems into actions merchants and marketers can actually execute.
