There's a version of a Google Ads audit that surprises almost everyone. Not because it uncovers some sophisticated structural flaw or a missed AI opportunity — but because the biggest problems are almost always the same ones that were causing issues five years ago.
The industry conversation has moved aggressively towards AI, Performance Max, and machine learning. Businesses are running to adopt the newest features, chasing automation, and assuming that if they can just configure the right campaign type, the results will follow.
Meanwhile, the fundamentals are being neglected. And the fundamentals still determine whether you win or lose auctions.
Account Structure Still Matters
The most common structural problem I see in accounts is fragmentation — too many campaigns each with too little conversion data to guide Smart Bidding effectively.
Smart Bidding needs signal. It needs to observe which users convert, what time of day they convert, what device they're on, how they came to your site. With ten campaigns each generating five conversions a month, the algorithm has nothing useful to work with. The result is unpredictable, erratic bidding that neither you nor Google can explain.
The opposite problem is equally common: one enormous campaign with no product segmentation, where high-margin products are competing for budget against low-margin ones, and there's no way to understand or adjust performance at a meaningful level.
Consolidation is the direction Google points everyone in, and in many cases that's right. But consolidation without segmentation isn't simplification — it's just less visibility. The correct structure depends on your conversion volume, your product range, and how differently your products need to be managed. There isn't one right answer, but there are plenty of wrong ones.
Negative Keywords Are Maintenance, Not a One-Off Task
Negative keyword lists get set up during account creation, then left untouched for months. I see this constantly.
Search intent changes. New products launch on your site and generate new queries. Seasonal events bring irrelevant traffic. Broad match and Performance Max reach further into query space than they did two years ago. Without regular review, irrelevant searches accumulate, each one spending a small amount of budget that would have been better spent elsewhere.
The compound effect matters. A query spending 50p per click with no conversions isn't alarming on its own. Across 200 such queries accumulating over six months, you're looking at meaningful wasted spend — and more importantly, budget that could have gone to queries that do convert.
Weekly negative keyword reviews, even brief ones, prevent this accumulation. They also serve a secondary function: new queries in the search term report often reveal demand you hadn't planned for. A term you hadn't thought to target might be generating revenue without any deliberate effort — that's a signal worth acting on.
Product Feed Quality Beats Bid Adjustments at Scale
I've put this in other articles because it keeps being true: for Shopping and Performance Max campaigns, the product feed is the primary lever.
A 10% bid increase improves your impression share by roughly 10%. Rewriting vague product titles to be specific and search-relevant can double it. The maths aren't comparable.
Google is matching your product titles and descriptions against search queries. If your titles are generic — "Blue Shirt" instead of "Men's slim-fit cotton shirt navy blue" — you're invisible for the searches that matter and appearing for ones that don't. Your bids can compensate for weak titles to some extent, but you're paying more to achieve less.
Feed quality work is unglamorous. It involves spreadsheets, product data exports, and often conversations with whoever manages the website. This is probably why it's frequently undone. But in accounts where it has been addressed properly, it consistently delivers more impact than campaign restructuring or bid strategy changes.
Ad Extensions Are Still Often Missing or Outdated
Sitelinks, callout extensions, structured snippets, and promotion extensions still expand your ad's footprint in search results, improve Quality Score, and give users more reasons to click your ad over a competitor's.
In the majority of accounts I audit, at least one of the following is true: sitelinks haven't been updated in over a year and reference promotions or pages that no longer exist; callout extensions are generic to the point of saying nothing ("Great service", "Shop now"); structured snippets haven't been set up for relevant product categories; or there are fewer than four sitelinks, which means Google can't show the full extension set.
Each of these is a quick fix. The cumulative effect of having complete, accurate, and relevant extensions is meaningful — Google rewards accounts that give users more useful information, and more ad real estate is always better than less.
Landing Page Quality Affects Quality Score — and Performance
A Shopping ad sending traffic to a category page when someone searched for a specific product is one of the most common and fixable mistakes I see.
Someone searching "men's waterproof walking boots size 10" has told you exactly what they want. Sending them to a general footwear category page with 300 products puts friction between them and the purchase. A higher proportion of those visitors will leave without buying, which over time damages your Quality Score, increases your cost per click, and reduces your ad rank.
Sending Shopping traffic to the specific product page — or the most relevant filtered category page — improves conversion rate directly and lowers your effective CPC through Quality Score improvements. It's a double benefit.
The audit question is simple: for your top 20 spending products in Shopping, where does the click actually go? Is that the most relevant destination on your site?
The Bigger Point About AI and Foundations
There's a logic that says: if Google's AI is handling bidding, targeting, and ad serving, the fundamentals matter less because the machine will compensate for weak foundations.
The opposite is closer to the truth. AI amplifies what you give it.
A well-structured account with clean conversion data, relevant product titles, accurate negative keyword lists, and correct landing pages gives Smart Bidding strong signal. The algorithm learns faster, bids more accurately, and delivers better results. The AI is working with a strong foundation.
A poorly structured account with fragmented conversion data, generic feed titles, months of accumulated irrelevant queries, and landing pages that don't match user intent gives the algorithm weak signal. It learns slowly, makes poor bidding decisions, and underperforms — not because AI is bad, but because it's been given bad inputs.
The businesses gaining ground in 2026 are mostly not the ones that found the cleverest automation. They're the ones that got the basics right and then let automation do its job from a position of strength.
The practical action: pick one area from this article — structure, negatives, feed quality, extensions, landing pages — and audit it properly. Fix what you find. Then move to the next. The compounding effect of getting multiple fundamentals right simultaneously is where the real gains are.
If you'd like an objective view of where your account stands across these areas, a Google Ads audit gives you a clear picture of where to focus first.