Over the past three years at BTB Media, we've audited more than a hundred Google Ads accounts for e-commerce and D2C brands. Skincare companies, fashion labels, supplement brands, home goods stores, pet product businesses — the verticals vary, but the mistakes are remarkably consistent.
What surprises most founders is that these aren't obscure, technical blunders. They're structural problems hiding behind dashboards that look perfectly healthy. A 6x ROAS that masks an acquisition problem. A conversion count that's 40% inflated. A Performance Max campaign that's spending thousands on queries no one has ever reviewed.
This article walks through the seven mistakes we see in almost every audit — not as abstract best practices, but as stories from actual accounts. For each one, we'll cover what we found, why it matters, how we fixed it, and what happened afterward.
If you're running Google Ads for an e-commerce brand and haven't had a proper audit in the last six months, at least two of these are probably happening in your account right now. If you want the full diagnostic framework, start with our complete Google Ads audit guide for e-commerce.
1. Running Performance Max With Zero Product Feed Optimization
Last month we audited a skincare brand doing $80K per month in revenue through Google Ads. They had three Performance Max campaigns running, a decent daily budget, and a ROAS that hovered around 3.5x. On the surface, it looked fine. Then we opened the product feed.
Their product titles were internal SKU names. Not "Hydrating Vitamin C Serum for Dry Skin — 30ml" but "VC-SRM-030-HYD." Their descriptions were manufacturer boilerplate copied from the supplier. Key attributes like skin type, key ingredients, product benefits, and GTIN were either missing or auto-populated with placeholder text. There was not a single custom label in the entire feed.
What We Found
Performance Max uses your product feed as one of its primary signals for targeting. When titles are SKU codes instead of keyword-rich descriptors, Google has almost no signal to match your products to the right search queries. The algorithm was guessing — and spending the budget while it guessed. We pulled the search term insights and found that over 30% of impressions were going to generic, low-intent queries that had almost nothing to do with their actual product range.
The fix was not complicated, but it was labor-intensive. We rewrote every product title to follow a structured format: Brand + Product Type + Key Attribute + Size/Variant. We rebuilt descriptions around search intent, filled in every missing attribute, added GTINs where available, and set up five custom labels for margin tier, best-seller status, price range, seasonal relevance, and new arrival flags.
After Fixing
Within six weeks, the same budget drove a 42% increase in qualified impressions and ROAS climbed from 3.5x to 5.1x. The irrelevant query rate dropped from 30% to under 8%. The feed didn't just improve Performance Max — it lifted Standard Shopping performance too, because Google finally understood what they were selling.
This is the single most common mistake we encounter. If you're running Performance Max with a default Shopify feed, you are leaving money on the table. Period. Our Google Ads audit checklist includes 12 specific feed checks you can run yourself.
2. No Separation Between Branded and Non-Branded Search
A fashion brand came to us celebrating an 8x ROAS across their search campaigns. They were spending around $25K per month and their founder genuinely believed the ads were performing brilliantly. He wasn't wrong that the dashboard showed great numbers. He was wrong about what those numbers meant.
We segmented their search terms into branded queries (people searching the brand name directly) and non-branded queries (generic product searches). Branded ROAS was 22x — people who already knew the brand were clicking ads and buying. That is not customer acquisition. That is paying for traffic you would largely get organically. Non-branded ROAS was 1.2x. Barely breaking even on an accounting basis, losing money on a true unit-economics basis once you factor in COGS, shipping, and returns.
What We Found
65% of the total ad spend was going to branded terms. The brand was essentially paying Google $16,250 per month to capture customers who were already searching for them by name. Meanwhile, the non-branded campaigns that were supposed to drive new customer acquisition had never been properly optimized because the blended ROAS looked healthy. This is one of the most common reasons we see brands plateau — they think ads are working when they're really just subsidizing existing demand.
We restructured the account into clearly separated branded and non-branded campaigns. Branded search got its own campaign with a much lower budget and a high Target ROAS target to maintain efficiency without overspending. The freed budget was reallocated to non-branded campaigns, which we rebuilt with tighter keyword grouping, more aggressive negative keyword lists, and landing pages matched to search intent.
After Fixing
Branded capture continued at similar volumes with 60% less spend. Non-branded ROAS improved from 1.2x to 2.8x over eight weeks as the reallocated budget found better placements. Total new customer acquisition increased by 34% — and the brand finally understood the true cost of acquiring a customer, not the vanity metric it had been reporting to investors.
If you don't know your non-branded ROAS right now, your Google Ads account is lying to you. This is one of the first things we check in every audit, and it is covered in detail in our full e-commerce Google Ads audit guide.
3. Conversion Tracking That Double-Counts (or Worse)
A home goods brand hired us because their agency kept reporting growth while actual revenue was flat. The agency dashboard showed 340 conversions in the last month. Shopify showed 195 total orders. That is not a rounding error. That is a 74% discrepancy.
We dug into the conversion actions configured in the Google Ads account. There were seven different conversion actions active: two purchase events (one from the Google Ads tag, one from an older Google Analytics import that nobody remembered setting up), an add-to-cart event marked as a primary conversion, a "begin checkout" event also marked as primary, a newsletter signup, a page view event on the thank-you page, and a duplicate purchase event imported through a third-party app. Three of these were set as primary conversions, meaning Google's bidding algorithm was optimizing toward a blended signal of purchases, cart adds, and checkout starts.
What We Found
The bidding algorithm thought it was getting 340 conversions when the real number was 195 actual purchases. This meant the algorithm believed it was performing almost twice as well as it actually was, so it never made the bid adjustments it should have. It was over-bidding on low-quality traffic because that traffic triggered cart adds and checkout starts (which counted as conversions) even when those sessions never completed a purchase. Every optimization decision the agency made based on this data was built on a broken foundation.
We cleaned the conversion setup down to a single primary conversion action: verified purchases, deduplicated using transaction ID matching between the Google tag and Shopify. All other events were either removed or reclassified as secondary conversions so they could be observed without influencing bidding. We also installed a server-side tag through Google Tag Manager to ensure tracking accuracy wouldn't degrade with browser-level ad blockers and cookie restrictions.
After Fixing
The reported ROAS dropped immediately — because the numbers were finally honest. But over the next six weeks, actual ROAS improved by 28% as the bidding algorithm learned from accurate data. Cost per acquisition dropped from $38 to $27. The brand finally had numbers they could trust to make scaling decisions. They also fired the agency.
Conversion tracking is the foundation everything else sits on. If this is wrong, nothing you optimize will produce reliable results. If your Google Ads conversions are more than 15% off from your store data, something is broken. This is also one of the clearest signs of bad agency management — any competent agency should catch this in the first week.
Seeing Yourself in These Mistakes?
These are the exact issues we uncover in our e-commerce Google Ads audits. If you suspect your account has similar problems, we'll find them — and give you a prioritized action plan to fix each one.
Request a Free Audit4. One Bid Strategy for All Campaigns, Regardless of Volume
A supplement brand had 12 campaigns in their account, all set to Target ROAS at 400%. It looked organized. It looked intentional. It was neither. When we pulled the data, three campaigns had more than 30 conversions per month — enough data for Target ROAS to actually work. Nine campaigns had fewer than 10 conversions per month. Some had three. One had zero in the last 30 days and was still set to tROAS 400%.
Smart Bidding algorithms are statistical models. They need data to learn. Google themselves recommend at least 30 conversions in the past 30 days for Target ROAS to function properly. When you give a tROAS target to a campaign with five conversions a month, the algorithm has almost no signal. It responds in one of two ways: it either barely spends at all (throttling delivery because it can't confidently hit the target) or it spends erratically, overpaying for some clicks and ignoring others with no discernible pattern.
What We Found
Of the nine low-volume campaigns, four were spending less than 15% of their daily budgets — the algorithm was effectively paralyzed. Three were spending the full budget but at a 1.8x actual ROAS (far below the 4x target), meaning the algorithm was ignoring the target because it didn't have enough data to learn from. Two had such erratic performance that week-over-week results swung by 300% or more. Combined, these nine campaigns were wasting approximately $8,500 per month.
The fix required restructuring campaigns by volume tiers. The three high-volume campaigns stayed on Target ROAS with refined targets based on actual performance data. Five mid-volume campaigns were switched to Maximize Conversions with a budget cap — giving the algorithm a simpler objective it could actually optimize for. The four lowest-volume campaigns were consolidated into two broader campaigns to pool conversion data, then started on Maximize Conversions with a plan to graduate to Target ROAS once they hit the 30-conversion threshold.
After Fixing
Total account conversions increased by 23% in the first month with roughly the same budget. The consolidated low-volume campaigns crossed the 30-conversion threshold within six weeks, at which point we migrated them to Target ROAS and saw a further 15% efficiency improvement. The supplement brand went from thinking they needed more budget to realizing they needed better structure.
5. Ignoring Search Term Reports for Months
A beauty brand was running both Search and Performance Max campaigns. When we asked for the last time someone reviewed search terms, the marketing manager paused and said, "I think the agency set up negatives when the campaigns launched." The campaigns had been running for nine months.
We exported the search term data for the previous quarter. The results were painful. $14,000 in spend over three months on queries like "free samples," "DIY face cream recipe," "how to make moisturizer at home," competitor brand names they weren't bidding on intentionally, and a surprisingly large cluster of queries related to a skin condition their products had nothing to do with. One query — "free skincare samples no purchase required" — had accumulated $1,200 in spend with zero conversions.
What We Found
22% of the total search spend in the previous quarter went to queries that had zero commercial intent. These were people looking for free things, DIY tutorials, medical information, or competitors — none of whom were going to buy. The Performance Max search term insights were even worse, because PMax casts a wider net by design, but nobody was reviewing the insights report to catch the irrelevant themes and add them as account-level negatives. The brand was effectively subsidizing Google's broad-match learning at $4,700 per month.
We built a comprehensive negative keyword list — 480 terms in the first pass — organized by theme: informational queries, DIY-related terms, free/sample seekers, competitor names, medical/condition terms, and geographic modifiers that didn't apply. These were applied at both campaign and account level. We also set up a weekly review process with a shared spreadsheet so the brand could see exactly what queries were being added and excluded.
After Fixing
Wasted spend dropped from 22% to under 4% within the first month. ROAS improved by 31% with no budget increase. Click volume actually decreased by about 18%, but conversion rate jumped by 40% because the remaining traffic was dramatically more qualified. The brand realized that more traffic was never the problem — better traffic was.
6. No Custom Labels Means No Margin-Based Bidding
A home and kitchen brand had about 600 products in their catalog, ranging from $12 kitchen gadgets with 20% margins to $89 premium cookware sets with 65% margins. Every single product was in one Performance Max campaign with one asset group. Google was spending to promote all of them equally.
Think about what that means in practice. The algorithm was spending just as aggressively to sell a $12 garlic press with $2.40 in profit as it was to sell an $89 cookware set with $57.85 in profit. The cost per click for both was roughly the same. But the garlic press needed a CPA below $2.40 to be profitable (nearly impossible in paid search), while the cookware set was profitable at any CPA under $58.
What We Found
38% of the total ad spend was going to products in the lowest margin tier — products that were mathematically unprofitable to advertise at any reasonable CPA. The high-margin products that should have been the stars of the account were getting less than 20% of total spend. The brand's blended ROAS looked acceptable at 3.2x, but when we calculated true profit-adjusted ROAS, it was closer to 1.4x. They were spending $42,000 a month and barely making a profit on the ad spend once you accounted for actual product margins.
We rebuilt the product feed with custom labels for margin tier (high, medium, low, and exclude), best-seller status, price range bracket, and seasonal flags. Then we restructured the campaigns: high-margin products got their own Performance Max campaign with an aggressive budget and a moderate ROAS target. Medium-margin products went into a separate campaign with a higher ROAS target. Low-margin products were either excluded from paid advertising entirely or placed into a very conservative campaign with strict CPA limits.
After Fixing
Within two months, profit-adjusted ROAS went from 1.4x to 3.6x — a 157% improvement — even though headline ROAS only increased modestly from 3.2x to 3.8x. The difference was entirely in what they were selling, not how much. Total ad-attributed profit increased by $11,400 per month on the same budget. The brand started expanding budget because for the first time, every additional dollar spent was going to products that actually made money.
7. Landing Page Mismatch — Sending Shopping Traffic to the Wrong Page
A children's clothing brand was running Shopping ads for "organic baby pajamas." We clicked on one of their ads. It went to the homepage. Not the product page. Not the category page for baby pajamas. The homepage — with a hero banner promoting a sale on toddler shoes.
This particular issue had multiple layers. Some products in the feed had the main website URL as the link instead of the specific product URL. Other products pointed to category pages that loaded slowly and required three clicks to find the actual product the person searched for. A handful linked to product pages that were out of stock, showing an "email me when available" button instead of an add-to-cart option. The brand was paying for every one of those clicks.
What We Found
17% of products in the feed had incorrect or suboptimal landing page URLs. Of those, about half went to the homepage and half went to broad category pages. The conversion rate for products with correct, direct product page links was 3.8%. The conversion rate for products landing on the homepage was 0.4%. For category pages, it was 1.1%. The brand was spending roughly $6,200 per month driving traffic to pages where almost nobody converted.
The fix started with a complete audit of every product link in the feed. We corrected all URLs to point directly to the specific product page. We set up automated rules to pause products that went out of stock so the brand wouldn't pay for clicks to unavailable items. We also worked with their dev team to ensure product pages loaded in under 2.5 seconds on mobile — several had been clocking in at 6-8 seconds, which tanks conversion rates regardless of how relevant the traffic is.
After Fixing
Conversion rate across all Shopping traffic increased from 2.1% to 3.4% — a 62% improvement. Bounce rate dropped from 64% to 38%. The out-of-stock automation alone saved an estimated $1,800 per month in wasted clicks. Combined ROAS went from 2.6x to 4.1x over the following month. This was arguably the simplest fix in this entire article, and it had one of the largest impacts.
The Pattern Behind All Seven Mistakes
If you look at these seven mistakes as a whole, the underlying pattern is the same: the dashboard says things are fine, but nobody is looking beneath the surface. Performance Max hides query-level details. Blended ROAS hides acquisition problems. Inflated conversion counts hide tracking failures. One-size-fits-all bid strategies hide volume mismatches. And when nobody reviews search terms or feed quality, the waste compounds month after month.
Every one of these brands had a Google Ads dashboard that looked reasonable. Some looked great. The problems only became visible when someone sat down and asked the right questions — not "what is the ROAS?" but "what is the non-branded ROAS?" Not "how many conversions?" but "do conversions match actual orders?" Not "what is the CPA?" but "what is the CPA on our high-margin products specifically?"
That is what an audit does. It asks the questions your dashboard doesn't. If you haven't had one in the last six months, the odds are high that at least two or three of these issues are active in your account right now. If you want to run the diagnostic yourself, our 50-point Google Ads audit checklist covers every check mentioned in this article and more.
Frequently Asked Questions
What is the most common Google Ads mistake for e-commerce brands?
Running Performance Max campaigns without optimizing the product feed is the single most common mistake we see. Most brands leave default product titles (often internal SKU names or manufacturer codes), generic descriptions, and miss key attributes like color, size, material, and GTIN. Since Performance Max relies heavily on feed data to match search intent, a poor feed means Google is essentially guessing — and spending your budget while it does. The fix is straightforward but labor-intensive: rewrite titles using a keyword-rich format, complete all attributes, and add custom labels for margin and performance tiers.
Why is my Google Ads ROAS high but sales are not growing?
This almost always comes down to branded versus non-branded search not being separated. When branded search (people searching your exact brand name) is mixed into the same campaigns as non-branded prospecting, the high-ROAS branded traffic inflates your overall numbers. You think you are acquiring new customers efficiently, but in reality your non-branded ROAS may be far lower — sometimes unprofitable. Segmenting these properly reveals the true performance of your acquisition spend and lets you optimize each segment independently.
How do I know if my Google Ads conversion tracking is accurate?
Compare your Google Ads reported conversions against your source of truth — typically Shopify, WooCommerce, or your order management system — over a 30-day period. A discrepancy of more than 10-15% indicates a tracking problem. Common causes include double-counting from redundant conversion actions, counting page views or add-to-carts alongside purchases, not filtering out test transactions, and missing cross-domain tracking. Our full audit checklist covers all 50 tracking and account health checks.
Should I use the same bid strategy for all my Google Ads campaigns?
No. Smart Bidding strategies like Target ROAS and Target CPA need sufficient conversion data to work effectively — Google recommends at least 30 conversions in the last 30 days per campaign. Applying a Target ROAS of 400% to a campaign that gets 5 conversions a month means the algorithm has almost no data to optimize from, so it will either spend erratically or barely spend at all. Lower-volume campaigns should use Maximize Conversions or even Manual CPC until they accumulate enough data to graduate to a target-based strategy.
How often should I check Google Ads search term reports?
At minimum, review search term reports weekly for active search campaigns and every two weeks for Performance Max. For new campaigns or periods of heavy spend, daily reviews for the first two to four weeks are advisable. The goal is to catch irrelevant queries before they consume meaningful budget. Brands that neglect search terms for months routinely waste 15-25% of their spend on queries that will never convert — such as informational searches, competitor names, or completely unrelated terms.
What are custom labels in Google Shopping and why do they matter?
Custom labels are optional fields (custom_label_0 through custom_label_4) in your product feed that let you tag products with business-level attributes — such as margin tier, best-seller status, seasonal relevance, or price range. They matter because without them, Google treats all your products equally. A product with 65% margins and a product with 20% margins get the same bid treatment, meaning you are over-investing in low-margin items and under-investing in high-margin ones. Custom labels let you create separate campaigns or bidding rules for different product groups based on profitability, not just category.
How Many of These Are in Your Account?
We've seen these seven mistakes cost e-commerce brands anywhere from $5,000 to $50,000 per month in wasted spend, inflated metrics, or missed growth. The worst part is that most brands don't know — because the dashboard looks fine.
At BTB Media, we run a thorough, no-obligation audit of your Google Ads account. We'll tell you exactly what's working, what's broken, and what to fix first — with the data to back every recommendation.
Get Your Free Google Ads AuditNo commitment. No sales pitch. Just a clear picture of where your budget is going.
Continue Reading
How to Audit a Google Ads Account for E-Commerce
The complete step-by-step audit framework organized by revenue impact.
ChecklistGoogle Ads Audit Checklist: 50 Points That Actually Matter
A printable, actionable checklist covering tracking, feeds, PMax, bidding, and more.
Performance7 Signs Your Ad Account Is Being Mismanaged
Red flags in the data that a proper audit catches immediately.