6 Google Ads mistakes that tank e-commerce campaigns (and what we do about them)

Most e-commerce brands I audit have one thing in common.

They're paying Google to close sales that were already happening.

It's rarely the brand's fault. The Google Ads interface has evolved to reward spend, not profit. Automation defaults favor volume. Dashboards highlight the numbers that feel good. And the people running the account, whether in-house marketers or agencies billing on retainer, don't always have skin in the game on whether the spend is actually incremental.

After 15+ years running paid media and auditing dozens of e-commerce accounts every year, the same six mistakes come up in nearly every pre-engagement audit. None of them are obscure. None of them require Google certifications to understand. But together, they quietly cost most DTC brands somewhere between 20% and 50% of their paid media budget.

Here's what to look for in your own account, and what to do about each one.

01 Treating Google like a retention channel

The mistake: You launch a Performance Max campaign. Early ROAS looks incredible. 6x, 8x, sometimes higher. Leadership is thrilled. A few months in, revenue plateaus. Nobody can figure out why.

The answer is almost always the same. Your PMax campaign is eating branded search, retargeting, and cart abandoners. Demand that was going to convert whether you paid for it or not. Strip those out and net-new customer acquisition is flat. You're not acquiring. You're taxing existing demand.

What to do instead: Separate your acquisition and retention plays at the structural level.

• Branded search gets its own campaign, with its own modest budget. Defend the brand, don't inflate it.
• Build dedicated Shopping campaigns targeting unbranded, high-intent keywords. This is where real acquisition happens.
• Configure PMax with audience signals and budget caps that prevent it from defaulting to the lowest-hanging fruit.

Google's reach into new audiences is enormous. Treating it as a closing channel leaves most of that opportunity on the table.

02 Collapsing upper and lower funnel into one campaign

The mistake: Someone searching "running shoes" and someone searching "buy Nike Pegasus 41 size 10" do not want the same thing. Top-of-funnel searchers are exploring. Bottom-of-funnel searchers have their credit card out. Lumping them into one campaign forces Google to average out bids, ad copy, and landing pages across wildly different intents.

The result: wasted budget on traffic that was never going to convert, and lost conversions from traffic that was.

What to do instead: Separate upper-funnel and lower-funnel campaigns at the structural level.

• Build dedicated campaigns for broad, research-phase keywords with lower bids and educational landing pages.
• Run separate campaigns for purchase-intent keywords with aggressive bids and direct-response pages.
• Match ad copy to the funnel stage. "Discover the new Nike Pegasus" for the top. "Nike Pegasus 41, free shipping, ships today" for the bottom.

03 Over-segmenting the account structure

The mistake: Instinct tells detail-oriented advertisers to segment everything. One campaign per category, one ad group per keyword, separate budgets for every audience. It feels like precision. It's actually the opposite.

Google's Smart Bidding needs conversion volume to learn. Spread your budget across ten underfunded campaigns and none of them get enough data to stabilize. You've built the appearance of control while eliminating the algorithm's ability to optimize.

What to do instead: Consolidate toward fewer, better-funded campaigns.

• Three well-funded campaigns with clear goals will outperform ten fragmented ones almost every time.
• Let Smart Bidding breathe. Give each campaign enough daily budget to generate at least 30-50 conversions per month.
• Reserve segmentation for scenarios where the bidding economics genuinely differ. Like separating high-margin from low-margin product lines, or separating acquisition from retention.

04 Neglecting the product feed

The mistake: The product feed is the single most leveraged asset in any Shopping or Performance Max account. It's also the least maintained. Weak titles, missing attributes, poor categorization, stale pricing, no GTINs on half the SKUs. These are silent budget killers that compound month after month.

The work of fixing a feed is unglamorous. No one posts about it on LinkedIn. But a well-optimized feed consistently outperforms a neglected one by 30-60% in impression share and CTR, and those gains hit the bottom line directly.

What to do instead: Treat the feed as a living asset, not a one-time setup.

• Audit product titles monthly. Front-load the brand, product type, and key attributes.
• Fill every optional attribute Google allows. More signals mean better matching.
• Monitor disapproval rates weekly. A 5% disapproval rate kills more revenue than most brands realize.
• Use feed optimization tools (DataFeedWatch, Feedonomics, GoDataFeed) if you have the SKU volume to justify them.

05 Running Max Conversion Value with no ROAS target

The mistake: Max Conversion Value tells Google: "spend my budget finding conversions, whatever it costs." Left without a ROAS target, the algorithm will happily optimize for volume at the expense of profitability. Conversion value goes up. Margin goes down.

The dashboard will look healthy. The P&L will not.

What to do instead: Set a realistic tROAS once you have enough conversion data to justify it.

• Wait until the campaign has accumulated at least 30 conversions before setting a target.
• Start with a target that matches your current account-level ROAS, then optimize up or down gradually.
• Never set an aggressive tROAS on a cold campaign. You'll starve it of traffic before it has a chance to learn.

A ROAS target is the guardrail that turns Google's optimization from "find conversions" into "find conversions that make us money."

06 Killing campaigns before the learning phase finishes

The mistake: You launch a new campaign on a small budget. CPAs come in high. Data is inconclusive. You pull the plug at day 7 and decide "Google doesn't work for us."

The campaign never had what it needed to succeed. Smart Bidding requires roughly 30-50 conversion events in a tight window to stabilize. An underfunded campaign stays in learning indefinitely, producing volatile results that look like failure but are actually just incomplete data.

What to do instead: Commit to adequate funding from day one.

• Calculate the budget needed to generate 30-50 conversions in 2-3 weeks based on your expected CPA.
• If that number is more than you can afford, run fewer campaigns at higher budgets rather than spreading thin.
• Resist the temptation to intervene in the first 14 days. Every change resets the learning phase.

Patience is expensive in the short term and enormously valuable over 90 days.

What ties these six mistakes together

None of these are rookie errors. I see them in accounts being managed by experienced in-house teams and well-known agencies. The reason they persist isn't incompetence. It's incentive misalignment.

When your agency bills on a retainer, pointing out that 40% of your spend is being wasted is bad for their revenue. When an in-house marketer's bonus is tied to total conversion volume, there's no reason to question whether those conversions are incremental. When a platform's defaults are set to maximize spend, the path of least resistance leads directly to waste.

This is why Digital Shepherds is built on pay-per-performance. When we only get paid on the revenue we actually generate, not on media spend, not on hours, not on retainers, these six mistakes become existential for us, not just for you. We catch them because we have to.

Yossi Levi is the founder of Digital Shepherds, a pay-per-performance growth agency working with e-commerce and professional services brands. We only get paid when we generate revenue. Everything else is just noise.