Why Most Ecommerce Paid Advertising Fails, And What the Best Brands Do Differently

An expert roundup on the most common paid advertising mistakes in ecommerce, and the strategies that actually move the needle.

There’s a version of ecommerce paid advertising that looks great on paper: high impressions, decent click-through rates, a ROAS number your platform dashboard is proud to show… And yet, at the end of the month, the revenue doesn’t match the story the data seems to be telling.

If that sounds familiar, you’re not alone. We talk to ecommerce brands every week who are spending real money on Meta, Google, TikTok, and beyond… and are quietly wondering why it isn’t working the way they expected.

The honest answer is rarely about budget, or the platform. Most of the time, the problem lives somewhere between strategy and execution: in the assumptions brands bring to their campaigns, the shortcuts they take with targeting and creative, and the metrics they choose to trust.

To get a clearer picture, we asked some of the sharpest paid media minds working in ecommerce today to share the mistakes they see most often, and what brands should do instead. We’ve added our own perspective from working with clients like The Helm and Iris Setlakwe, where we’ve seen firsthand what separates campaigns that perform from campaigns that just spend.

Here’s what we found.

The Ecommerce advertising landscape right now

Paid advertising for ecommerce has never been more complex, or more consequential. Between iOS privacy changes, the rise of AI-driven campaign types like Performance Max and Advantage+, the growing influence of TikTok as a commerce channel, and an attribution landscape that grows murkier by the year, brands are making bigger decisions with less reliable data than ever before.

At the same time, the cost of getting it wrong has gone up. Ad costs have risen across every major platform. Competition is stiffer. And consumers, particularly younger ones, have developed a finely tuned radar for advertising that doesn’t feel native to the platform they’re on.

The brands pulling ahead aren’t necessarily the ones spending more. They’re the ones spending smarter, thanks to clearer strategies, tighter targeting, and a better understanding of what their numbers actually mean.

The core problem: trying to be everywhere, say everything, to everyone

Before we get into the tactical mistakes, it’s worth naming the mindset that causes most of them.

In our experience at Casa Media, the single most common thread running through underperforming ecommerce campaigns is a lack of specificity. Brands come to us having cast the widest possible net (broad audiences, vague messaging, campaigns spread thin across every available channel) and hoping that volume will compensate for precision.

It doesn’t.

What we’ve seen, both in our own client work and in the expert responses below, is that narrowing down almost always improves performance. A tighter audience, a clearer message, a more intentional creative:  these are the conditions under which good advertising actually works.

Svetlana Stotskaya, a global marketing executive and mentor at Techstars — named one of the top 100 retail technology influencers by Retail Technology Innovation Hub in both 2025 and 2026 — puts it plainly: “Many new brands pour big budgets into digital ads before truly identifying and validating their target audience segments. This often leads to wasting massive promotional resources on broad, untested campaigns that don’t convert.”

Her recommendation is to start small, test audience segments and creatives with modest budgets, and only scale what’s already proven to work. It’s not a complicated idea. But most brands skip it.

Mistake #1: Optimizing for ROAS without understanding what ROAS actually measures

This one comes up more than any other. And it’s worth spending time on, because the problem is subtle.

ROAS (return on ad spend) is the metric most ecommerce brands live and die by. It’s visible, it’s easy to report, and it feels like a direct measure of whether your advertising is working. The issue, as Rafael Sarim Oezdemir explains, is that ROAS has a serious blind spot.

Oezdemir is Head of Growth at EZContacts, a US eyewear retailer with over 5 million customers, and manages paid media budgets across Google, Meta, TikTok, and multiple retail media networks.

“ROAS is usually more favorable to the channels and campaigns that capitalize on existing customer interest,” he says. “Branded search and retargeting campaigns show great results. But branded search and retargeting aren’t actually delivering incremental revenue — they’re harvesting revenue already driven by upper-funnel activities.”

In other words: your best-performing campaigns by ROAS are often your least valuable in terms of actual growth. You’re paying to capture customers who were already going to buy.

Sébastien Godin, leader behind ROAS Secrets, makes the same case with a concrete example. If you spend $1,000 on ads at a 4:1 ROAS with a 25% margin, you make $0 in profit. Spend the same $1,000 at a 2.2:1 ROAS with a 50% margin, and you pocket $300. The campaign with the lower ROAS wins.

“In order to avoid this,” Godin says, “you should split campaigns by product margins so you can control the budget going to each type of product, and set target ROAS thresholds by margin.”

Chris Coussons, founder of Visionary Marketing, has seen this play out in practice with a homewares client running a single Performance Max campaign across 14,200 SKUs. The reported ROAS looked fine at 3.1x. But Google’s algorithm was quietly pouring budget into high-volume, low-margin products because those produced the cheapest conversions. Margin was bleeding while the dashboard looked healthy.

Coussons split the campaign into three asset groups by margin band, each with its own budget cap and ROAS target. Same total spend. After 11 weeks: blended ROAS of 4.6x, profit per pound spent up 71%, total revenue up 19%.

The lesson isn’t that ROAS is useless. It’s that ROAS without margin context is an incomplete picture, and making decisions based on an incomplete picture is how budgets disappear.

Mistake #2: Handing full control to the algorithm

Performance Max on Google. Advantage+ on Meta. These campaign types are powerful, and the platforms have done an excellent job of convincing advertisers to trust them completely.

That trust has limits.

“This strategy works great when your brand sells only one kind of low-margin but high-converting SKU,” Oezdemir explains. “For businesses with diversified offerings and multiple categories with different margins, the algorithm could spend the bulk of the allocated budget on best-selling items that would have been sold anyway.”

Godin is even more direct about the audience targeting side of this equation. Giving full control to AI over audience targeting often leads to overspending on low-incremental audiences and a loss of control between short-term profit generation and long-term business growth. His recommendation: separate acquisition and retention audiences into distinct campaigns, each with different ROAS targets based on predicted lifetime value. New customers, who have more remaining LTV, should generally receive higher investment.

Both experts recommend running incrementality tests (also called audience holdout tests or conversion lift studies) to understand how much of your ad spend is generating true added revenue versus capturing conversions that would have happened regardless.

This is something we pay close attention to at Casa Media. For The Helm, a premium Canadian menswear retailer, we made a deliberate shift away from traditional retargeting and toward a prospecting-focused strategy. Rather than spending heavily on a small pool of people who’d already visited the site, we invested in reaching new, high-potential audiences with precision targeting and strong creative. The result was a Meta CTR of 5.40% against an industry average of 1.24%, and a CPC of $0.21 against an industry range of $0.45–$1.07.

Our intentional structure helped us with this, not the algorithm. 

Mistake #3: Treating the product feed as an IT task

For brands running Google Shopping ads, Performance Max, or Meta’s Advantage+ Shopping campaigns, the product feed is the foundation everything else sits on. Most brands treat it like a backend export. The best brands treat it like marketing material.

“In practice, the task of updating product feeds is left to the PIM or development team,” Oezdemir notes. “A lot of SKUs feature titles such as ‘Brand Name Model Number Color Variant SKU-12847’ that don’t convert due to lack of keywords matching customer queries.”

His finding: writing category and use-case-driven titles, filling in every optional field, and filtering the feed by strategic priorities or profitability leads to a 20–40% CTR uplift. In the current environment, he argues, feed optimization matters more than bid strategies for most product catalogs.

Coussons saw a version of this problem with a fashion client whose 8,200-SKU feed was sending 28% of Advantage+ retargeting impressions to either out-of-stock items or products under £15. Average order value on retargeting was £41: well below the site’s AOV of £74. After adding feed-level exclusions for low stock, image quality, and margin band, the active feed shrank to 5,400 SKUs. AOV on retargeting moved to £69 in six weeks. Retargeting ROAS climbed 38% with no change in spend.

A clean, curated feed gives the algorithm better inputs, and better inputs produce better outputs. We know it sounds obvious, and yet, most brands still don’t do it.

Mistake #4: Misreading attribution data

Post-iOS 14, attribution has become significantly less reliable. Platforms use modeled conversions to fill in the gaps left by missing signal. View-through attribution inflates the credit Meta and Google claim for purchases. Last-click models in GA4 over-credit the bottom of the funnel.

The brands getting burned by this are the ones making major budget decisions based on platform dashboards alone.

Coussons watched a food brand cut Meta prospecting spend from £6,200 to £1,400 per month because last-click data showed a weak 1.8x ROAS. Three months later, total revenue had fallen 22% — and branded search volume, their supposedly best-performing channel, had dropped 31%. The “underperforming” prospecting was feeding the “high-performing” branded search all along.

“Relying on platforms’ data alone when managing a 7-digit budget,” Oezdemir says, “is as risky as flying with one instrument only.”

He argues that brands often chase “Green ROAS” in Meta or Google dashboards, but those metrics don’t translate into real contribution margin in the bank. He recommends shifting toward media mix modeling and incremental lift testing to measure true incrementality instead of relying on platform-claimed conversions.

The fix isn’t to stop using platform data. It’s to triangulate. Use geo holdout tests, compare platform numbers against your finance system, and watch branded search volume as a leading indicator of upper-funnel health.

Mistake #5: Sending paid traffic to pages that aren’t ready for it

Brands spend enormous energy optimizing their ads. They test headlines, audiences, creative formats, bid strategies. And then they send that hard-won traffic to a product page nobody has touched in months.

“A £1 click landing on a 1.8% conversion page costs £55 per order,” Coussons explains. “The same £1 click on a 3.2% conversion page costs £31. The cheapest media optimization is almost always a CRO fix, not a bid adjustment.”

His example: a home fragrance client paying £1.42 average CPC with a 1.6% conversion rate on paid landing pages. Three changes — a sticky add-to-cart button on mobile, a social proof block above the fold, and one FAQ section addressing objections — pushed the conversion rate to 2.7% in five weeks. Same media spend. Revenue per click up 69%.

Oezdemir frames this as a structural problem in the agency model: “Most advertisers ignore landing page optimization because the ownership falls outside the scope of ad agencies’ responsibility.” The ad and the landing page are treated as separate problems managed by separate teams. But the customer experiences them as a single, continuous moment. When that moment breaks down, no amount of ad optimization fixes it.

At Casa Media, we take this seriously. It’s one reason we consistently recommend that paid media be supported by a well-optimized Shopify experience. Our results with Iris Setlakwe, a Montréal womenswear brand, showed this clearly. In their first year working with us, ecommerce revenue grew by 146%, reaching the highest performance in the brand’s history. By 2025, that growth held and climbed further, with revenue up 16% over 2024 and Meta ROAS reaching 10.58. That kind of sustained performance requires the full ecosystem to work: ads, landing pages, and retention channels together.

Mistake #6: Retargeting everyone the same way

Retargeting is one of the most misused tools in ecommerce advertising. The logic makes sense on the surface: someone visited your site but didn’t buy, so you show them an ad. But in practice, most brands execute this badly.

The most common version of the mistake is treating all website visitors as a single audience. Someone who spent 12 seconds on a homepage and someone who added to cart and abandoned at checkout are not the same person. Showing them the same ad with the same message and the same creative is a waste of budget on the former and a missed opportunity with the latter.

Beyond segmentation, three other retargeting errors consistently drain budgets: failing to exclude existing customers from prospecting ads, not capping ad frequency, and running the same creative indefinitely. The result is an experience that feels less like relevant advertising and more like being followed around the internet by a brand that doesn’t know you.

At Casa Media, our retargeting approach is built around exclusion as much as inclusion. We only target visitors who haven’t purchased. We cap frequency to avoid fatigue. And we rotate creative so the experience stays fresh. The goal is to feel helpful, not relentless.

Godin adds another layer to this: people on social media aren’t there to shop. They’re there to be entertained. Cold retargeting audiences shouldn’t be hit with a hard sell. The messaging should meet them where they are, with a benefit or a result, not a product feature.

Mistake #7: Misunderstanding what TikTok is (and isn’t)

TikTok is growing as an ecommerce advertising platform, particularly for brands targeting younger audiences. But many brands bring the same playbook they use on Meta or Google and wonder why it doesn’t land.

Our view is direct: TikTok audiences don’t want to be advertised to. They want to be entertained, inspired, and spoken to as peers. The more your ad looks like an ad (corporate tone, high production value, product-focused messaging), the faster it gets scrolled past. Some users will actively filter ads from accounts that feel inauthentic, effectively removing you from their feed.

The content that performs on TikTok for ecommerce is native-feeling. It’s current, lightly playful, and unpretentious. If your brand can’t commit to that register, TikTok probably isn’t the right primary channel for paid investment. But if you can, and if your product has a natural fit with the platform’s culture, it can be genuinely effective.

The principle Godin articulates across all platforms applies here especially: match user intent with user experience. Someone in entertainment mode doesn’t want to be sold to. Start with the result your product delivers. Build the emotional connection. Then show the product as the vehicle that gets them there.

Our contrarian view: you don’t need a massive budget

Most conversations about paid advertising eventually come back to budget. Spend more, reach more, grow more.

We don’t buy it.

Our experience (that’s backed by the results we’ve delivered for brands like The Helm and Iris Setlakwe) is that a focused budget applied with precision consistently outperforms a large budget spread thin. A niche audience, a message that connects with something culturally relevant to them, and a seamless post-click experience can drive real revenue without the kind of spend most agencies will tell you is necessary.

This isn’t a pitch for spending less. It’s an argument for spending intentionally. The brands that win at paid advertising aren’t the ones throwing the most at it. They’re the ones who’ve done the harder work of getting specific: specific about who they’re talking to, what they’re saying, and what happens after someone clicks.

Reigns captures a version of this when he talks about the “single-touch conversion fallacy”: the assumption that cold audiences will convert immediately on a “Buy Now” CTA without any prior value exchange. His agency has seen a 30% conversion lift by leading with education and value before the sales pitch. The audience arrives at the conversion moment pre-qualified and trusting.

What good ecommerce paid advertising actually looks like

Across all the expert insights in this roundup, a few principles show up consistently:

1. Segment by margin, not just by product. Whether you’re structuring Google Shopping campaigns, Performance Max asset groups, or Meta catalog ads, margin should drive your budget allocation and ROAS targets. Revenue metrics without margin context are decorative.
   
2. Test incrementality, not just performance. Before you scale a channel or cut one, run a holdout test to understand whether your spend is generating new revenue or simply claiming credit for revenue that would have happened anyway.
   
3. Treat the feed as a marketing asset. For any catalog-based campaign, the product feed is where significant performance is either won or lost. Optimize titles for search intent, exclude low-margin and out-of-stock items, and update it regularly.
   
4. Separate acquisition from retention. New customers and returning customers have different LTV profiles and respond to different messages. Running them through the same campaigns muddles both objectives.
   
5. Fix the landing page before adding budget. A conversion rate improvement is almost always cheaper and faster than a CPC reduction. Audit your paid landing pages first.
   
6. Match the platform to the mindset. TikTok, Meta, and Google attract users in very different states of intent. The creative, tone, and offer need to reflect where the user is in that moment.

And underneath all of it: be specific. About the audience. About the message. About what success actually means for your business, not just your dashboard.

Working with a paid advertising agency that gets this

If you’re an ecommerce brand actively evaluating paid advertising partners, the questions worth asking are harder than “what’s your average ROAS?”

Ask how they think about margin. Ask how they handle attribution. Ask what their process looks like for landing page performance, product feed management, and creative testing. Ask whether they separate acquisition and retention audiences, and whether they run incrementality tests.

The answers will tell you a lot about whether an agency is managing your campaigns or actually thinking about your business.

At Casa Media, paid advertising is one part of a broader approach to ecommerce growth. We’ve helped a premium menswear retailer achieve a 5x+ ROAS while protecting brand equity. We’ve helped a womenswear brand turn years of declining online revenue into 146% growth in a single year, then build on that into a sustained new baseline. Neither result came from a big budget or a clever hack. Both came from strategic thinking, precise execution, and the patience to let good decisions compound.

If that’s the kind of partnership you’re looking for, we’d love to talk.