Ecommerce is the hardest category on Meta. A plumber competes with a handful of local firms. A Shopify store selling a skincare serum competes with every well-funded brand on the continent, all bidding with the same objective, the same placements and increasingly the same creative playbook.
That has one useful consequence. Your rivals are sophisticated, national and spending real money, and every ad they run sits in a public library. You are not short of teachers. You are short of a way to tell which of their ads work.
Why generic ecommerce Facebook ads advice fails
Most advice on ecommerce Facebook ads treats the click as the goal. Here the goal is a click that pays for itself, a different problem entirely.
Unit economics decide everything. A store with a low average order value and a thin margin has almost no room to acquire a customer before the first order is underwater. A store selling a high-value bundle has many times that room. The identical ad, at the identical cost per click, is a winner in one and a slow bleed in the other. Nobody who does not know your AOV, margin and repeat rate can tell you what a good CPC is.
Three more constraints make generic advice worse here than in local services:
- Creative volume is the real lever. Targeting is largely automated now. What is left is how many distinct angles you test per month, and how fast you read the results.
- Seasonality is severe. Peak retail compresses a large share of annual demand into a few weeks, so an ad that worked in a quiet month is suddenly bidding against every rival's biggest budget of the year.
- The creator format dominates. In most consumer categories, ads that look like a person talking to a camera do the heavy lifting.
So the useful question is not "what makes a good ecommerce ad." It is "which ads in my category survived long enough that somebody with a P&L kept paying for them."
What you can actually see in the Meta Ad Library, free
The Meta Ad Library is public by law, so researching it is legal and anonymous. The advertiser is not notified. For free it shows every ad a competitor has live, every variant of each creative, the full copy, the start date on each ad, and the landing page it points to, which reveals whether they send traffic to a collection, a product page or a dedicated offer.
Be equally precise about what it never shows you.
| What you can verify from outside | What nobody outside can see |
|---|---|
| Which ads are live right now | CTR, CPC, CPM |
| How long each ad has been running | Spend and budget |
| Creative, copy, format and offer | ROAS and conversion count |
| Landing page and funnel shape | Audience and targeting setup |
Everything in the right column lives inside the advertiser's own account. Any tool claiming to show a rival's exact ROAS invented the number. That absence is exactly why longevity matters.
Longevity is the only performance signal an outsider can verify
Advertisers kill ads that lose money and keep ads that make money. That is the entire mechanism.
So an ad live for four months survived four months of somebody with budget on the line deciding, repeatedly, not to switch it off. In a category this heavily measured, that is close to a verified result. The newest, glossiest ad has passed no such test, and it is the one the library surfaces first.
The catch is that the Ad Library has no sort by run time, no longevity column and no filter for "live over 90 days". Each card buries a start date in its detail panel, so finding the oldest survivor by hand means opening every ad and doing the arithmetic yourself. For a brand with a few hundred live ads that is an afternoon, and it is stale a week later, because Meta keeps no run-time history.
This is the gap AdWhispr fills: it ingests a brand's entire Meta ad library and re-snapshots it daily, so it holds the run-time history Meta's own API does not return. The mechanics are in how to find a brand's longest-running Meta ad, the reasoning in days running as a performance proxy. You ask it the way you would ask a colleague:
Show me every ad this brand has had live over 90 days, sorted by days running, and tell me what those ads have in common.
Then the follow-up most people never ask:
How many of their ads are under 30 days, 30 to 90, and over 90 days?
That distribution is the strategy. A dozen ads past 90 days means a durable formula they are scaling. An oldest ad of three weeks means they are still hunting, and the category angle is still open.
The patterns that repeat in durable ecommerce ads
Once you have long-runners from several competitors, stop seeing ads and start seeing structures. The durable ones fall into a small number of shapes.
| Pattern | Structural shape | Why it survives |
|---|---|---|
| Creator monologue | Phone camera, problem stated aloud before the product appears | Reads as opinion, not advertising, so it earns the first three seconds |
| Before and after | Visual state change in the opening frame, product explained after | Proof arrives before the pitch, so scepticism has nothing to grab |
| Objection first | Names the reason people do not buy, then answers it | Filters unqualified clicks, protecting the unit economics |
| Founder explainer | Direct to camera, why the product exists, usually unpolished | Origin credibility a studio spot cannot buy |
| Bundle framing | Leads with the multi-item offer, not the hero SKU | Lifts AOV, which often makes the maths work |
Notice what is not on that list: production value. What repeats is the order of information, not the budget.
Read the offers structurally too. Whether a category is dominated by a percentage discount, a free gift, a bundle or free returns tells you what its buyers are anxious about. If every long-runner leads with free returns, the objection is fit, not price.
Seasonality changes which ads count as proven
Peak retail distorts the library. Ahead of a major shopping season most brands flood their account with offer-led creative, much of which disappears afterwards. An ad that ran six weeks across a peak is not the same evidence as one that ran six months through quiet trading and a peak.
So when you research during or just after a peak, treat only the evergreen survivors as proven. And build your peak creative from angles that already worked off-peak, because peak is the worst possible time to find a new angle at full auction prices. Daily snapshots make that checkable: ask which ads were live before the season, which appeared for it, and which are still live.
Turning a proven pattern into your own ad
Research that ends in a screenshot folder does nothing. The output should be creative in your account.
The honest version is not copying a rival's ad. It is taking the structure that survived, the hook shape, the format, the order the proof arrives in, the offer mechanic, then rebuilding it with your product and your evidence. AdWhispr classifies every ad by hook, format, tone and offer so that pattern is explicit, and can clone a proven ad for your brand as an image or video with original copy grounded in the real winner.
Clone the structure of their longest-running creator ad for my product, keep the objection-first opening, and swap the offer to a two-pack bundle.
The walkthrough is in cloning a competitor image ad, and if you have never run a research pass, start with your first competitor research run. Work in the web app, or from Claude, ChatGPT, Cursor and Claude Code over MCP.
The free plan covers one tracked brand, no credit card. Enough to test whether the patterns you assumed held in your category actually do.
FAQ
Why do UGC and creator ads work so well for DTC brands?
They earn the first few seconds by reading as a person's opinion rather than an advertisement, and they deliver proof before the pitch. That is a structural advantage, not a production one, which is why unpolished creator ads often outlast expensive studio work.
What is a good cost per click for ecommerce Facebook ads?
There is no universal answer. What matters is cost per click relative to your average order value, margin and repeat purchase rate. The same CPC is comfortable for a high-value bundle and fatal for a low-margin single item. Work backwards from what you can afford to pay for a customer.
Can I see how much a competitor is spending on Facebook ads?
No. Spend, budget, ROAS, CTR and conversion counts sit inside the advertiser's own account and are never published. What you can verify is which ads are live and how long each has run, which is why longevity is the signal worth building on.
How many Facebook ads should a Shopify store be testing?
There is no fixed number, but creative volume is the lever that still responds to effort now that targeting is automated. Aim at a cadence rather than a count: enough new angles per month that you always have something running against your best ad.