Instagram ads for clothing brands

Every other guide in this cluster has to argue for Instagram. This one does not, and that is the problem. Apparel is native here: the platform's whole visual grammar was built by and for clothing, which means your ad is not competing for attention against other ads, it is competing against content in an identical idiom — a competitor's campaign, a creator's haul, and a photograph a friend posted this morning, all with the same light and the same crop. In that context a beautifully shot garment is not a strong ad, it is camouflage. Indistinguishable from organic means indistinguishable, and a paid impression that reads as organic content is one you bought for nothing.

So the working rule for this platform is that the ad has to make a claim the surrounding content is not making. Look at what the ten ads on the Browse page actually do: one argues that a size chart lied about inseam, one puts six reviews side by side labelled XS to 4XL, one prices a pair of jeans against the four pairs it replaces, one states what organic cotton costs and where the rest of the markup goes. Not one of them is a product photograph doing product-photograph work. That is the platform tax, and it is worth paying because the alternative is invisible.

The shape half is short and mostly familiar from the other Instagram pages: 4:5 for feed, full-screen 9:16 for Stories and Reels — eight of the ten ads are 4:5 and two are 9:16 — and no 1.91:1, because that ratio routes to Facebook's right-hand column and search results and our publisher gives it no Instagram position. (Instagram Feed itself accepts up to 1.91:1 — corrected 7 September 2026; what this app lacks is the surface, not the shape.) For apparel the consequence is specific: the shapes that matter beyond those two are catalogue formats rather than aspect ratios, which is a campaign-type decision and is covered below rather than here.

The last Instagram-specific point is the one most brands get backwards. The organic playbook on this platform — post often, ride the trend, let the garment speak — is a genuinely good playbook, and it is the wrong one for paid. Organic reach is granted to content that fits the feed; paid reach is bought, and what you are buying is a chance to say something the feed will not say for you. Brands whose paid creative is their best-performing organic post, boosted, are the ones who conclude Instagram ads do not work.

Sales, and this is the only cluster on this site where that sentence is true. Six of the ten ads on the Browse page were written for it, three for Traffic and one for Leads — against dentist's seven Leads out of ten and real estate's eight. The reason is not that apparel advertisers are more ambitious; it is that a purchase can complete inside the session the ad started, so Sales is the objective describing the thing that can actually happen. Buying Leads for a product with a checkout is asking the delivery system to optimise for a lesser outcome and then being surprised it delivered one.

The three Traffic ads on that page are all doing the same job, and it is worth recognising: they argue something the reader has to read before they could want the product — what 17.5-micron merino actually is, what organic cotton costs, how a shoulder seam should sit. Those are the ads that need a page, not a checkout, and Traffic is honest about it.

The single Leads ad is the most instructive thing on the page, because it went wrong in a way that teaches the rule. The sneaker drop was pinned as a Leads concept — build a list before the drop — and the copy came back saying "no raffle, no waitlist… just a countdown". Objective is recorded metadata here and is never fed to the model, so the model followed the angle to its natural conclusion: first come, first served, live now. The ad is fine. The pairing is not, and you cannot run that copy against a signup form, because it has just told the reader there is no list. Match the objective to the mechanic the offer actually has, not to the stage of the funnel you wish it were at.

Apparel has no service menu, so the unit of planning here is not an offer but an argument — which claim about the product, aimed at which body or which habit. What the ten Browse ads have in common is instructive: eight of them are about fit, sizing, durability or cost, and none is about how the garment looks. Looks are what the photograph is for; the copy exists to answer the two questions that stop a purchase completing, which in apparel are always "will it fit me" and "what happens if it doesn't".

Each row pins four decisions: who it is for, the angle family that argues it (each is a page in the angle library), the objective to buy it under, and the finished ad that shows it assembled. The two rows with no example are the families this batch skipped, and one of them — the founder's story — is the single most over-used and worst-executed argument in this industry, which is why it is worth writing carefully rather than not at all.

One recurring shape worth naming before the table: the strongest ads here narrow the audience by body rather than by taste. Tall women. Petite professionals. Big and tall. Lifters who have been let down at 2XL. That is a real, unglamorous segmentation the rest of the category ignores, and it is why those ads can say something a general apparel ad structurally cannot.

This is the one industry in the cluster where the first decision is not who to target but which kind of campaign to run, because Meta offers a trade that suits apparel better than any other business here: the purchase happens in the ad's own session, so there is a purchase signal worth handing the targeting to. An Advantage+ sales campaign hands the targeting to the delivery system in exchange for that signal doing the work instead. It is a Sales campaign that Meta marks as Advantage+ while its budget, audience and placements stay inside Meta's limits. In Ads Manager that means Advantage+ audience, Advantage+ placements and, whenever there is more than one ad set, a campaign budget; Meta's developer guide asks for a campaign budget in every case, so use one. While all three hold, Advantage+ shows as On at the top right.

It allows multiple ad sets, provided every one of them qualifies. Exclude a placement in any of them, or opt out of mobile or desktop devices, and Advantage+ turns off for the whole campaign. So an Advantage+ sales campaign cannot be Instagram-only: its ads can run on Facebook's placements too, including the ones built for a 1.91:1 image, so export that shape as well.

Under Advantage+ audience Meta names four things it never widens: location, a minimum age (which you can set no higher than 25), language and custom-audience exclusions. Most of the rest becomes a suggestion that Meta can deliver beyond: an age range above the minimum, gender, lookalikes, interests, and any custom audience you include unless you turn on "Further limit the reach of your ads". Beyond those four, and an included audience with that setting on, what you put in the picker guides delivery rather than drawing its boundary.

That trade is worth reading against the housing guide in this cluster. A real estate agent loses age, ZIP codes, lookalikes and exclusions by regulation and gets nothing back. An apparel brand keeps a minimum age and its custom-audience exclusions, hands over lookalikes and most of the rest of the picker voluntarily, and gets a system optimising against actual purchases in return. The trade is only good if the signal is good: it is bad if your purchase events are broken, because then you have given away the targeting and received noise. And if you run catalog ads in the campaign, the catalogue has to line up with those events too.

So: catalogue and pixel before campaign type. Advantage+ catalog ads need a product catalogue and a pixel firing ViewContent, AddToCart and Purchase, and each of those events has to carry either content_ids or contents, with IDs that match the ones in your catalogue. That last clause is where most implementations quietly fail — the events fire, the dashboard fills in, and the product IDs do not line up, so the dynamic half of the system has nothing to be dynamic about. Check that the IDs match before you check anything else.

If you have read about these campaigns before, including on this page, three of their old rules no longer apply. Its predecessor, the Advantage+ shopping campaign (ASC), allowed exactly one ad set, nothing in its targeting but geography, and an optional cap on spend to existing customers. (This page described all three as current until 19 September 2026, seven months after version 25.0 of Meta's API stopped accepting new ones; corrected on that date.) Meta's developer guide says version 26.0, released on 29 July 2026, blocks edits to the rest and pauses the ones that used the cap. If you still have one, the FAQ below covers what to do.

The cap was the difference between a campaign that grows a brand and one that flatters it, and Meta has retired it: its help centre says the existing customer budget cap "is no longer available". The problem it solved has not gone anywhere. The two objections this page keeps coming back to, will it fit and what happens if it doesn't, are ones a past customer has mostly answered: they know their size in your cut and have already risked your returns once. That makes them the easiest purchases in the room, and left alone a Sales campaign can spend a large part of your acquisition money re-selling to people who bought a fortnight ago. Your ROAS will look excellent and your customer count will not move.

Meta's help centre gives two replacements. The first is an exclusion: on every acquisition ad set, under Audience, open "Show more controls" and add your customers as a custom-audience exclusion. Custom-audience exclusions are one of the four things Advantage+ audience never widens, and Meta's setup guide says using "Show more controls" "will not turn Advantage+ off immediately", so the campaign should spend next to nothing on anyone matched into those audiences. The second puts the campaign's ad sets under spending limits, each a daily or lifetime maximum: one ad set includes your customers with "Use as a suggestion" unticked, which Meta says may turn Advantage+ off, and a copy of it excludes them. Neither of those Meta pages says what a campaign loses when that happens. A spending limit is a maximum in money, where the old cap was a maximum percentage that moved with the budget, so the limit has to be changed by hand, and changing it can count as a significant edit.

Use the first, and buy repeat orders in a customer campaign of their own, as in the funnel below. They want their own creative, such as the drop, the reorder and the complementary item, and a budget you set for them rather than a ceiling inside the acquisition campaign.

Two details decide whether the exclusion works. Build it from your customer list and a website custom audience of purchasers together, and refresh the list, because an upload holds only the customers you had that day. Exclude exactly the people your customer campaign targets, so nobody falls between the two. And adding it is a targeting change, which restarts the ad set's learning phase: do it once, with the next batch of ads rather than in the week before a drop, then redo the budget arithmetic below on your cost per purchase from new customers, which is likely to be higher than the blended figure you had.

Then check it from both sides. Meta's own advice for Advantage+ sales campaigns is to build audience segments, which break a campaign's results down into new, engaged and existing audiences; with your customers defined there, an acquisition campaign should show an existing-customer share close to zero. That shows Meta honoured the audiences you gave it, not that they hold every customer, so check the other side in your store as well: orders from the campaign's tagged links placed by people already in your order history.

For everything outside an Advantage+ sales campaign — the Traffic ads, the arguments, the top of the funnel — the whole picker is available, because apparel is not a special ad category, though a Sales ad set still treats an included custom audience as a suggestion unless you turn on "Further limit the reach of your ads". The recommendation is the same as the SaaS page's, for a related reason: go broad and let the creative do the narrowing, because the segmentation that actually matters here is by body and habit rather than by interest. No picker has ever offered "women who are five foot ten", and under Advantage+ audience even "women" is only a suggestion.

The product page, and specifically the page for the item in the ad — not the homepage, not a collection, not a filtered grid. This is the least ambiguous destination recommendation in the whole cluster and the most frequently ignored, because collection pages feel generous. They are not: an ad that argued about the inseam on one pair of jeans, landing on forty pairs of jeans, has handed the reader the exact search problem the ad just solved for them.

Catalog ads solve this by construction — the ad is generated from the catalogue entry and the click goes to that item — which is one of the underrated reasons to get the catalogue right. For hand-built ads it is a discipline you have to keep yourself.

There is no lead-form debate to have here, which is unusual for this cluster. A form between a reader and a checkout is a step that costs a sale and gains an email address. The single exception is a real list mechanic — a waitlist for a made-to-order run, a restock alert on a size that is genuinely out — and the test is whether the thing you are collecting an address for actually exists. The Browse page's drop ad is the cautionary version: its copy promises there is no waitlist, so a form behind it would be contradicting the ad that produced the click.

One thing to put on the destination page rather than in the ad: the returns policy and the size guide, both above the fold on mobile. Those are the two questions that stop the purchase, and the ad can only gesture at them.

The rule that matters most here is the one that sounds wrong: do not make the ad beautiful. Or rather, do not let beauty be the only thing it does, because the feed you are buying into is already full of beautifully photographed clothes that cost nothing to post. An ad has to make a claim, and the claim is what buys the second of attention that the photograph then converts. Every one of the ten ads on the Browse page argues something, and it is not a coincidence that eight of the arguments are about fit, sizing, durability or price.

The second thing is a format worth stealing wholesale, and the to-do for this cluster was right to call it out. A Reviews/Ratings ad normally asserts volume — 4.9 stars, twelve thousand reviews — and volume is wallpaper. The version on the Browse page instead shows six review cards, one per size from XS to 4XL, under a headline naming the pattern. The structure does the arguing: it answers "is it good in my size", which is the actual anxiety, rather than "is it good", which nobody was asking. Any brand with a real size range and real reviews can build that ad this week, and almost none of them have.

Which leads directly to the hazard. That same ad is the one on the Browse page that must not be run as generated, because Jordan R., Morgan T. and the rest do not exist and their verified purchases never happened. A review grid works because readers assume real customers are behind it — that assumption is the mechanism — so filling it with invented ones is fabricated social proof, which is a policy violation on Meta and a legal problem in most markets. Take the format. Fill it with people who said it and agreed to appear.

Two structures, and which one you are running depends on whether your purchase signal is clean. If it is, an Advantage+ sales campaign is most of the funnel by design — usually one ad set, your customers excluded, the system doing the rest — and the manual work sits either side of it: the argument ads that create demand at the top, and the customer programme at the bottom. Run those as campaigns of their own, so their audience settings cannot switch the Advantage+ campaign off. If it is not, run the manual three-stage version below and fix the events in parallel, because handing targeting to a system fed by a broken purchase signal is worse than keeping it.

Either way the third stage is where apparel margin actually lives. Repeat purchase in this category is the difference between a business and a hobby, and it is the cheapest revenue in the account — which is also why the customer exclusion in the targeting section matters so much. You want to buy repeat purchases deliberately, in their own stage, rather than have an acquisition campaign quietly buy them for you and report them as growth.

Start with the learning-phase arithmetic, which lands somewhere between the two extremes in this cluster. Meta says an ad set usually leaves the learning phase after about 50 results in the week after its last significant edit. For a dental practice, 50 booked implant consultations a week is fantasy; for a software company, 50 trial starts is routine. For an apparel brand it depends entirely on your average order value: 50 purchases a week is achievable at €40 and hard at €400. Multiply your own cost per purchase from new customers by 50 and you have roughly the weekly spend at which one ad set can be expected to leave the learning phase. Each extra ad set needs that spend again, so unless your budget covers it once per ad set, keep your Advantage+ sales campaign to one: one ad set puts every purchase into a single pile.

Resist the obvious workaround. Optimising for AddToCart because purchases are too thin is the same mistake the SaaS page warns about with trial starts, in a different costume: the system will find you people who add to cart, which is a different population from people who complete a checkout, and the gap between those two groups is exactly where apparel loses money. Use the shallower event only as a temporary measure while volume builds, and know that you are doing it.

Then leave it alone. Changing an ad set's targeting, creative or optimisation event, adding a new ad, changing its bid strategy, or pausing it for seven days or longer is a significant edit that restarts the learning phase. Changes to the budget, a spending limit or a bid or cost goal can be one too, depending on size. In a business with a drop calendar the temptation to intervene daily is stronger than in any other industry here, and the calendar itself is a trap: every drop brings new ads, and each one added to a running ad set is a significant edit. Add a drop's ads in one batch, not as each is finished (if you publish them paused, switch them on together), and change the customer campaign's budget for the drop in the same sitting.

Finally, the number Ads Manager cannot show you. Returns in apparel arrive weeks after the sale, they are concentrated in exactly the categories where fit is uncertain, and they do not appear in reported ROAS. A campaign that looks profitable in the dashboard and unprofitable in the warehouse is the normal case rather than an edge case, so reconcile against net revenue on a lag before you scale anything — and treat a high return rate on a specific ad as a creative problem, because it usually is one: an ad that oversold the fit produced the return.

The first two are the reason most apparel accounts underperform on this platform specifically, and they are the same mistake: bringing the organic playbook to a paid slot.