Meta Ads are the sponsored posts, Stories and Reels you see on Facebook and Instagram. They are bought through one tool, Meta Ads Manager, and they show up wherever Meta has an audience: the Facebook feed, Instagram feed, Stories, Reels, Messenger and a network of partner apps. Unlike a Google search ad, nobody asked to see them. The ad interrupts a scroll and has about a second to earn attention.

They concern creators and brands whose product is visual and can be understood in a glance. A skincare line, a print shop, a course with a strong before-and-after, a handmade bag: all of these are Meta-native. If your product needs a paragraph of explanation before someone cares, Meta is harder, though not impossible.

What is Meta Ads?

Meta Ads is the advertising system for every Meta property. You create a campaign with an objective (sales, leads, traffic, awareness), define an audience, upload creative (images, videos, carousels), set a budget and let Meta's delivery system decide who sees the ad, when and at what price. Billing is mostly on impressions, expressed as a cost per mille, even when you optimize for purchases.

The structure has three levels:

  • Campaign: the objective and, with Advantage+ budgets, the daily spend.
  • Ad set: the audience, the placements, the schedule and the optimization event (purchase, add to cart, lead).
  • Ad: the creative itself, the copy, the headline and the destination URL.

Targeting used to be the main lever: interests, lookalikes of your buyers, age and location. Since 2021 and Apple's tracking changes, Meta pushes broad targeting and lets the creative do the selecting. Today the practical rule is that the ad is the targeting: a video about pottery aprons finds potters on its own.

What Meta Ads is not: it is not the same as boosting a post from the Instagram app. Boosting is a simplified front end with fewer objectives and no proper purchase optimization. Serious selling happens in Ads Manager with a Pixel or the Conversions API installed on your store.

Related vocabulary: Pixel (the script that reports views, add-to-carts and purchases from your site), Conversions API (the server-side version of the same reporting), lookalike (an audience built to resemble your existing customers), frequency (how many times the same person saw the ad) and ROAS (return on ad spend).

Why it matters

Meta is the largest source of cold, visual, scroll-based demand for small sellers, and the numbers behind it are simple enough to model on a napkin.

Say a brand sells a $45 candle set with a 60% margin, leaving $27 per order before ads. A campaign buys impressions at a CPM of $14. The ad gets a 1.4% click-through rate, so 1,000 impressions produce 14 clicks, and each click costs $1. The product page converts cold Meta traffic at 1.8%. A sale therefore needs 56 clicks and costs $56 in ads. That loses $29 per order.

Now change two things. A better hook video pushes click-through to 2.2%, so a click costs $0.64. Adding a second candle as an offer lifts the average order to $62, with $37 of margin. A sale still needs 56 clicks but costs $36, and the order leaves $1 of profit. Add one email a week later that brings back 15% of buyers for a second purchase, and the campaign becomes worth running.

That is what Meta Ads matters for: it forces you to look at creative, offer and average order together. No single one saves a campaign.

How it works

Here is the path from an empty account to a first sale-optimized campaign:

  • Install tracking. Add the Meta Pixel to your store and, if possible, the Conversions API. Verify that "Purchase" fires on your order confirmation page with the right value. Without this, Meta optimizes for clicks or views and you fly blind.
  • Choose the Sales objective. Optimize for Purchase from day one, even at low volume. Optimizing for add-to-cart or traffic brings people who add and leave.
  • Start broad. One ad set, your country, ages 18 to 65, no interests. Let the delivery system learn from the creative and the Pixel data.
  • Upload three to five creatives. A short video (6 to 15 seconds, hook in the first second), a carousel of product angles, a plain photo with a price, a customer quote. Meta shifts spend to whichever wins.
  • Set a daily budget of at least $15 to $30. Meta's learning phase looks for about 50 purchases in a week. Below that, results stay noisy, but a small store can still read the direction.
  • Send clicks to the right page. A product page or a dedicated landing page, never the home page. For a creator, the link-in-bio page works if the product is above the fold.
  • Read results after 7 days, not 2. Look at cost per purchase, ROAS and frequency. Kill creatives with high spend and no purchases, duplicate winners with a slightly higher budget.

Retargeting sits on top of this. Anyone who visited your product page and did not buy can be shown a follow-up ad for the next 7 to 30 days, usually at a much lower cost per purchase. See retargeting ads.

Benchmarks and examples

Costs move with the season, the country and how many other advertisers want the same people. Q4 is 30% to 60% more expensive than spring.

  • CPM: $8 to $20 in the US and Western Europe for broad audiences, higher for retargeting because those audiences are small.
  • Click-through rate: 0.9% to 2% on feed and Reels. Below 0.7% the creative is not working.
  • Cost per click: $0.50 to $1.50 for most consumer goods.
  • Conversion rate from cold Meta traffic: 1% to 2.5%. Retargeting traffic converts at 3% to 6%.
  • ROAS: 1.5 to 2.5 on cold traffic for a first month is normal. Established brands with strong creative and repeat purchases run at 3 to 5.

Three typical cases:

A creator launching a $120 online course runs Reels ads to a free 20-minute workshop. Cost per registration is $4, and 6% of registrants buy. A sale costs $67, leaving $53 on a product with almost no cost of delivery. The creator adds an email sequence and pushes the buy rate to 9%, dropping the cost per sale to $44.

A ceramics brand with a $58 average order runs one broad ad set at $25 a day. Week one: 4 sales, ROAS 1.1. Week three, after replacing the product-only image with a 10-second video of the glaze being poured: 11 sales, ROAS 2.6. Same budget, same audience, different creative.

A small store selling $22 socks cannot make cold ads work: CPM $12, conversion 1.5%, and a sale costs $30 on a product with $11 of margin. They switch to a $39 three-pack as the advertised product and the sale now costs $32 against $21 of margin. Still thin, but positive with the second order.

Common mistakes

  • Optimizing for the wrong event. Traffic and engagement objectives fill the top of the funnel with people who never buy. If you sell, optimize for Purchase.
  • Testing audiences instead of creatives. In 2026, ten interest-based ad sets with one ad each is backwards. Run one broad ad set with five ads.
  • Changing things every day. Each edit resets learning. Let an ad set run 5 to 7 days before judging.
  • Ignoring frequency. When the same person has seen your ad 4 or 5 times in a week, click-through drops and costs rise. Refresh creative or widen the audience.
  • Trusting Meta's reported ROAS alone. Attribution windows count purchases that might have happened anyway. Cross-check with your store's orders and a campaign discount code.

Best practices

  • Hook in the first second. Show the product doing its thing before anything else. No logo intro, no "hi guys".
  • Make the ad look like content. Vertical 9:16, phone-shot, subtitles burned in. Polished studio ads often lose to a customer unboxing on Reels.
  • Show the price in the creative. It filters clicks. Fewer clicks at a higher conversion rate beats many cheap clicks that bounce.
  • Pair every campaign with a discount code. A code like REEL10 shown in the ad makes attribution honest and gives a reason to buy now.
  • Bring buyers into email. Meta traffic is rented. Every order should land on your list so the second sale costs $0 in ads. See email marketing.
  • Feed the algorithm with clean purchase data. Correct order values, currency and event deduplication between Pixel and Conversions API. Bad data teaches Meta the wrong customer.
  • Rotate creative every 2 to 3 weeks. Keep the winning angle, change the footage. Fatigue is the most common reason a good campaign stops working.

In Roctify

Roctify is not an ad platform, but it gives you the two things a Meta campaign needs on your side: a fast landing page and honest numbers. Each product page and your link-in-bio page has built-in checkout on your custom domain with free SSL, so a Reels click goes straight to a page that loads and converts on mobile. Discount codes (REEL10, STORY15) tag each campaign in your orders. The one shared catalog keeps prices, stock and variants identical across your store and your link-in-bio page, so an ad never sends someone to an outdated price. Your reports give orders and revenue per period, which is what you put next to your Ads Manager spend to compute a ROAS you can trust, and on the Pro plan, audience analytics and exports let you match campaigns to customers.

FAQ

How much should I spend to test Meta Ads?

Enough to buy about 1,500 to 3,000 clicks over two to three weeks, so that a conversion rate near 1.5% produces 20 to 45 purchases. At $0.80 per click, that is $1,200 to $2,400. A creator with a smaller budget can test at $15 a day for 21 days ($315) and read direction, not certainty. Below that, the numbers are noise.

Should I boost posts from Instagram or use Ads Manager?

Ads Manager. Boosting is fine to push a post to existing followers for a few dollars, but it cannot optimize for purchases and gives you almost no control over placements or reporting. Any campaign meant to sell should be built in Ads Manager with the Pixel installed.

Is Meta Ads worth it with 0% transaction fees on my store?

The fees you pay your commerce platform are separate from ad cost, but they add up in the same margin calculation. A platform taking 2% of a $45 order costs you $0.90 per sale; an ad costs $30 to $60 per sale. The ad is the number to optimize, and a platform without transaction fees simply leaves more margin to pay for it.