Contact

Tools · Ad platforms

Meta Ads

Facebook and Instagram, checked against your real orders.

Meta Ads is Meta's ad system for Facebook, Instagram and its other apps, run in Ads Manager and measured with the Meta Pixel and Conversions API. Donut Analytics builds and runs Facebook and Instagram campaigns with a few well-funded campaigns, creative tested in a grid, and conversions set up per product. Before budget moves, we check what Meta credits against your own orders.

Let’s Chat All tools

How we use it

How we use Meta Ads

  • Fewer, better-funded campaigns

    We'd rather run a few well-funded campaigns than many starved ones. Campaigns built around product lines and buyer lookalikes give Meta's delivery enough data to learn, and budget grows only as results hold.

  • Creative tested in a grid

    We test several creative concepts, each with more than one message. A cell has to spend enough to judge before we call a winner, and we flag results that lean on view-through credit.

  • A conversion for each product

    If one pixel covers many products, a campaign optimizing for purchases can collect credit for all of them. We build custom conversions so each campaign is judged on its own product's sales.

  • Automatic extras, chosen on purpose

    Meta can add its own touches to an ad, like site links, summaries and reworked formats. We decide which to use instead of accepting the defaults, so the ad people see is the ad that was approved.

  • Policy checked before launch

    We check copy against Meta's ad policies before it ships. Financing offers, for example, belong on the landing page or in an on-image disclaimer, not in the ad text.

  • Conversions API, scoped with your developers

    When a pixel only runs in the browser, we scope a Conversions API setup with your developers or site vendor: which events go server-side, how they're matched to the pixel so nothing counts twice, and how each event respects the visitor's cookie choice.

In practice

From Meta accounts we run

Real work, kept anonymous.

  1. 01

    A split test read against real orders

    For a client that sells event tickets, each event's budget was split between a purchase-optimized campaign and a traffic campaign. In Ads Manager the purchase half looked many times better, but most of that credit came from purchases of other events on the same pixel. Read on 7-day click and real orders, the split didn't prove itself. We recommended going back to traffic campaigns and pooling the purchase budget in one line where it could learn.

  2. 02

    Concept mattered more than message

    For one advertiser, we tested several image concepts, each with an urgent and an informational message. Cost per result came out level between the two messages. The concept made the difference.

  3. 03

    A dormant account brought back

    For a brand that sells online, we took over a Meta account nobody was managing, paused what wasn't working and consolidated to two campaigns by product line on purchase lookalikes. Daily budget grew as results held.

  4. 04

    Competitors' ads, reviewed

    We pulled a competitor set's ads from Meta's Ad Library into a short deck, to show what they were running and where the space was open.

What it can’t tell you

What Meta Ads can't tell you

  • Which sales would have happened anyway

    Ads Manager credits Meta's ads under Meta's own attribution settings, including people who saw an ad and never clicked. It can't separate sales the ads added from sales that were coming regardless.

  • Which product a purchase was for

    On a shared pixel, a campaign set to optimize for purchases can be credited with every product the site sells. Ads Manager won't flag it. You find it by checking against your own orders.

  • A clean comparison with older reports

    In March 2026 Meta narrowed click-through attribution for website and in-store conversions to link clicks only. Likes, shares and saves moved to engage-through attribution, Meta's new name for engaged-view, so reports from before and after the change don't line up exactly.

FAQ

Questions about Meta Ads

Do I need the Conversions API if I already have the Meta Pixel?
It helps when browser tracking loses signal. The Conversions API sends events from a server, and Meta matches them to Pixel events by event name and event ID so the same action isn't counted twice. We scope which events to send and work with your developers or site vendor to build it.
Why does Meta report more conversions than Google Analytics?
Meta counts people who saw an ad and later bought, even without a click, and a shared pixel can credit one campaign with purchases of other products. Google Analytics only sees visits to your site. Before budget moves, Donut Analytics reads Meta on a click-only window and checks it against real orders.
Should we use Advantage+ campaigns?
It depends on how much conversion data the account has and how much control you need. Advantage+ sales campaigns, formerly Advantage+ shopping, let Meta's AI set the audience, placements and budget. In the accounts we run, we decide each Advantage+ setting on purpose, and Meta's automatic creative extras stay off unless we've chosen them.
How many creatives should we test on Meta?
A few concepts, each with more than one message, and no more than the budget can fund. A test only means something once each version has spent enough to judge. In one test we ran, the concept made the difference and the message angle didn't.
How do I know if Meta ads are adding sales?
Ads Manager alone can't tell you, because it can't see which sales would have happened anyway. A lift test with a holdout group, or a media mix model that measures advertising against total sales, can estimate what Meta added.

Pairs with

Guides: Incrementality vs. attribution: what's the difference? FAQ

All tools

Know what Meta is really driving.

Tell us what you’re working on. We’ll build a measurement-first plan for your brand. Email hello@donutanalytics.com or call (407) 394-9720.

Let’s Chat See our results