Ecommerce Conversion Tracking: How It Works and How to Set It Up for Native Ads

Ecommerce conversion tracking is how you know which ad produced which order. On Meta you probably never think about it, because the pixel, the catalog and the reporting come as one system. Native ad networks like Taboola and Outbrain work differently. Traffic often lands on an article-style page first, buyers tend to take longer to convert, and you set up much of the tracking yourself. If it is wrong, you will either kill campaigns that work or fund ones that don't. This guide covers what to track, how the pieces connect, a step-by-step setup using a made-up example, and the mistakes that cause the most damage.

What ecommerce conversion tracking covers

Ecommerce conversion tracking records what a visitor does after clicking an ad, such as viewing a product, adding to cart and buying, and sends that data, including order value, back to the ad platform and your analytics. It lets you tie revenue to a specific campaign, ad or traffic source.

A conversion is any action you want a visitor to take. In ecommerce, the one that pays the bills is the purchase. But a good setup also records the steps before it, so you can see where people drop off and give the ad network more to learn from.

The standard events, in order of the shopper's journey:

• Landing page view: the visitor arrived. This is your check that clicks became visits.
• Product view: they looked at a product page.
• Add to cart: they showed buying intent.
• Initiate checkout: they started paying.
• Purchase: the order went through. Send the order value, the currency and a unique order ID with it.

The purchase event with its value is the one that matters most. Without value, you can count orders but you can't calculate return on ad spend (revenue divided by what you spent on ads), and you can't judge your cost per acquisition. Our guide to DTC customer acquisition cost covers how to calculate that figure and decide what a good one looks like for your margins.

How ecommerce conversion tracking works

Every tracking setup solves the same problem: the ad click happens on one site, the purchase happens on another, and something has to connect them. The connection is an identifier.

The basic flow

1. A person clicks your ad on a publisher's site. The network adds a unique click ID to your destination URL.
2. Your page loads. The network's pixel (a small snippet of code, also called a tag) reads that click ID and stores it in the browser.
3. The visitor browses, adds to cart and buys. Your store fires events, and the pixel reports them back with the stored click ID.
4. The network matches the click ID to the original ad and shows the conversion in its reporting.

Two ways to send the data back

A browser pixel runs in the visitor's browser. It is quick to install, but it can be blocked by browser privacy settings, ad blockers or a visitor who closes the tab before the code runs.

A server-to-server postback (S2S, sometimes called a server-side event) skips the browser. When an order is confirmed, your server sends a message directly to the network's server with the click ID and order details. It takes more work to build, but it doesn't depend on what the shopper's browser allows. Many brands run both and let the network deduplicate them using the order ID or event ID.

The exact setup differs between networks. Our post on Taboola conversion tracking walks through how the pixel, click ID and postbacks fit together for that platform.

What matters for a DTC brand running native ads

If you come from Meta, these are the points where native tracking will surprise you.

The click ID must survive the whole path

Native campaigns often send traffic to an advertorial (an article-style page that sells while it reads like editorial content) before the product page. Our advertorial guide explains the format. For tracking, it adds a hop. If the click ID is in the URL on the advertorial but gets dropped when the visitor clicks through to your store, the purchase can't be matched to the ad. Check that the ID is passed along on every link and button that leads to checkout.

Native shoppers often buy later

Native ads reach people while they read, not while they shop. Many won't buy in the first session. That makes the attribution window (how long after a click the network can still credit a sale to that ad) important. If the window is shorter than your buyers' typical delay, real sales will go unreported and campaigns will look worse than they are. Check the default on your network and set it deliberately.

Numbers will not match Meta or your store

Each platform counts in its own way: different windows, different handling of repeat visits, different treatment of blocked pixels. Your store's order data is the closest thing to the truth, because it records actual payments. Use the network's numbers to compare ads and campaigns against each other, and use store revenue to check the total. Decide this before launch so you don't argue about it afterwards. If you want a refresher on how the Meta side compares, see our piece on Facebook ads for ecommerce.

Give the network more than one signal

Native campaigns on a new account usually produce fewer purchases than Meta does at similar spend, so purchase-only optimization can be slow to learn. Tracking add to cart and checkout events lets you spot which sources bring engaged visitors even before orders pile up. Only optimize toward a higher-funnel event as a temporary step, and check that it still leads to sales.

Values and currency need to be right

Send the actual order total in the right currency, and decide whether it includes shipping and tax. Pick one rule and apply it to every platform, or your return on ad spend comparisons won't be fair.

A worked example: setting up tracking for a first native campaign

The brand below is a made-up skincare company, not a client case. It sells on a hosted store platform and plans to send native traffic to an advertorial, then to a product page.

1. Define what counts. The team decides the primary conversion is a completed purchase with order total, currency and order ID. Add to cart and initiate checkout are secondary events.
2. Install the network's pixel on every page a visitor can reach: the advertorial, the product pages, the cart and the confirmation page. Use the store's native integration or tag manager if one is available.
3. Add the click ID to the ad's destination URL using the network's tracking parameter. Check the network's documentation for the exact name, because it differs between platforms.
4. Make the advertorial pass the click ID forward. Every call-to-action button links to the product page with the parameter still attached.
5. Store the click ID with the order. When the visitor reaches checkout, the ID is saved in a cookie or as a cart attribute so it ends up on the order record.
6. Set up the server-to-server postback. When the store marks an order as paid, a webhook sends the click ID, order value, currency and order ID to the network.
7. Name your campaigns and ads clearly, and add a source label to the URL, so that your analytics can separate native traffic from Meta and from email.
8. Place a real test order by clicking a live ad or a test link built the same way. Confirm the event shows in the network's reporting, in your analytics and in the store's order list, and that the values match.
9. Compare after a few days. Check network-reported orders against store orders from the same traffic. A gap is normal. A very large gap points to a broken step, usually step 4 or step 5.

If step 9 shows a problem, repeat steps 4 to 8 and test again before you raise budgets. Our Taboola campaign setup guide covers the rest of launching a campaign once tracking works.

Common ecommerce conversion tracking mistakes

• Never testing with a real order. A pixel that shows as installed isn't proof that purchases are recorded. Buy something.
• Losing the click ID between the ad and the checkout, usually on the advertorial or at a redirect.
• Firing the purchase event twice. Reloading the confirmation page can record the same order twice. Use the order ID to deduplicate.
• Sending the purchase event without a value, or with the wrong currency.
• Running the pixel and the postback with no deduplication, so every order counts double.
• Judging campaigns too early. Delayed purchases haven't arrived yet, so early numbers undercount.
• Comparing native results to Meta's reporting one-for-one. The windows and counting rules differ.
• Changing the setup in the middle of a test. If you edit tags and URLs while a campaign runs, you can't tell whether a change in results is real.
• Ignoring consent and privacy rules. Make sure your cookie banner and tag setup follow the laws that apply to your customers.

What to do next

Before you spend on a new network, run through the list above with a single test order and write down what each system reports. Then pick your source of truth, set your attribution window and decide which events you will use to judge ads.

If you are new to the channel, the complete guide to native advertising explains how the networks, placements and ad formats fit together. If you would rather hand off the build, an agency that runs native campaigns can handle tracking and campaign management for you.

Frequently asked questions

The bottom line

Ecommerce conversion tracking is plumbing, and it decides whether the rest of your native testing means anything. Define the events and the value you will send, keep the click ID intact from ad to order, use a pixel plus a postback, and test with a real purchase. Then judge campaigns against store revenue, not any single platform's dashboard.

Keep reading

• DTC customer acquisition cost
• Taboola conversion tracking
• advertorial guide
• Taboola campaign setup guide
• complete guide to native advertising
• an agency that runs native campaigns