Installing a pixel on your own domain: what changes in week one

Installing a first-party pixel takes an afternoon. Understanding what it shows you takes about a week. This post walks through the first seven days on a typical Shopify or BigCommerce store: what appears on day one, what the numbers mean, where the gaps are, and which of them you can close. No campaign changes, no feed work. Just measurement, so the decisions that follow rest on something solid.

Day zero: the install

On Shopify, the pixel is a custom pixel pasted once in the store's customer events settings. On BigCommerce, it is a script added through Script Manager with coverage set to all pages, including checkout and order confirmation, with the store's data layer switched on so add-to-cart and purchase events are available. Either way the script is served from your own domain on a neutral path.

Two settings matter before you look at any number:

• Consent category. Which category the pixel sits in decides how many visitors it can count. Under "analytics" or "marketing" it is withheld from every visitor who does not click Accept, and on many stores that is most of them. Some regions allow first-party analytics under "essential" and some do not, so check your consent rules, or ask whoever manages your banner, before you choose.
• Consent by region. The pixel decides a visitor's default posture from where they are. The EU, UK, Switzerland, Brazil and Quebec wait for the banner. US states track by default and honor the Global Privacy Control browser signal. You do not configure this by hand; you check that it is doing what you expect.

Then you open the Real-Time page, visit your own store, and watch your session appear. That is day zero.

Day one: sessions and sources

The first full day gives you two numbers and one surprise.

Sessions. The overview shows qualified sessions, with obvious bots and empty drive-by hits filtered out. Compare it to what your previous analytics tool reported. If the pixel sits in the essential category and the old tag sat behind the banner, expect the pixel's count to be noticeably higher. Nothing about your traffic changed. The measurement did.

Sources. Every session carries a source and medium: `google / cpc`, `facebook / paid_social`, `newsletter / email`, `direct`. Paid sessions also carry the platform's click ID, captured from the landing URL on arrival. On day one you will see your Google, Microsoft, Meta and TikTok traffic split cleanly, before any spend data is connected.

The surprise. The "direct" row is usually bigger than you would like. Direct means the pixel did not see a source on that visit. Some of it is real brand traffic. Some of it is missing tags from an app, a redirect, or a stripped referrer. Note the size of it now. You will come back to it.

Days two to three: orders start matching

Orders arrive on two paths. The browser records the purchase when the shopper reaches the confirmation page. The store platform sends the order separately, by webhook, independent of the browser. The Conversions page shows both:

• Total conversions the browser saw.
• Verified conversions, confirmed by the store platform.
• Verified revenue, refunds, and the net figure after refunds.

The gap between total and verified is your first diagnostic. If verified is much higher than total, the browser is missing purchases (blockers, an early tab close, a checkout on a different domain). If total is much higher than verified, the webhook is lagging or not installed. Say your store does 80 orders a day and the browser saw 50 of them. That is a normal starting point, and it is exactly why the server-side record exists.

By day three the identity layer has started joining sessions to customers. A shopper who browsed on a phone Tuesday and bought on a laptop Wednesday becomes one person, joined by the email or customer ID on the order. Orders that arrived with no browser source can be matched to an earlier session that carried a click. That is where the "direct" row starts to shrink.

Day four: the ad-blocker gap, measured

By now you can put a number on something that used to be a guess.

Open the Sources page and filter to conversions with an Unknown source. The drilldown explains why each one has no signal: a blocked browser, a checkout on a different site identity, an order placed by phone or at a register. Then compare verified orders to orders with a known source.

Say your store takes 500 orders in the week. Say 320 of them carry a traced source, 60 came through the register or by phone, and 120 arrived with no browser signal at all. That last group is your blocker gap. Some of it will close over the following weeks as identity matching connects late orders to earlier visits. Some of it will not, because the visitor never let a script run.

That number is worth knowing on its own. It tells you how much of your revenue any browser-only tool was silently dropping, and it tells you what share of your sales the platforms could not have been bidding on.

Days five to seven: the first real report

By the end of the week you have enough to read three pages with confidence.

Two things you will not have yet. Ad spend and return on ad spend appear once the ad connectors are linked, which is usually the next step. And attribution windows longer than a week are still filling in, because click history begins the day the pixel went live. Order history is imported from the store, so the customer record starts with what you already know, but the touch trail starts now.

You will also see a data quality grade and a consent breakdown. The consent page shows two tiers: visitors whose click IDs can be captured for your own dashboard, and the smaller group whose conversions may also be sent back to ad platforms. A gap between the two is normal. It reflects people who allowed analytics but declined ad sharing, and the system is honoring that.

What to do with week one

Resist the urge to change bids. The point of week one is a baseline. Three actions are worth taking:

1. Fix the direct row where it is fixable. Check that email links, affiliate links and app redirects carry tracking parameters. The pixel can only read what the URL brings.
2. Move the script if it is behind the banner. If sessions look low against your platform's own visitor count, the consent category is the first suspect.
3. Connect the ad accounts. Spend lands beside revenue, return on ad spend appears per source, and the platform-versus-pixel comparison starts.

After that, the pixel keeps running, the touch trail lengthens, and the attribution pages get more useful every week. Everything else in the workspace, from feeds to email lanes to the CRM, reads from this record. Week one is the foundation.

Want to see your own week one? Talk to us.

---

Want this for your business? Talk to us How the pixel works