Track ROI from Facebook & Instagram Ads: Meta Pixel, Conversions API & GA4 Setup

Key Takeaways

• Meta Pixel misses 40 to 60% of iOS conversions. Conversions API recovers 10 to 20% of that lost data

• Shopify and WooCommerce native setups take 30 minutes and track purchases automatically

• Your ROAS looks 20 to 30% lower than 2020 numbers even with identical performance

• True ROI equals revenue minus all costs (product, shipping, returns, fees, ad spend) divided by ad spend

Why Facebook and Instagram Numbers Are Fiction

You spend £1,000 on Facebook and Instagram ads. Meta’s dashboard says you made £5,000 back. Check your bank account and the maths doesn’t work. Here’s why: Apple changed the rules in 2021. Roughly half of iPhone users are now invisible to Facebook ad tracking. Your ads work, people buy, but Meta can’t see most of it.

The fix to properly measure ROI needs three things: Meta Pixel tracks browsers, Conversions API catches what Pixel misses, and Google Analytics gives you independent verification. On Shopify or WooCommerce? Thirty minutes to set up. Worth it to know if you’re actually making money from your Facebook ad campaigns.

Meta Pixel Setup for Facebook Ad Tracking (15 Minutes)

The Pixel is code that watches your website. When someone views a product, adds to basket, or buys, it tells Facebook and Instagram. Without it, Meta has no clue whether your ads work.

Shopify: Settings, Customer Events, Connect Facebook. Pick your Pixel, done. Takes 10 minutes.

WooCommerce: Install “Facebook for WooCommerce” plugin. Connect Facebook, choose Pixel, handles tracking automatically.

Custom sites: Grab code from Facebook Events Manager, paste into site header. Need your developer for 20 minutes.

Check it works with Meta Pixel Helper Chrome extension. Visit your site, look for green tick. Buy something from yourself and confirm it shows up in Events Manager with the actual amount, not £0.

Critical Tracking Events

Your Pixel tracks three things: product views (ViewContent), basket adds (AddToCart), and purchases (Purchase). Each needs product names, prices, what they bought.

Shopify and WooCommerce handle this automatically. Custom sites need code pulling real prices and order totals.

Big mistake: tracking set up twice. App plus manual code counts every sale twice. Facebook thinks you’re making double. Pick one method. Check Events Manager for duplicate timestamps.

Conversions API (The iPhone Fix)

Half of iPhone users are invisible to Pixel. Conversions API sends sales data from your server to Facebook, bypassing phones and browsers completely.

Think of it this way: Pixel watches through your shop window. Someone pulls curtains (blocks tracking), can’t see anything. Conversions API checks your till at day’s end. You know who bought what because transactions happened.

Result? Ten to twenty percent more sales show up. Not new sales. Sales happening anyway that Facebook couldn’t see.

Shopify: Settings, Apps, Facebook, toggle “maximum data sharing.” Done.

WooCommerce: Marketing, Facebook, enable server events. Grab Access Token from Facebook Events Manager, paste it in.

Critical: when running both Pixel and Conversions API, prevent double counting. Shopify and WooCommerce handle deduplication automatically. Custom setups need matching ID numbers so Facebook knows it’s the same purchase.

Why Your ROAS Dropped (And It’s Not You)

Run Facebook and Instagram ads for years? Numbers got worse around 2021 to 2022. Apple changed the rules.

Facebook used to count sales within 28 days of a click. Now? Maximum 7 days. Someone sees your ad, thinks about it ten days, then buys? Meta can’t count it even though your ad influenced the sale.

For products over £50 where people think before buying, Facebook’s numbers look 20 to 30% worse than they used to. Same actual sales. You were happy with £3 back per £1 in 2020. Now seeing £2.30 and panicking. If profit margins work at £2.30, you’re still making money. Meta just can’t prove it helped with all of it.

Quick task: verify your domain in Facebook Business Manager. Business Manager, Brand Safety, Domains, add website, verify. Takes 10 minutes. Without this, iPhone tracking is even worse.

Google Analytics (Your Second Opinion)

Facebook’s numbers are biased. Their job is making their platform look good. Google Analytics just tells you what happened on your website.

You need both. Facebook shows what Facebook thinks it deserves credit for. Google shows actual website activity. They’ll differ. That’s normal. Having both spots when something’s broken.

Tag your ads so Google knows where traffic came from. Add this to ad links:

?utm_source=facebook&utm_medium=paid_social&utm_campaign=summer_sale

Change campaign name to match yours. This tells Google “this person came from Facebook.” Otherwise Google thinks they typed your address directly.

In GA4: Reports, Acquisition, Traffic Acquisition. Find “facebook / paid social” to see sales Google attributes to Facebook. Compare to Facebook’s numbers. Expect 15 to 30% difference. If Facebook says £10,000 and Google says £800, tracking’s broken.

Check GA4 marks “purchase” as conversion. Admin, Events, toggle purchase on. Otherwise GA4 won’t count sales. For running both platforms together, both tracking systems give the full picture.

Why Numbers Don’t Match (It’s Normal)

Everything set up correctly. Facebook says 50 sales, £2,000. Google says 38 sales, £1,600. Both tracking same website. What gives?

Both tell truth from their perspective. Example: someone sees your ad Monday, clicks, browses, doesn’t buy. Wednesday they remember your brand, search Google, click organic result, buy. Facebook says “my ad caused this.” Google says “came from Google search.” Both contributed.

Track both for months. Facebook typically reports 15 to 30% higher than Google (or reverse). Pattern stays consistent. If Facebook says £5,000 and it’s usually 20% optimistic, real Facebook revenue is closer to £4,000.

Truth test: actual sales. Check Shopify reports, WooCommerce orders, bank. Facebook says £10,000, Google says £7,000, but you received £8,500? Both systems are off. Trust neither completely. Oddly the smartest position. Proper ROI reporting means accepting uncertainty whilst making informed decisions.

Calculating Actual Profit to Measure ROI (Not Just ROAS)

Facebook says £5 back per £1 spent. Sounds brilliant. But are you making money? To properly measure ROI, you need the full picture.

Real ROI: Sales revenue minus product costs minus shipping minus returns minus payment fees minus ad spend equals actual profit.

Example: £1,000 on ads. Facebook says £5,000 sales. Products cost £3,000, shipping £600, returns £400, Stripe £100. That’s £5,000 minus £4,100 minus £1,000 equals minus £100. Lost money despite Facebook showing good numbers.

Profit margin matters more than ROAS. Thin margins under 30%? Need really high ROAS to profit. Over 60% margins? Profitable at much lower ROAS.

For repeat customers, first purchase doesn’t tell whole story. Spend £40, customer buys £30? Looks like minus £10. But they buy three more times this year? Made £120 against £40 acquisition. Very profitable.

At 100 plus purchases weekly, switch Facebook to “value” optimisation instead of “conversions.” Shows ads to people likely spending more per order. Works brilliantly for different price points when combined with data-driven marketing strategies. Check realistic targets for your industry.

What Numbers Actually Matter

Facebook’s dashboard overwhelms. Loads of graphs, numbers, percentages. What matters:

ROAS: Money back per pound spent. £4,000 sales from £1,000 spend equals 4.0 ROAS. What Facebook thinks it caused, not guaranteed truth. Your main judging number.

Cost Per Purchase: Spent per sale. £1,000 for 40 sales equals £25 per sale. Must be lower than profit per order. Make £30 profit, pay £25 to acquire? Making £5 per sale. Pay £35? Losing £5 every time someone buys.

Frequency: Times average person saw your ad. Hits 3 or above? People are tired. Costs jump 15 to 30%, fewer buy. Change creative every couple weeks. Retargeting? Change weekly.

Don’t panic daily. Facebook tracking delays. Monday sales show Wednesday or Thursday. Check weekly or monthly for trends. Daily checks only spot breaks (tracking stopped, ad rejected, budget blown).

Monthly Maintenance

Tracking breaks. Facebook updates, Apple changes privacy, themes update and break code. Check monthly everything works.

Simple checklist: Buy from own website. Check shows in Events Manager with correct amount. Look for duplicates. If Conversions API set up, check says “Received from Server.” Check Analytics shows Facebook traffic. Compare purchase counts. Facebook says 100 sales but got 70 orders? Something’s wrong.

When breaks: Pixel stops after theme update? Reconnect Facebook in settings. Purchases show but £0.00 value? Checkout not telling Facebook order amounts. Check integration permissions. Conversions API stops? Token expired. Generate new one in Events Manager.

UK compliance note: Make sure your cookie consent banner delays Pixel firing until users accept marketing cookies. Required for GDPR. Most consent tools (Cookiebot, OneTrust, Shopify’s Customer Privacy) handle this automatically.

Facebook has Test Events feature showing real time what fires. Nothing appears within a minute? Broken.

Honestly, if this feels overwhelming whilst running your business, you’re not alone. At Lucky Penny, we set up tracking as part of campaign management because accurate data is everything. Can’t make smart decisions with fiction numbers. Get in touch if you’d rather focus on what you’re good at whilst we handle technical bits.

Caio for now

Ali Puglianini

Frequently Asked Questions

How long to set up? Shopify or WooCommerce? About an hour for Pixel, Conversions API, Google Analytics together. Custom sites take 2 to 4 hours. Then another hour over days checking purchases track correctly. For businesses wanting simpler ROI tracking without complex software, this setup covers the essentials.

Why different numbers between Facebook and Google? Different measuring. Facebook: “showed ad, bought within 7 days, I caused it.” Google: “came from this source, bought, this caused it.” Someone clicks Facebook ad, later searches brand on Google, buys? Both claim credit. Both contributed. Expect 15 to 30% difference.

Need Conversions API or just Pixel? Need both for 2025. Pixel alone misses half of iPhone users (40 to 60%) with tracking turned off. Conversions API catches 10 to 20% more via server data. Half of UK shoppers use iPhones. Without Conversions API, deciding on incomplete data.

What ROAS to aim for? Depends on profit margins. Most UK e-commerce with 40 to 50% margins needs 2.0 to 2.5x just breaking even after costs. Profitable scaling above 3.0 ROAS. Seeing 5.0 or higher? Either exceptional or retargeting people already knowing your brand. Important: actually making money after all costs?

Purchases show but £0.00 value? Tracking knows purchase happened, not how much. Shopify or WooCommerce? Check Facebook app has permission seeing order details. Custom sites? Purchase code needs pulling actual order amount, not placeholder. Common: tracking fires before checkout finishes processing order value.