Lumiceo | Meta Ads Tracking Audit

Meta ads tracking audit for Shopify brands. Pixel + CAPI dedup, purchase values, ROAS you can trust. $150 audit, credited to fixes from $750.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

Is your Meta ROAS real?

You can check most Shopify stores in about 15 minutes, without any tools.

Here's the exact order I check things in 🧵

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

Black Friday is 16 days away.

Before you raise budget: does Meta count the same sales as Shopify? Same 7 days, same time zone.

If Meta shows more, some sales are counted twice and your ROAS is lying.

$150 tracking audit for Shopify brands. DM AUDIT.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@holmisthename Part of it is what Meta is learning from. If it only gets the form fill, it finds more people who fill forms. Send the $19 purchase back through CAPI with the lead's hashed email and phone, and it starts finding buyers. Were they optimizing for leads or purchases?

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@CamRConsulting Number 6: reporting ROAS without checking the purchase value Meta gets. If the pixel sends the subtotal and CAPI sends the total with shipping and tax, ROAS is off before you even get to margin.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@autoady_ Solid list. I'd add: on that test purchase, check that the value and currency in Events Manager match the Shopify order exactly. Dedup can be perfect and ROAS still wrong if one side sends the value before discounts.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

A significant edit — audience, creative, or optimization event — can restart Meta's learning phase. Meta says the learning phase usually ends after about 50 optimization events since the last significant edit. BFCM Nov 27: lock what works. #metaads #shopify #bfcm

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

If you send purchases from both the Meta Pixel and the Conversions API, Meta only treats them as one when the event name and event ID match, within 48 hours. Use the order number as the ID on both sides, then check Events Manager for duplicates.#metaads #shopify #ecommerce

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@DerRobertsEcom Nice day. On the 20% bump: watch frequency on the winner. If CPA keeps creeping while frequency climbs, it's fatigue not budget, so the static ads you're planning are the right lever. Get them live before the next increase.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@EcomMexico 4.2 at $225/day means there's room. Raise ~20% every 48-72h on the same ad set, and keep a fresh creative test running alongside. Ads "not holding spend" is usually thin creative depth, not budget, so build that before the big jump.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@ifourth Instead of new campaigns for the winner, try bumping the existing one 15-20% every 2-3 days. New campaigns split data and each restarts learning; consolidated budget on a proven ad set usually holds ROAS better while you test new products.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@fromadstoassets On the early BF offer: put it in the creative (headline + first 3 sec), not just on site. Same winning ad + new offer overlay tells you in ~48h if it's the offer or the angle. And on the 1.2k CBO, judge on 3-day before the next budget move.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@willlthomas Worth checking: link clicks vs landing page views in Ads Manager since Oct 24. If LPV per click dropped, it's the link/site or in-app browser, not the auction. Fix that before touching creative or you'll be testing ads on a broken path.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@bronnybuilds Good catch. Also check delivery: partnership ads that lose FB placements shift spend to IG only, so CPMs jump before you spot the errors. Once creators connect a Page, relaunching as fresh ads with the same creative is usually faster than waiting on old IDs.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@MicoMartinez @lukasenECOM Intra-day swings are mostly attribution lag. Meta backfills purchases for hours, so 10x at noon that "cools off" often evens out by next day. Judge on a 3-day rolling view + Shopify MER, and avoid mid-day edits that reset learning.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@ecommadam One 1.1 day isn't a trend. Check last 3 days: CPM up, CTR steady = auction (post-election), leave it. CTR down + frequency up = fatigue, ship new hooks. We kill sets same day when CTR craters, e.g. a $169 set at 3.4% vs 17-21% on healthy ones.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@Daggerecom Seen it. Top spenders get re-reviewed as reach grows, so one borderline claim kills your best ad. Check the exact policy reason, request review, and keep a cleaned-up variant (softer claim, no before/after) live in the same ad set so spend doesn't crash.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@aalanazarr Usually not a bug. Meta counts view-throughs + Shopify orders credited to email/organic. Check Events Manager: pixel + CAPI should share an event_id so purchases dedupe. Then compare 7-day click only vs Shopify, that's closer to the real number.

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@DerRobertsEcom If those 5-6 repeat buyers also clicked or saw an ad, Meta counts them as ad purchases too. Then Ads Manager looks better than what the ads really brought in new. Do you compare Meta purchases with new-customer orders in Shopify?

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@arnavsawantt @vytisbareika Broken bottles are the worst kind too, because it isn't the customer changing their mind, it's the box. One fix: send Meta the purchase value minus your usual breakage rate. Then the ads chase sales that stay sales, not just checkouts. What's your breakage rate, roughly?

[Lumiceo | Meta Ads Tracking Audit @lumiceolabs]

@arnavsawantt @vytisbareika That refund lag is the hard part. Meta counts the sale on day 1, the refund lands weeks later, so ROAS always looks better than it was. A weekly sheet with revenue after refunds next to Meta's number closes that gap. Are most refunds coming from one courier or region?