Mcfly Ads

Why pixels fail after privacy

Browsers thinned the identity graph. Suites still sell path credit — and overlapping claims inflate platform ROAS. Mcfly stops at Shopify sales after returns ÷ ad spend, then break-even, then allocate to grow. Decide any day with numbers you can trust.

What broke

The graph needed one continuous person.
Privacy cut the rail.

Path credit assumes a durable join from ad click → site → checkout. That join is sparse by design now. These are structural constraints — not Mcfly lab stats.

  1. Safari ITP

    JavaScript cookies get a short leash — roughly a week in many cases, and often about a day after an ad-click landing. Storage that attribution models treat as durable is not.

  2. iOS link tracking protection

    Tapped links can lose the click IDs suites need to stitch sessions. When the ID is stripped, the model still prints neat credit — from a thinner graph.

  3. CNAME cloaking collapsed

    First-party DNS workarounds that tried to look “same-site” got capped when IP distance and browser heuristics caught up. The cloaking era is over as a reliable fix.

Noise in, noise out

CAPI is not a resurrection spell.
Triple Pixel is not magic.

Server-side event streams still ingest junk. Industry reporting often cites invalid / bot-adjacent traffic risk in a rough ~25–30% band — treat that as a risk narrative, not a Mcfly measurement of your store. Feed that into Conversion API and delivery can optimize toward noise.

Triple Whale’s “Triple Pixel” is still the standard stack: JS tag + UTMs + cookies. Renaming the pipe does not restore a thinned identity graph.

The overlap problem

Channels claim the same sale.
Stacked credit exceeds reality.

Pixel and path models often double-count. The same order shows up in Meta, Google, and email — so platform ROAS looks great while total claimed revenue exceeds what Shopify actually kept after returns. Overlap is measurable in aggregate; chasing which click “deserved” the sale does not fix the P&L.

What matters more: statistical spend allocation that yields the highest sales and profit — move budget toward channels that clear break-even on real Shopify sales, not toward whoever claimed the last touch. Mcfly ships that cash desk: sales after returns ÷ spend, then allocate to grow. No path theater.

Platform variance sample → · Product →

Price honesty

GMV tax vs a flat desk

Suites that scale with your sales tax the merchants who grow. Mcfly’s path is App Store Free (Meta + Google) → Pro $39/store/mo when Billing is announced — not a cut of GMV. See Pricing.

Product refuse

We explain the failure. We don’t ship it.

This site may demystify why pixels fail. The Shopify app does not ship them. Full desk: Product. Data diet: Privacy.

  • No tracking pixels as a Mcfly product
  • No path / multi-touch / view-through “true ROAS”
  • No identity-graph clone or connector zoo as the core offer

What to do instead

Measure sales.
Make the budget call.

Total ROAS = sales ÷ ad spend for the same period. Break-even from your margin. One allocation call. Path credit trains a model; Total ROAS trains the week.

Install free App Store install steps How it works →

Quick answers

Questions we get

Does Mcfly replace my Meta or Google pixel?

No. Platforms keep their own tags for delivery. Mcfly never ships a path pixel or claims to restore cross-site identity. We reconcile money out against money Shopify recorded.

If the graph is thin, why do suites still look precise?

Sparse joins still produce confident-looking credit. Stacked “driven revenue” that never ties to banked sales is the failure mode — not a sync bug you can patch with another tag.

What number should you trust any day?

Total ROAS = Shopify sales after returns ÷ spend, then compare to break-even from profit margin. That is the whole desk — see Product. Overlap is why path credit fails; allocation on cash math is what grows sales and profit.