Why Andrew Faris chose Bily after a head-to-head Meta tracking test

Andrew Faris has spent his career inside the operating reality of ecommerce growth: the campaigns, forecasts, creative decisions, inventory constraints, and measurement gaps that sit between a promising brand and durable scale.

His public biography traces that work back to 2014, when he joined a bootstrapped ecommerce business during an approximately 18-month journey from zero to $20 million. He later became a growth strategist at Common Thread Collective, where he worked on growth problems for brands including The North Face, Melin, TheraGun, DIFF Eyewear, and Born Primitive. He was promoted to VP of Growth in 2018.

Andrew then served as CEO of 4x400, an ecommerce aggregator operating a portfolio of as many as six brands. Today, AJF Growth brings that operator perspective to cohort forecasting, managed Meta media buying, creative, and ecommerce education. Andrew also hosts The Andrew Faris Podcast, where he breaks down the mechanics behind growth decisions rather than reducing them to headline tactics.

That background makes his Bily testimonial particularly useful: it is not a generic endorsement. It documents a direct comparison and the decision that followed.

The test design

Andrew compared Bily with Elevar on a managed ecommerce account. The two sides used the same ad sets and the same ads so he could inspect how the signal setups behaved under comparable media conditions.

In the testimonial, Andrew describes Bily as the clear winner. He points to stronger overall performance, cheaper clicks, higher click-through rate despite a slightly higher CPM, and materially stronger Event Match Quality.

The delayed-attribution comparison

Andrew compared each setup's seven-day-click ROAS with its 28-day-click ROAS. On Bily, the values moved from 0.94 to 1.04, which he calculated as an 11% increase. On Elevar, the comparison moved from 0.70 to 0.74, approximately 6%.

The 11% figure is therefore a delayed-attribution ROAS comparison inside the tested account. It is not an 11% revenue claim, an 11% CPA reduction, or proof of independently measured incrementality.

The signal-quality result

The Meta Event Match Quality view added a second layer of evidence. Andrew reported:

• PageView Event Match Quality of 6.9 on Bily versus 6.3 on Elevar.

• Purchase Event Match Quality reaching 9.0 on Bily.

• AddPaymentInfo Event Match Quality reaching 8.8 on Bily.

Those scores gave him confidence that the performance difference was supported by stronger event matching, not just a single top-line number.

The decision

Andrew said the result justified moving the tested account toward Bily and rolling the stronger setup into the team's operating process. Private onboarding correspondence documents Bily work across multiple AJF-managed Shopify brands.

The public version deliberately does not name those client brands. Each brand owns its own customer relationship and result, and each should approve public identification before appearing in Bily marketing.

This story is also dated honestly. It describes Andrew’s customer-reported test and rollout work; it does not claim that AJF Growth or every referred account remains a current Bily customer in 2026.

Watch Andrew's testimonial

Andrew Faris's four-minute Bily Pixel vs Elevar test update. Select the image to watch on Loom.

How to read the result

The matched ad sets and ads make the comparison more useful than an unrelated before-and-after snapshot, but the recovered material is still a customer-reported account test rather than a randomized experiment audited by an independent third party.

Auction conditions, attribution behavior, creative delivery, and other account changes can affect platform results. The exact figures belong to the tested account and period; they are not a guarantee for other brands.

Next steps

Repeat the comparison over a predeclared window with both signal paths running simultaneously. Freeze material campaign changes, verify deduplication, and reconcile Meta-reported purchases with paid Shopify orders, refunds, and contribution margin. If the signal-quality and delayed-attribution advantages remain aligned with the order ledger, publish the expanded result with the client's explicit approval.