How Bullseye Sellers Dropped Joanna Vargas’ nCAC by ~40%
Premium skincare brand Joanna Vargas used a modern performance stack—Meta’s Andromeda creative protocol, Wicked’s Advanced Signal, and 5 Forces AI—to cut new customer acquisition cost from roughly $116 to $69 in November, while getting clearer, trusted attribution across channels.
At a glance
~40% nCAC Drop
New customer acquisition cost (nCAC) in one month
Premium Brand
Skincare DTC brand with celebrity founder
Signal Architecture
Creative as targeting + enriched conversion events
5 Forces AI
Powers weekly Scale/Chill/Kill calls for the media team
“We dropped our customer acquisition cost by 40%. We went from acquiring customers at about $116 on average down to $69. Strategy plus Wicked’s Advanced Signal are 50/50 in terms of credit for that success.”
THORSTEN LOTTO
Lead Strategist & E-commerce Developer, Bullseye Sellers
The Brand & The Agency


“Analyzing data was one of the first tasks I was given. To do my job as a strategist, I need data that’s on point, accurate, and pulling correctly. That’s how we ended up on Wicked Reports.”
THORSTEN LOTTO
Lead Strategist & E-commerce Developer, Bullseye Sellers
The Challenge:
Great-looking Platform ROAS, Stalled New-Customer Growth
Before Wicked Reports, the team was stuck in the usual post-iOS mess:
Platform bias everywhere
-
Meta, Shopify, Google all claiming credit
-
Lots of view-through and retargeting inflation
Messy, manual reporting
-
Multiple tabs: Ads Manager, Shopify, GA, spreadsheets
-
Numbers rarely matched, no granular campaign-level truth
Decisions made on incomplete KPIs
-
Some campaigns looked “good” on CTR or a single KPI
-
But high nCAC, low revenue, or weak profitability when looked at holistically
Risk of “paying for the same customer twice”
-
Retargeting and Advantage+ setups quietly pulling budget away from true TOF
The team needed:
A single source of truth for all channels
A way to see new vs repeat customers clearly
A stack that could train Meta to prioritize new customers, not just easy retargeting wins
Why They Chose Wicked Reports
Bullseye Sellers rolled out Wicked Reports across their key brands—especially Joanna Vargas—because they needed scientific, click-based, omnichannel attribution that matched how serious performance marketers actually think.
WHAT STOOD OUT:
First-party, click-based attribution
Real order IDs, real click paths, real customer IDs—not blended view-through guesses.
New vs Repeat baked in
Clear visibility into new customers vs repeat customers by channel, campaign, ad set, and creative.
FunnelVision & 5 Forces AI
-
FunnelVision as Thorsten’s “daily playground” for analysis
-
5 Forces AI to apply intentions, zones, and consistent decision logic so the team can scale/chill/kill without guesswork.
“Wicked gives us continuous data to see what campaigns and strategies are working, and which ones are driving new customers to the brand. It brings all our channels into one platform, gives us granular control, and the data is trustworthy.”
THORSTEN LOTTO
Lead Strategist & E-commerce Developer, Bullseye Sellers
The Breakthrough:
From Audience Targeting to Signal Architecture
Instead of another “we tested a new audience and ROAS went up” story, Thorsten framed this as a shift in architecture:
Most advertisers test image vs video. Bullseye used a Persuasion Matrix approach, building large, creative-heavy ad sets (20+ creatives) where each focused on a specific pillar:
-
Emotional drivers
-
Logical reasons
-
Social proof & product benefits
-
Founder story & authority

Meta’s pixel alone couldn’t distinguish cheap clicks from valuable customers. Bullseye “enriched” the data using Wicked's Advanced Signal.
-
Sent new customer-specific purchase events into Meta
-
Helped curb retargeting bias
-
Rewarded Meta only when it found new buyers
Implementation:
The New-Customer Growth Engine
-
Connected Shopify, Meta, Google, Klaviyo into Wicked
-
Verified attribution health and variance (<2% in comparisons)
-
Set up new-customer conversion events for Advanced Signal
-
Built creative-heavy ad sets (~20+ ads per set)
-
Each creative mapped to a specific persuasion pillar
-
Let Meta’s Andromeda model distribute spend dynamically across the matrix
-
Advanced Signal began sending high-fidelity new-customer events
-
5 Forces AI ingested performance data, attribution, intentions, and zones each night
-
Media buyers started using 5 Forces AI as a daily/weekly playbook
-
Review FunnelVision + 5 Forces AI each week
-
Tag campaigns: Scale / Chill / Kill
-
Reallocate budget to the best new-customer drivers
-
Retire “feel-good” campaigns that looked good on a single KPI but lost money in reality
“If I’m not around to validate data or guide the media buying, they can just jump in, click the 5 Forces AI report, and it’s all there for them—what to scale, what to chill, what to kill”
THORSTEN LOTTO
Lead Strategist & E-commerce Developer, Bullseye Sellers
Results:
nCAC Down ~40% in a Single Month
When they zoomed in on November, the impact was obvious:
Why It Worked
Shared Truth, Not Siloed Dashboards
Wicked gave them a unified, trustworthy view of performance across channels, within 2% of reconciled numbers.
New-Customer Signal, Not Just Any Conversion
Advanced Signal stopped rewarding Meta for easy retargeting. The algorithm learned to chase new buyers.
AI With Guardrails
5 Forces AI didn’t just spit out random suggestions; it followed a methodology built around:
-
Intentions
-
Payback windows
-
Zones
-
Cohorts and attribution models
Creative as Targeting
The Persuasion Matrix meant they weren’t obsessed with new audiences—they let Andromeda + Wicked figure out which hooks actually convert.
Delegatable, Repeatable Process
If the strategist is busy, the media buyers always know what to do. It eliminates hours of manual analysis and bogus spreadsheets.
Want to Drop Your nCAC Too?
-
An eCommerce brand tired of paying to reacquire your own customers, or
-
An agency strategist who wants trusted data, new-customer clarity, and delegatable AI guidance…
First-party, click-based attribution
Advanced Signal to train Meta on new customers
5 Forces AI to turn messy data into clear next actions





