HiJenny

Starting from Absolute Zero

No Ads, No Tracking, No App — and Five Weeks to Demo Day

Building an entire acquisition engine before the product was even live.

The Challenge:

HiJenny had literally started the Monday before our first call. The website had been built over a single weekend, the app wasn't live in the App Store yet, and there was zero marketing infrastructure in place. To make things harder, the team's Meta Business Manager was tangled — two separate business accounts owned different assets, and the primary admin got locked out of Facebook during the verification process. The previous contractor had set up Google Ads with a Target CPC bid so restrictive they were getting 30 impressions per day. With Demo Day on March 24th and the goal of 100 projects on the platform, there was no room for a slow ramp.

The Matchbox Solution:

We treated Day 1 as an emergency deployment — getting on calls the same day, troubleshooting account access live, and standing up campaigns within hours of signing the contract.

• Diagnosed and worked around Meta Business Manager lockout by leveraging a secondary admin account with page-level access to build campaigns in real time
• Completed advertiser verification and beneficiary status submissions on the same call, reducing typical 1–2 week delays to 48 hours
• Immediately fixed the Google Ads setup — removed the restrictive Target CPC bid that was suppressing all impression volume
• Launched a $1/day test ad within the first session to validate account health and clear any policy flags before scaling
• Assigned a dedicated team member (Michael Hood) locally in San Francisco for in-person coordination with the HiJenny team at HF0

Building the Full-Funnel Acquisition Engine

From First Dollar Spent to 36,000 Weekly Visitors

Multi-channel, geo-targeted campaigns optimized for app downloads in the Bay Area.

The Challenge:

HiJenny needed to drive app downloads for a consumer product that didn't exist in the App Store yet, in a hyper-local market (three Bay Area counties), against a homeowner demographic that was unfamiliar with the brand. There was no historical performance data to optimize against, no retargeting audiences to leverage, and the conversion event — an actual app download — couldn't be tracked initially because the app was still pending Apple's review. The team needed to spend aggressively to learn quickly, but couldn't afford to waste budget while the funnel infrastructure was still being built.

The Matchbox Solution:

We designed and launched a three-stage funnel architecture that could start generating signal immediately while building toward deeper conversion optimization over time.

• Built a top-of-funnel awareness campaign driving to the homepage, generating landing page views at $0.13 per visit to populate retargeting audiences
• Created a middle-of-funnel conversion campaign optimized for download page visits, using a custom conversion event fired on page load at the app download step
• Uploaded an 80,000-record homeowner list (all Bay Area home purchases in the prior 24 months) as a custom audience for precision targeting
• Geo-targeted campaigns to Marin, San Mateo, and Santa Clara counties with age targeting of 40–65 based on first-time homeowner demographics
• Launched YouTube awareness campaigns in Marin County to flood the local market with video ads alongside the street team's in-person efforts
• Implemented 48-hour reporting cadence with a custom dashboard so the HiJenny team had full visibility despite being locked out of Facebook

Cracking App Install Attribution

Engineering Tracking Through iOS Privacy, SDK Gaps, and Tag Manager Chaos

Building reliable measurement when every layer of the attribution stack needed fixing.

The Challenge:

Tracking app installs from paid campaigns was the single biggest technical hurdle. The initial Meta SDK implementation didn't account for iOS 14+ App Tracking Transparency, meaning the vast majority of installs couldn't be attributed back to ads. The Facebook pixel was duplicated — one hard-coded on the site, one in Google Tag Manager — creating potential double-counting. At one point, the engineering team accidentally killed the entire Google Tag Manager container during a site update, wiping out all web-side tracking for three days. And the conversion events that were firing weren't being sent back to the ad platforms, so campaigns couldn't optimize for the outcomes that mattered.

The Matchbox Solution:

We treated attribution as an ongoing engineering project, shipping fixes iteratively while keeping campaigns running and optimizing against the best available signal at each stage.

• Sent detailed iOS 14+ ATT compliance instructions to the engineering team, enabling post-iOS 14 install attribution through the Meta SDK
• Detected the Tag Manager outage within hours using our AI-powered anomaly detection tool, which flagged a conversion drop while spend remained constant
• Guided the team through AppsFlyer integration as the mobile measurement partner, replacing the limited Facebook SDK with a proper MMP
• Walked the client through remapping Facebook Events Manager from Facebook SDK to AppsFlyer on a live call, fixing the data pipeline in real time
• Audited the Tag Manager JSON and identified the duplicate pixel issue, providing a remediation plan to consolidate tracking into a single clean container
• Created interim manual tracking by appending AppsFlyer download data to the performance dashboard so optimization decisions weren't delayed while the full fix was implemented

Creative Optimization at Accelerator Speed

Fear-Based Hooks, UGC-Style Video, and Systematic Testing

Discovering what resonates when you're inventing the category in real time.

The Challenge:

HiJenny was creating a new product category — an AI remodeling assistant — which meant there was no playbook for what messaging would resonate. The team was producing video content using AI tools and needed rapid feedback loops on what was working. Early batches included ads that looked too polished or AI-generated, drawing negative comments on Instagram. With dozens of creatives cycling through across multiple funnel stages, there was no systematic way to identify which messages, hooks, or formats were actually driving results versus just generating cheap impressions.

The Matchbox Solution:

We built a performance-driven creative testing framework that connected ad metrics to messaging insights, enabling the team to produce more of what worked and kill what didn't.

• Identified that fear-based, pain-point hooks dramatically outperformed abstract or aspirational messaging — ads about timeline anxiety, contractor distrust, and renovation cost fears consistently drove the highest CTRs
• Discovered that handheld, POV-style UGC videos with clean audio outperformed polished or AI-generated content across both Meta and YouTube
• Achieved a blended 4.5% CTR across all campaigns and a 16–20% CTR on retargeting — rates the client described as "astronomically high"
• Implemented systematic creative rotation with automatic pausing of underperformers, cycling through 7+ batches of video and static ads over 5 weeks
• Proposed a naming convention and messaging pillar framework (control, confidence, value, fear) to enable structured A/B testing of themes across audience segments
• Retired specific ads generating negative sentiment (e.g., AI-looking visuals drawing "AI slop" comments) before they could damage brand perception

‍