A DTC brand had outgrown the agency model: four vendors, last-click attribution, and ad spend continuing on out-of-stock SKUs. We brought paid media in-house, built first-party analytics from scratch, and instrumented inventory-aware ad serving. ARR went from $20M to $54M in 18 months.
average ROAS sustained over 18 months
vs 2.1× baseline across four prior agencies
Client
DTC consumer brand, $20M ARR at engagement start
Timeline
18 months ongoing (audit + foundation + run rate)
Team
2 engineers, 1 paid media strategist, 1 analyst, 1 designer
Engagement
Long-term retainer with quarterly strategy reviews
A DTC consumer brand had scaled to roughly $20M ARR through paid media, but the model had stopped working. Four different agencies were running channels in parallel — Google, Meta, TikTok, and Pinterest — each optimizing for their own dashboard, none coordinating with each other or with the brand's actual inventory. ROAS was deteriorating quarter over quarter, attribution was fundamentally last-click, and the CFO was getting genuinely concerned.
The core decision was structural before it was tactical: consolidate paid media in-house with one team accountable across channels, build the first-party data layer that the agencies had no incentive to build, then optimize.
01
Six-week audit across all four channels: spend allocation, creative-iteration velocity, audience overlap, attribution model, inventory integration. Recommendation: terminate three of the four agencies, retain the most performant one as a creative-only partner, build everything else in-house with our team.
02
Server-side Google Tag Manager + Snowplow event collection feeding a Snowflake warehouse. Schema designed in collaboration with the brand's finance team so the same numbers reconciled across paid media, finance, and the board reports. This took 10 weeks and was the most important thing we did.
03
Implemented a position-based attribution model (40/20/40 first-touch / mid-touch / last-touch) with channel-specific weighting based on lift studies. The model was validated against incrementality holdout tests on Meta and Google quarterly.
04
Real-time inventory feed (every 15 minutes) automatically pausing ad sets, product feeds, and shopping campaigns on stocked-out SKUs across all four channels. Roughly 8% of ad spend was being burned on stockouts pre-launch. After: 0%.
05
Brought channel management in-house with our paid media strategist + analyst leading day-to-day operations. Weekly cross-channel reallocation based on the attribution model, not channel ROAS in isolation. Creative production retained at the surviving agency.
06
Looker Studio dashboards built off the warehouse, not off agency feeds. Daily decisions, weekly leadership review, monthly board pack — all from the same source of truth. The CFO stopped asking which number was real.
average ROAS sustained over 18 months
vs 2.1× baseline
in annual paid media spend managed in-house
inventory-aware ad serving across all four channels
ARR growth over 18 months
Paid media + organic compounding
reduction in agency management fees
The most important outcome wasn't the ROAS number — it was that the brand finally trusted its own numbers. Every weekly leadership meeting now starts from one dashboard, with finance, marketing, and operations all looking at the same figures. The CFO described it (informally) as 'the first time in three years I haven't had to argue about which version of the truth we are using.'
Paid media stopped being a black box and started being a manageable lever. The team can now answer 'if we add $200K of spend next month, where does it go and what should we expect?' with a defensible answer instead of a guess. That visibility is what unlocked the additional growth — leadership was willing to invest because they could see what they were investing into.
We continue as the analytics + paid media team on a long-term retainer, augmented by quarterly strategy reviews. The brand's in-house team has grown to handle creative direction and brand work, while we run the data + media operation.
01
Audit + structural recommendation
Cross-channel spend audit, attribution diagnostic, agency-vs-in-house decision framework, transition plan.
02
First-party analytics build
Server-side GTM, Snowplow events, Snowflake warehouse, schema reconciled with finance.
03
Attribution + inventory integration
Multi-touch attribution model, lift-test validation, real-time inventory feed, automated ad-set pausing.
04
Agency transition + in-house ops
Three agencies offboarded, paid media operations established in-house, creative agency retained on revised scope.
05
Steady-state operation
Weekly cross-channel reallocation, monthly board pack, quarterly strategy reviews + lift-test refresh.
09
Technical SEO that engineers understand. Core Web Vitals, structured data, indexing audits — and the PPC and analytics work to back it up.
Learn more10
Get recommended by ChatGPT, Claude, Perplexity, Gemini, and Google AI when buyers ask for companies like yours. Entity clarity, structured data, AI-readable evidence, third-party validation.
Learn more06
LLM features, RAG systems, document AI, and workflow automation — integrated where they pay for themselves, not bolted onto everything.
Learn moreStagnant 18-month organic plateau, broken Core Web Vitals, JS-heavy SPA blocking crawlers. Migration to Next.js with comprehensive schema, performance overhaul, and indexation cleanup.
A 12-person ops team manually classifying 800+ shipping documents a day became a 4-engineer + LLM pipeline doing it in minutes, with a 0.3% error rate. Customs holds dropped 80%.
Most engagements start with a 30-minute discovery call. No pitch deck, no NDAs on day one — just an honest conversation about your situation.
Schedule a Call