Media Strategy & Analytics
I help brands build the strategy, measurement infrastructure, and channel architecture that connect media spend to real business outcomes — online and in-store.
"Most brands have a media execution problem disguised as a media strategy problem."
About
I'm Corinna Wu — an independent media strategist. I build the analytical frameworks and channel strategies that agencies often skip, connecting media spend to real sales outcomes — whether that's in-store traffic, online conversions, or both.
My background spans DIRECTV, OMD, Time Inc., Rational Interaction, and Xevo (connected vehicle tech) — across US and international markets, managing budgets from $1M to $25M+. I've led campaigns in integrated media strategy, attribution modeling, and performance marketing across more than 10 countries.
I work with brands that are between where they are and where their media could take them — and I help them cross that threshold with confidence and clarity.
What I do
I trace where your attribution breaks down — UTM gaps, GA4 misconfiguration, data pipeline failures — and build the infrastructure that connects media spend to real business outcomes. You get a reporting framework your team can actually use.
Full-funnel channel architecture defining the role of each channel across awareness, consideration, intent, and conversion — with targeting methodology, budget allocation logic, and a test-and-learn framework built in from day one.
I don't just advise — I buy. Fluent in programmatic platforms, Meta campaign management, paid search, and direct mail. I can run the media, optimize in-platform, and manage vendor relationships directly when that's what the engagement calls for.
Performance analysis, optimization recommendations, vendor management, and executive reporting — on a continuous basis. I work alongside your agency or as a standalone strategic resource, keeping strategy and measurement connected over time.
Who I work with
Your agency runs campaigns but nobody is connecting media spend to actual business outcomes. You need someone who can see across channels and tell you what's working and why.
You're ready to expand beyond search — into social, programmatic, radio, or direct mail — but you want a structured test-and-learn approach before committing budget.
Your dashboards are full. Your reports are long. But nobody is telling you what the numbers actually mean or what to do about them. That's where I come in.
Selected Work
Built UTM framework, GA4 reporting protocol, POS-to-media correlation methodology, and data quality fix — transforming a program running blind into one with location-level accountability.
Designed and launched a 5-channel integrated media system across programmatic, social, CTV, audio, and direct mail — from carrier route selection to channel role framework to budget architecture.
Developed a proprietary impressions-per-household coverage metric, ran Pearson correlation analysis across 8 media variables, and identified over/under-served locations driving reallocation decisions.
Architected a paid media strategy across 267 locations and 6 brands — spanning paid search, Meta, programmatic, direct mail, radio, and CTV. Included tier allocation, phased bidding strategy, radio format evaluation framework, and competitive intelligence.
Case Studies
Five engagements across franchise automotive, experiential retail, and multi-brand national programs. Each involved ambiguous data environments, tight timelines, and the need to produce both strategic frameworks and day-to-day executional precision.
Regional Automotive Service Brand
Inherited a campaign structure with legacy 2-mile targeting radii at multiple underperforming locations. Car counts were declining YoY while delivery data showed frequency spikes — the same households being served 10+ times per month with nowhere left to expand reach.
The Problem
Coverage was the issue, not creative or spend. Programmatic display and Meta were oversaturating core zones while adjacent trade areas with real demand went uncovered. The signal was clear: reach suppression and frequency saturation at the same locations, year-over-year.
Approach
Built a three-criteria location prioritization model — applied in sequence, not simultaneously:
Cross-referenced platform weekly reach/frequency data against POS car count files across the full portfolio. Proposed expansions in two tiers: 8 locations from 2mi to 5mi; 2 rural outliers from 8mi to 20mi. Held spend constant so any change was attributable to structure. The rest of the portfolio served as a natural control group.
The framework produced two additional optimization candidates using the same three-criteria model — systematizing the learning rather than treating the test as a one-time fix.
Regional Automotive Service Brand
Client had direct mail budget but no methodology for route selection, no UTM tracking, and no way to attribute customer visits to specific mail pieces. Built the full program infrastructure from scratch — carrier route prioritization, UTM framework, and Q1 matchback analysis.
Approach
Developed a two-ratio carrier route selection methodology:
Vehicle targeting criteria defined explicitly: mass market brands, model year 2023 and older, 30K+ mileage, personal use. New mover recency rule: 12-month home purchase window. Built UTM tracking for all locations with unique store-level parameters and validated with client analytics contact before launch.
The carrier route methodology is replicable for any automotive, QSR, home services, or retail franchise with a radius-based trade area. Built to be repeatable — not rebuilt each quarter.
Regional Automotive Service Brand
Three separate attribution failures were discovered during a performance audit — each from a different source, each silently corrupting the data used to make optimization decisions. Identified and resolved all three, then built a stronger measurement model to replace what had been broken.
Failure 1 — SFTP Data Pipeline
An automated IT move rule was pulling files out of the vendor's SFTP inbound folder before the programmatic platform could read them — making 15 weeks of POS data appear missing. Root-caused via terminal SFTP navigation and timestamp comparison. Coordinated the fix across the programmatic vendor, agency IT, and client liaison. Recovered all files; established three-party monitoring protocol.
Failure 2 — GA4 Channel Misclassification
Paid social campaigns were tagged with utm_medium=social, causing GA4 to classify a significant paid social budget as Organic Social. Paid traffic appeared to represent only 20.9% of California sessions. Fix: utm_medium=social → utm_medium=paid-social.
Failure 3 — Unsupported Attribution Claim
An internal draft claimed coupon creative was driving car count lift without POS-level redemption tracking. Corrected the framing: attribution requires a POS-level match. Without it, the observation is correlation, not cause.
Measurement Upgrade
After fixing the environment, ran Pearson correlation across 8 media variables. Impressions-per-household coverage (r=0.479) outperformed GA4 sessions (r=0.343) — establishing it as the primary optimization signal going forward.
Data infrastructure problems often masquerade as performance problems. Diagnosing the environment before drawing optimization conclusions is a discipline that protects both the budget and the client relationship.
National Multi-Brand Automotive Franchise
Inherited a 6-brand, 267-location franchise network mid-transition from an incumbent agency. Built the H2 investment framework, launched Meta as a net-new channel, developed bidding strategy positions across four Google channels, and established measurement infrastructure — simultaneously.
Investment Framework — WPI Analysis
Reviewed the H2 allocation model and identified five blind spots before the plan was finalized: Demand Gen CPA seasonal volatility (Dec–Feb CPAs $8–12, March–April $17–24); zero Demand Gen coverage for two brands; an inverse WPI score/budget correlation at several locations; minimum budget viability concerns at sub-$1K locations; and a gravity score model refinement identifying moderate-probability, large-household-count markets as the highest marginal return for new channel investment.
Bidding Strategy Positions
Developed clear positions on five proposals: EC4L (agree, implement now); Dynamic Conversion Values (decline — contradicts car count objective); Geo-Clustering (defer to Q3 — needs baseline data first); Brand Exclusions on PMax (agree, implement at launch); Capacity-Driven Dynamic Budgeting (stretch goal — requires POS integration).
Meta Launch Architecture
Designed the franchise network ROP program Meta launch: $230K across 25–30 locations, tire services May–August / oil change September–December. Separate campaigns per service line to prevent algorithmic mixing. Identified and corrected a vendor pixel misunderstanding before launch. Recommended 2-year POS pull for lookalike audience seeding and offline conversion matching.
Operating simultaneously at the strategic and executional layer — and pushing back on the plan with specific, documented concerns — is where the most durable value gets created.
Experiential Retail Brand — New Market Launch
A new retail concept entering a suburban New York market had the highest Average Order Value in the chain but the lowest transaction counts — clear evidence the store was attracting the right buyer, just not enough of them. Before recommending a single line item of spend, three compounding sources of brand confusion were diagnosed.
The Brand Confusion Trifecta
Owned Channel Audit
The store's own locator page described it using product language that reinforced the competitor's category — not its own. The Google Business Profile routed to the parent brand's website rather than the store's own page. Both issues were flagged and escalated before any media spend was recommended.
Media Strategy
Sequenced to fix the leaky bucket before filling it: resolve CRM routing and GBP errors first, then activate always-on foundation (direct mail at 5–7mi targeting female homeowners $125K+ HHI ages 45–70; Meta/Instagram; Nextdoor for conversational brand differentiation; Paid Search bridging the SEO gap), then layer seasonal creative on top.
Scale Framing
With 10–15 additional locations planned for 2027, the single-store engagement was reframed as a repeatable grand opening playbook — not a one-time fix. The trade area targeting model, audience segmentation, seasonal calendar, and brand education creative framework developed here become the template for every subsequent opening.
High AOV + low transactions = a traffic and awareness problem, not a product problem. Diagnosing what the media needs to do before prescribing what channels to buy separates a strategic recommendation from a line item.