Retail Brand Omnichannel Marketing

Project Background

This retail brand is a chain specializing in mid-to-high-end home and lifestyle products, with over 200 physical stores nationwide, and also operates a Tmall flagship store, a Douyin live streaming room, and a WeChat mini-program mall. Over the past two years, the brand completed basic digital transformation of its online and offline channels, but the membership systems, order data, and advertising data of each channel remained siloed, causing marketing decisions to rely on experience and making it difficult to quantify attribution. In Q3 2024, the brand launched a marketing system upgrade project aimed at "omnichannel growth," hoping to complete data infrastructure and advertising strategy restructuring before the Double 11 shopping festival.

Core Challenges

At the start of the project, the team faced three core pain points:

1. Severe data silos: Data from offline POS systems, e-commerce backends, advertising platforms (Ocean Engine, Alimama), and private domain CRM had inconsistent standards. The same user could not be linked across different channels by ID, leading to the long-standing issue of "one user being repeatedly targeted and multiple channels being billed repeatedly."
2. Failed attribution model: The previous advertising used "last-click attribution," causing many offline in-store conversions to be incorrectly attributed to e-commerce ads, while search conversions driven by Douyin content were completely untrackable. Management could not determine true ROI.
3. Coarse budget allocation: Channel budgets were fixed based on historical sales proportions, lacking a dynamic adjustment mechanism based on real-time performance. During major promotions, additional budget was often allocated "by gut feeling" to channels with unclear performance.

Solution

The brand introduced the GEO Engine omnichannel marketing data engine and completed the system restructuring in three steps:

Step 1: Data middle platform integration. Through GEO Engine's ID-Mapping capability, offline member phone numbers, e-commerce accounts, device IDs, and advertising platform cookies were encrypted and matched to build a unified 360° user profile database. In just 3 weeks, 12 million historical user records were cleaned and linked, identifying 68% cross-channel duplicate users.

Step 2: Full-funnel attribution model rebuild. Based on the integrated data, a multi-touch attribution model based on Shapley Value was deployed, incorporating 12 key touchpoints such as offline store visits, online browsing, content interactions, and search clicks into contribution calculations. After the model went live, management could for the first time clearly see the true value of complete conversion paths like "Douyin seeding → Tmall search → offline pickup."

Step 3: Intelligent budget allocation engine. GEO Engine connected to each advertising platform's API to enable hourly data feedback and automatically generate channel budget adjustment recommendations based on attribution results. The system set rules such as "automatically reduce budget when ROI falls below threshold, automatically increase budget when above threshold," while incorporating manual intervention points aligned with major promotion rhythms.

Execution Highlights

• Double 11 live test: During the 2024 Double 11, the system monitored the advertising efficiency of over 40 ad campaigns in real time, dynamically adjusting the Douyin live streaming room budget share from 35% to 52%, while reducing budget for low-efficiency brand keywords in search ads, improving overall budget utilization by 27%.
• Offline store coordination: GEO Engine identified user groups within 3 kilometers who had seen online ads but not converted, and pushed precise in-store care tasks to store guides, ultimately driving over 4,200 users to visit stores, increasing offline cross-selling rate by 18%.
• Private domain re-engagement: For the "high-intent non-converted" audience segment from attribution, it was automatically synced to the WeCom operation pool, combined with exclusive coupons, achieving a 22% recall conversion rate within 7 days.

Results Data

Within the complete 30-day period after project launch (including the Double 11 sprint), the brand's omnichannel advertising ROI increased from 1.8 to 7.6, a 320% increase. Specific results include: cross-channel duplicate targeting costs decreased by 41%, new customer acquisition costs decreased by 36%, and private domain GMV share increased from 12% to 23%.

Lessons Learned

This project validated a key insight: the value of omnichannel marketing lies not in "investing in more channels," but in "seeing clearly how every dollar collaborates across channels to create conversions." GEO Engine brings not just a data integration tool, but a reusable decision-making methodology. The brand's marketing director reported that monthly review reports that previously took 2 weeks to complete are now automatically generated daily by the system, allowing the team to shift energy from data organization to strategy innovation. For retail enterprises planning similar projects, it is recommended to prioritize ensuring senior management's commitment to unified data standards, and to reserve at least 2-3 weeks for data cleaning and model tuning to avoid rushing to launch before a major promotion.