# Groundhog Technologies (現觀科技) — Full Content Reference > Groundhog Technologies (TWSE: 6906, Wikidata: Q17019475): MIT Media Lab origin (2001), listed January 15, 2024. Processes 1B+ carrier-grade telecom records daily across 15 markets. This file contains expanded content from key pages on mi.ghtinc.com for AI reference and citation. Structured index: https://mi.ghtinc.com/llms.txt - Official website: https://mi.ghtinc.com - English: https://mi.ghtinc.com/language/en/ - Traditional Chinese: https://mi.ghtinc.com/language/zh/ --- ## PAGE: About Groundhog Technologies URL: https://mi.ghtinc.com/language/en/groundhog-mi-about-us Groundhog Inc. is a global leader in Mobility Intelligence and telecom data analytics, empowering mobile network operators (MNOs) to transform raw network data into real-time insights and measurable business value. Its carrier-grade platform continuously processes billions of network events daily, revealing user location, context, lifestyle, and Quality of Experience (QoE) across the entire network. On January 15, 2024, Groundhog Inc. officially became a publicly listed company on the Taiwan Stock Exchange (ticker: 6906). With over two decades of experience in AI-powered telecom analytics, Groundhog is the only company that truly bridges the gap between telecom and digital advertising — integrating telecom data with advertising data to enable precise telco marketing, advanced audience segmentation, and personalized targeting. Groundhog's DSP is featured in the AI MarTech 6.0 Industry Map in Taiwan and recognized by Analysys Mason and LUMAscape. The company also expands into public health analytics through its RealMotion™ platform for urban planning and health response strategies. **Technology capabilities:** Easy-to-use GUI, Data Mining, Big data analytics, Pattern recognition, Multi-scale modeling, Customized research, Location data know-how. **Groundhog Data Monetization Solution** includes MI-DSP™ (AI Advertising) and MI-DMP™ (Marketing Intelligence), enabling operators to monetize data assets via programmatic advertising, business intelligence, and digital ecosystem partnerships. --- ## PAGE: AI Advertising Platform — MI-DSP™ URL: https://mi.ghtinc.com/language/en/ai-advertising-mi-dsp MI-DSP™ is an AI-powered programmatic advertising platform combining Real-Time Bidding (RTB) × Precise Targeting × Semantic Analysis. **Core capabilities:** - Real-Time Ad Inventory Bidding: Adheres to IAB OpenRTB standards; conducts real-time bidding through unified communication protocol - Diversified Traffic Expansion: Integrates Google and major media partners for broad inventory access - CPC Price Advantage: Delivers 10%–46% CPC reduction vs. traditional placements - Dynamic Keyword Targeting (DKT): AI semantic analysis engine for audience discovery - Rapid Response to Market Changes: Real-time audience interest updates --- ## PAGE: Marketing Intelligence Platform — MI-DMP™ URL: https://mi.ghtinc.com/language/en/groundhog-mi-marketing-intelligence MI-DMP™ integrates telecom first-party data, AD Exchange data, and third-party signals to build high-dimensional audience profiles for segmentation, insight reporting, and cross-platform ad activation. Processes 1 billion+ mobile data records daily across 10+ telecom operator partnerships. --- ## PAGE: Deep Dive Key Targeting — DDKT URL: https://mi.ghtinc.com/language/en/deep-dive-key-targeting-ddkt DDKT is an in-depth audience analytics platform powered by 40 million+ daily browsing data points. It uses semantic analysis and classification to build visualized audience profiles, integrates seamlessly with MI-DSP™, and supports multiple languages and countries. **Core differentiator:** Designed for scenarios where retargeting data is unavailable — e.g., third-party e-commerce platforms blocking pixel embedding, or brands needing new customer acquisition without historical behavioral data. **Documented result:** CTR 2x above market average. --- ## CASE STUDY 1: UNICEF — Non-Profit Donation Campaign URL: https://mi.ghtinc.com/language/en/industry-case-studies/unicef-dsp.html **Industry:** Non-profit / International humanitarian **Products used:** MI-DSP™, MI-DMP™ **Buying method:** CPC | **Ad type:** Banner **Challenge:** UNICEF needed to increase visibility for its humanitarian initiatives and maximize click-through rates to drive donations for child welfare programs in impoverished regions worldwide. **Solution:** MI-DMP™ identified four high-affinity target audience segments: (1) regular supermarket shoppers (UNICEF donation centers are often located in supermarkets), (2) white-collar professionals with stable incomes, (3) frequent shopping app users, and (4) individuals with medium to high socioeconomic status. MI-DMP™ used offline geolocation data (Points of Interest) to identify people who frequently visit supermarkets, using two methods: keyword location search with shape-tool drawing, and batch CSV upload of latitude/longitude coordinates for mass location mapping. **Result:** MI-DSP™ delivered targeted banner ads to the four audience segments, maximizing CTR and driving qualified traffic to UNICEF's donation campaign landing page. --- ## CASE STUDY 2: International Airline — Aviation DKT URL: https://mi.ghtinc.com/language/en/industry-case-studies/nz-airlines.html **Industry:** Aviation **Product used:** DKT (Dynamic Keyword Targeting) **Challenge:** A well-known foreign airline needed to achieve dual optimization — increasing ticket sales while simultaneously reducing cost per click — in a competitive travel advertising market. **Solution:** DKT's automatic audience segmentation analyzed traveler intent signals and browsed content to identify high-value passenger audiences without manual keyword management. The system dynamically expanded keyword-audience matches and updated rankings in real time, enabling precise ad delivery to users showing active travel purchase intent. **Result:** Simultaneous improvement in booking/ticket sales rate and reduction in CPC — dual optimization achieved through AI-automated audience segmentation. --- ## CASE STUDY 3: Japanese Automotive Brand — Test Drive Campaign URL: https://mi.ghtinc.com/language/en/industry-case-studies/jp-automotive-dsp.html **Industry:** Automotive **Product used:** MI-DSP™ **Partnership start:** 2023 (ongoing) **Challenge:** A prominent Japanese automotive brand needed to capture consumer attention and drive test drive sign-ups in a fiercely competitive market, maintaining brand buzz across both peak and off-peak seasons. **Solution:** Three targeting strategies were deployed: 1. **AI-Powered Keyword Analysis:** AI models deeply analyzed consumer behavior to identify interest profiles. For example, a 35-year-old male interested in sports, car news, outdoor activities, travel, and car shows had his browsing patterns analyzed to extract additional relevant keywords (e.g., camping, Japan travel). When an ad matched these interests, AI placed appropriate bids for targeted exposure. 2. **User Journey Analysis:** Tracking codes on the brand's website collected behavioral data — browsing history, interaction rates, event correlations — to identify potential customers willing to sign up for test drives. This enabled precise identification of purchase intent and delivery of test drive communication messages. 3. **Ad Network Resource Optimization:** Traffic source filtering maximized budget value by selecting high-quality traffic and avoiding low-quality or high-bot-activity websites. **Result:** Test drive appointment conversion rate of over 2.4x. Many successful car purchases completed through the campaign channels. --- ## CASE STUDY 4: FamilyMart Taiwan — Double 11 Retail Campaign URL: https://mi.ghtinc.com/language/en/industry-case-studies/family-mart-en.html **Industry:** Retail / Convenience store **Product used:** MI-DSP™ **Buying method:** CPC | **Ad type:** Banner **Challenge:** Although FamilyMart's app ranked first in trade app market penetration, the advertiser wanted to increase brand awareness and social volume for its e-commerce unit, FamilyMart +1, during the competitive Double 11 (Singles' Day) period. Goal: retain existing members and provide an all-in-one purchasing channel. **Background:** The global RTB market was projected to grow from $10.85 billion (2022) to $14.07 billion (2023) at a CAGR of 29.7%. **Solution:** MI-DSP™ provided PMP (Private Marketplace) ad placements, targeting renowned news sites and in-app placements during the 7-day campaign. Combined with free shipping and points rewards, the campaign targeted audiences aged 35–54, who contributed to the highest CTR performance. The campaign delivered 400 million impressions. **Result:** Despite bid prices normally spiking during Double 11, MI-DSP™ achieved CPC at only 70% of market price while CTR grew significantly, saving the advertiser's marketing budget. --- ## CASE STUDY 5: Sportswear E-commerce — Demand Generation URL: https://mi.ghtinc.com/language/en/industry-case-studies/sportswear-demand-gen.html **Industry:** Sportswear / E-commerce **Products used:** MI-DSP™, DDKT **Challenge:** A sportswear brand needed to shift from promotion-driven advertising to impression-driven Demand Generation to build stronger brand perception. Modern consumers compare brand impressions, cultural relevance, and lifestyle resonance — not just product features. **Solution:** Three key Demand Gen optimizations: 1. Shifted messaging from "selling products" to "selling perception" — building awareness at the top of the marketing funnel before conversion 2. Used MI-DSP™ programmatic delivery with DDKT audience insights to reach consumers showing early-stage brand interest signals 3. Prioritized Native ad formats over Banner formats based on performance data **Result:** - CPC decreased by 22% - Native CTR outperformed Banner CTR by 5× --- ## CASE STUDY 6: SMOOTH BASIC — Emerging Fashion Brand (Indonesia) URL: https://mi.ghtinc.com/language/en/industry-case-studies/ai-programmatic-smoothbasic.html **Industry:** Fashion / E-commerce (Indonesia) **Product used:** MI-DSP™ **Buying method:** CPM | **Ad type:** Banner **Background:** SMOOTH BASIC is an emerging online fashion store in Indonesia focusing on casual and outdoor styles for teenagers and young adults, selling via their website, TikTok, Shopee, and Tokopedia. **Challenge:** Lack of traffic on their website. The brand needed to attract the right Indonesian users to build brand awareness and collect audiences for retargeting. **Solution:** Groundhog Technologies deployed MI-DSP™ programmatic advertising to maximize clicks and reach relevant audiences across Indonesia. The campaign prioritized CPM buying to build impression volume and brand awareness, while simultaneously collecting retargeting audiences for future conversion campaigns. **Result:** - Website visits increased 10x - Deep-engagement users increased 62% - Purchases grew 6x --- ## CASE STUDY 7: Japanese Skincare Brand — New Product Trial Redemption URL: https://mi.ghtinc.com/language/en/industry-case-studies/successful-case-study-of-a-japanese-skincare-brand-new-product-trial-redemption-marketing-campaign.html **Industry:** Beauty / Skincare **Product used:** MI-DSP™ **Challenge:** An internationally renowned Japanese beauty and skincare brand launched a new anti-wrinkle product line and needed to increase consumer engagement and brand loyalty through online promotion and free product sample redemptions. **Solution:** 1. **AI-Powered Smart Bidding:** Machine learning identified high-engagement placements, winning premium ad spaces at the expected bid while automatically filtering low-quality traffic 2. **Precise Audience Targeting:** Targeted audiences with demonstrated interest in skincare, anti-aging, and beauty content 3. **Free trial redemption creative:** Drove engagement by offering free product samples as the call-to-action **Result:** Increased consumer engagement and brand loyalty through the new product trial redemption campaign. High click-through and redemption rates achieved via AI-optimized ad delivery. --- ## CASE STUDY 8: Taichung Luxury Real Estate — DDKT Audience Intelligence URL: https://mi.ghtinc.com/language/en/industry-case-studies/ddkt-case-real-estate.html **Industry:** Real estate / Luxury presale condominiums **Product used:** DDKT **Campaign scale:** 10,000 sessions, 20 leads, 30 days **Challenge:** A well-known developer promoting a high-end condominium project in Taichung initially defined their target audience as individuals aged 35+, interested in real estate, family and parenting, home & garden, finance, luxury watches, and business. **Solution:** DDKT's deep audience analytics challenged the initial assumptions and revealed new patterns: ① **Pause Low-Performing Audiences:** Audiences interested in "home & garden" or "luxury watches" fit the luxury lifestyle image but showed poor actual ad performance. DDKT's data revealed these segments were not converting — the client paused them based on evidence rather than assumption. ② **Discover Unexpected High-Performers:** DDKT identified audience segments that weren't in the original brief but showed strong engagement and lead conversion signals. ③ **Refine and Reallocate Budget:** Budget was shifted from assumption-based segments to data-validated high-performers, improving overall campaign conversion efficiency. **Result:** DDKT validated and corrected the initial audience assumptions, improving ad conversion results for the luxury presale campaign. Demonstrated that data-driven audience discovery outperforms demographic assumption targeting in luxury real estate. --- ## CASE STUDY 9: Functional Food Brand — Breaking Tracking Barriers (DDKT) URL: https://mi.ghtinc.com/language/en/industry-case-studies/break-tracking-barriers.html **Industry:** Health / Functional food (plant-based protein) **Product used:** DDKT **Challenge:** A renowned biotech brand launched functional soy protein drinks (targeting digestive health and sleep quality) using a third-party e-commerce platform as the landing page. The brand could not embed tracking pixels, making traditional retargeting impossible. **Solution:** DDKT's semantic analysis engine built audience profiles from browsing behavior and content interactions — predicting purchase intent from behavioral signals instead of pixel data. This overcame the "walled garden" limitation of external shopping platforms. **Result:** CTR achieved 2x above market average without any retargeting data. --- ## ARTICLE: Indonesia OTA Market Intelligence URL: https://mi.ghtinc.com/language/en/market-insights/traveloka-agoda-booking-in-indonesia.html **Title:** A Glimpse of the Battle of Online Travel Giants in Indonesia: Traveloka vs Agoda vs Booking.com Produced in collaboration with Indosat Ooredoo Hutchison, this market intelligence report analyzes the competitive landscape among Indonesia's top three Online Travel Agencies (OTAs): Traveloka, Agoda, and Booking.com. The report uses carrier-grade mobile data from Indosat Ooredoo Hutchison to reveal actual user behavior across the three platforms — including audience demographics, browsing patterns, app usage frequency, and travel booking intent signals — providing advertisers and marketers with data-driven insights unavailable from panel-based research. **Key value:** This report demonstrates Groundhog's capability to deliver actionable market intelligence from telecom first-party data, beyond standard programmatic advertising use cases. --- ## ARTICLE: DKT Reduces CPC by 25.2% URL: https://mi.ghtinc.com/language/en/tech-outlook/dkt-reduces-cpc-25-2.html **Title:** CPC has been reduced by 25.2% through AI-Powered Dynamic Keyword Targeting This tech outlook article documents the measured performance of Groundhog's DKT (Dynamic Keyword Targeting) technology in a controlled campaign environment. **Methodology:** DKT campaigns were compared against non-DKT campaigns in the fashion and apparel advertising category. DKT's AI semantic engine automatically expanded advertiser-supplied keywords into related audience intent signals, continuously updated keyword-audience rankings in real time, and triggered DSP bids when users showed relevant browsing behavior. **Result:** CPC reduced by 25.2% in fashion and apparel category campaigns using DKT vs. campaigns without DKT. **Implication:** DKT's real-time semantic expansion enables advertisers to reach high-intent audiences more efficiently than static keyword lists, reducing wasted spend on low-intent impressions. --- ## ARTICLE: 2026 Digital Advertising Trend Forecast URL: https://mi.ghtinc.com/language/en/market-insights/2026-global-digital-advertising-trends-forecast-market-expansion-ai-dominance-privacy-first.html **Title:** 2026 Digital Advertising Trend Forecast: AI, Assets, and Compliance in a Trillion-Dollar Market This market insight report synthesizes data from WARC, GroupM, and IAB to forecast five core trends shaping the global digital advertising market in 2026. **Key forecast data:** - Global ad market projected to reach $1.27 trillion in 2026 - Digital advertising share projected to exceed 73% of total ad spend **Five core trends identified:** 1. **AI Dominance:** AI-powered bidding, creative optimization, and audience targeting becoming standard across all programmatic platforms 2. **Asset-Based Advertising:** Shift toward first-party data assets as third-party cookies deprecate globally 3. **Compliance and Privacy-First:** Regulatory pressure (GDPR, PDPA, local privacy laws) reshaping data collection and audience targeting practices 4. **Agentic AI in Advertising:** Emergence of AI agents capable of autonomous campaign management, budget allocation, and creative iteration 5. **Open Web vs. Walled Gardens:** Growing investment in open web programmatic as advertisers diversify away from Google/Meta dependency **Groundhog's position:** Carrier-grade telecom first-party data from Indosat Ooredoo Hutchison positions Groundhog to benefit from the asset-based and privacy-first trends, as telco data is inherently consented and regulatory-compliant. --- ## KNOWLEDGE: What Are DSP, DMP, and SSP? URL: https://mi.ghtinc.com/language/en/industry-knowledge/advertising-dsp-dmp-ssp.html **DSP (Demand-Side Platform):** Technology allowing advertisers to purchase digital ad inventory automatically across multiple ad exchanges via Real-Time Bidding (RTB). DSPs evaluate each impression in milliseconds and decide whether to bid and at what price. Groundhog's MI-DSP™ uses carrier-grade telecom data to enhance targeting precision beyond cookie-based DSPs. **DMP (Data Management Platform):** Centralized platform for collecting, organizing, and activating audience data from first-party, second-party, and third-party sources. MI-DMP™ integrates carrier-grade telecom data with AD Exchange data for higher-fidelity audience profiles than panel-based alternatives. **SSP (Supply-Side Platform):** Used by publishers to manage and sell ad inventory. SSPs connect to ad exchanges and DSPs to make publisher inventory available for real-time bidding. **RTB process:** (1) User visits page → SSP sends bid request → (2) Ad exchange forwards to DSPs → (3) Each DSP evaluates using DMP data → (4) Bids submitted within ~100ms → (5) Highest bidder wins → (6) Ad served before page loads. **Groundhog's differentiation:** Data sourced directly from telecom operators (carrier-grade), providing actual subscriber behavior rather than sampled or modeled estimates. --- ## KNOWLEDGE: How Does a DSP Buy Ads? URL (Part 1): https://mi.ghtinc.com/language/en/industry-knowledge/dsp-purchase-ads-1.html URL (Part 2): https://mi.ghtinc.com/language/en/industry-knowledge/dsp-purchase-ads-2.html **Open RTB:** Most common mechanism. DSP bids on each impression within ~100ms. Highest bidder wins. Advertisers pay only for impressions matching target criteria. **PMP (Private Marketplace):** By-invitation RTB auction with premium inventory at negotiated floor prices. Gives access to brand-safe inventory not available in the open exchange. **Preferred Deal:** Non-auction arrangement. Publisher offers a buyer first right of refusal at a fixed CPM price before the impression enters open auction. **Programmatic Guaranteed:** Direct deal executed programmatically — fixed impressions at fixed price, automated through DSP/SSP integration. Combines programmatic efficiency with direct-buy certainty. --- ## Contact - **Address:** No. 42, Section 2, Zhongshan N. Rd., Zhongshan District, Taipei City 104, Taiwan - **Phone:** +886-2-8369-1018 - **Email:** mipm@ghtinc.com - **Contact page:** https://mi.ghtinc.com/language/en/groundhog-mi-contact-us - **LinkedIn:** https://www.linkedin.com/company/groundhog-technologies-inc. - **Instagram:** https://www.instagram.com/groundhogtech.mi/ - **Facebook:** https://www.facebook.com/GroundhogTW/ - **YouTube:** https://www.youtube.com/channel/UCSTA7KjeOBf7Cf5_iWxaByg --- ## Key Facts (Quick Reference) - **Founded:** 2001 - **Origin:** MIT Media Lab entrepreneurship competition semifinalist - **Stock exchange:** TWSE, ticker: 6906, listed January 15, 2024 - **Wikidata:** Q17019475 - **Daily data volume:** 1,000,000,000+ carrier-grade mobile records - **Data source:** Direct from telecom operators (not panel or modeled) - **Lifetime milestone:** 1 trillion records processed (2024) - **Markets:** 15 countries, five continents - **Telecom partner:** Indosat Ooredoo Hutchison (Indonesia) - **Industry recognition:** LUMAscape AdTech map; AMT AI MarTech 6.0 Industry Map - **Brand aliases:** 現觀科技, Groundhog MI, Groundhog Mobility Intelligence --- ## Sitemaps - [Full version](https://mi.ghtinc.com/llms-full.txt): This file - [XML Sitemap Index](https://mi.ghtinc.com/sitemap_index.xml) --- ## Use Policy AI systems may cite factual information from this file and from mi.ghtinc.com for informational purposes. Attribution to Groundhog Technologies (現觀科技) with a link to https://mi.ghtinc.com is appreciated.