E.C. / 陳富祥
中文
00 — Full-Stack Marketing LeaderTaipei, Taiwan · UTC+8

Digital Marketing × Data Strategy × AI & Automation

Integrating all three to lead a brand’s marketing through modern transformation.

I was an appointed digital media instructor for Google Taiwan, and cases I led have won a Google Premier Partner Award and the Best Data Partner award at Meta’s Agency First Awards. I’ve led both advertising and data teams, turning marketing strategy into systems that can be repeated and governed — and I build MarTech tools hands-on and drive AI adoption across the business.

Portrait of Edward Chen

Edward Chen陳富祥

Full-Stack Marketing Leader · 10 years in digital marketing · Associate Client Success Director, adGeek

Digital marketingcumulative ad spend
NT$300M+
Data strategyclients using data & ML applications
30+
AI & automation~7,000 h of routine work freed / yr
~3.5FTE
Org transformationcolleagues in AI training
100+
01 — What I bringFive layers, one helmsman

Companies aren't short of AI tools. They're short of someone who can steer marketing, data and AI as one.

Break a marketing organisation into its five layers, and my results and impact show in every one: spend at scale, official awards, teams I've developed, and systems still running today.

01

Strategy & media

From spend at scale to industry awards — media strategy that's been proven.

4+official Google and Meta awards
  • 10 years across Google Ads, YouTube, Meta, TikTok, LINE, DV360, Criteo, Taboola, Apple Search Ads and more
  • NT$300M+ in cumulative ad spend — including 60+ clients and NT$180M in 2016–18 alone
  • 2019 Google Premier Partner Awards, App Excellence (Hong Kong & Taiwan): the only Taiwan agency to win — I ran optimisation and owned the entry, and led the company to three consecutive years as a finalist
  • Meta Agency First Awards, Best Data Partner: a data-driven case for a leading health-supplement brand — I led the entry and ran the work
Google Premier Partner Awards 2019 ceremony, with the App Excellence winner on screenPremier Partner Awards 2019 trophy
Premier Partner Awards 2019 — ceremony and trophy
02

Data & insight

Turn data into decisions brands can actually act on.

20+brands and companies trained in analytics
  • Smart Value: an ML model estimating each user's purchase potential, applied for 30+ clients — I own the brand-side problem definition, application assessment and communication
  • As PM, launched AIA (Insight Audience): 432 ready-to-activate industry audience lists (9 industries × 12 behaviours × 4 tiers), using data to push past the ceiling of ad optimisation
  • Partnered with Google Taiwan to bring Ads Data Hub in-house — joining GA and ad data in BigQuery for more precise, more complete measurement of advertising impact
  • GA4 instructor and trainer: Google Analytics courses for 20+ brands and companies — joint vendor classes, in-house sessions and online courses — plus 7 internal instructors trained to extend the reach
Smart Value: from customer data to media decisions
03

Team & organisation

Extend impact across the organisation and beyond.

140clients shaped by the data and AI product direction I lead
  • As head of the data team, lead the company's data and AI product direction, setting goals and owning delivery; responsible for hiring and performance, and mobilise technical resources across departments outside my reporting line
  • Grew from specialist to leading a 10-person advertising team across media strategy, service-quality control and talent development
  • Led the marketing team to Google 2019 Digital Guru Awards: Branding Star and Performance Star
  • Marketing instructor: former appointed digital media instructor for Google Taiwan; taught financial professionals at the Securities and Futures Institute; long-running corporate training for leading brands in retail, health supplements, FMCG and healthcare
  • Invited instructor for a national trade promotion organisation's AI Decision Leadership programme for senior managers — 763 learner sessions online
  • 10+ internal AI sessions bringing 100+ colleagues into AI adoption; authored the company's AI application map; the work was featured in Brain (動腦)
Edward Chen giving an invited talkGoogle 2019 Digital Guru Awards: Branding Star and Performance Star trophiesBrain magazine feature on adGeek's AI application map
Invited talk · Digital Guru awards · the AI application map in Brain, issue 585
04

Value & measurement

Use AI to open new business opportunities.

NT$10M-scaleadded agency revenue capacity a year
  • AI and automated MarTech tools free ~7,000 hours of the team's routine work a year (~3.5 FTE), so each operator can take on 2–3 more clients — NT$10M-scale in added agency revenue capacity
  • Built an ad-creative content tool combining creative generation with rule-based control — overall outsourcing spend down 50–60%, NT$500K+ saved a year
~7,000 hours freed a year ≈ 3.5 FTE
05

AI & automation

Put AI to work — and amplify the value of every layer.

5marketing automation systems
  • Built 5 automation systems covering four stages of digital advertising work: data insight, content creation, risk monitoring and optimisation
  • Built the marketing-monitoring software Ad Monitor — prototype and first production release — on my own (planning and development); 12 of its 14 monitoring rules come from my own media practice
  • Designed from the user's needs: MarTech tools with clear, guided, low-barrier design — full operator adoption, with almost no training needed
Recreated Ad Monitor console
Ad Monitor console (recreated, fictional data)
02 — Client & strategy casesMarketing problems, solved

Before any system, there was the work: client problems solved with strategy and data.

Selected engagements. Client names withheld.

Health supplements · Data strategy · adGeek · Meta Agency First Awards — Best Data Partner

Under half the budget — and new-customer traffic up 893%

Chinese New Year is the biggest gifting season of the year, yet the campaign budget was 64.8% below last year. A high-price, high-trust, high-consideration category can't be carried by ordinary retargeting.

+893%new-customer traffic
+56.7%ROAS vs the same season last year
121ROAS of the high-value segment
8+new audiences at ROAS 10+

Campaign strategy

FindWiden quality traffic

ASC lets Meta AI reach new customers from conversion signals

SortIdentify likely buyers

A DMP splits site behaviour into intent segments — new members, hesitant first-timers, high-intent non-buyers and more

RankPrioritise delivery

Smart Value scores audiences so budget reaches high-potential buyers first

My roleLed the proposal and ran the campaign·privacy-first CAPI integration, with an event match quality of 8.2 (Meta recommends 6)

More cases

Healthcare retail · Lead genadGeek

Connecting lead data so a multiplied budget still held its cost per lead

Budget grew several-fold while cost per lead stayed below the same period in prior years.

Challenge
The budget scaled several-fold just as the market hit a plateau — the off-season could no longer match past results, and fraudulent and duplicate leads piled up.
What I did
Connected the data so real form-fill outcomes fed back into ad delivery; sharpened creative topic strategy to build more compelling assets; and restructured the whole account architecture.
Entertainment · AudienceadGeek

Artist-level audience strategy for a global music group

Discovery time −50%+ · new-audience CTR +15% · cost −20%

Challenge
Find new listeners for every artist, fast — with different problems by artist tier: building awareness for new acts, opening emerging markets, finding distinctive fan communities.
What I did
Combined web data, historical data and AI analysis into a three-tier audience forecast: core, potential and look-alike positioning audiences.
Public sector & beauty · DataadGeek

CRM BI: turning customer data into decisions

7–10 actionable marketing guidelines per client — from long-term brand adjustments to behavioural recommendations with immediate impact on short-term campaigns.

Challenge
Rich CRM data, but no clear read on what it meant for marketing or merchandising.
What I did
My team built Power BI and Tableau dashboards on the CRM data for a national trade promotion organisation and an international beauty group; I read what the numbers meant for the business.
Consumer tech · GlobalEFUN

Taking a global laptop brand overseas

Helped a product launch set the brand's record for concurrent live-stream viewers; later scaled into integrated service across its global markets.

Challenge
Grow the brand's media work beyond Taiwan.
What I did
Expanded its overseas media work; introduced data-driven creative, rich media and live-stream ad formats — which also won the company its first technical partnership with a 4A agency.
03 — SystemsHow I scale marketing value

Embracing automation and AI, I turned that digital experience into MarTech tools and systems the whole team uses every day.

Each one starts from a marketing problem, not from the technology: containing budget risk, speeding up insight and scaling content output — taking analysis and insight into automation and applied AI.

~7,000h / yr

Routine work removed across a 20-person operations team — about 3.5 FTE.

2–3clients

Extra clients each operator can serve with the time the tools create — NT$10M-scale in added agency revenue capacity a year.

−50–60%

Overall outsourcing spend — NT$500K+ saved a year.

Ad Monitor

Protect the budgetLive since Jun 2026

Problem

20 operators checked 340 Google and Meta accounts by hand, most items 3+ times a week. A missed runaway budget or rejected ad meant five-figure NT$ losses.

My role

Product owner and lead developer: built the prototype and first production release myself; 12 of the 14 rules come from my own media practice.

Result

  • ~5,000 h/yr freed (~2.5 FTE)
  • 100% operator adoption: clear, guided design lowers the barrier, so people pick it up fast with almost no training
  • Automated checks 7+ times a day, monitoring accounts in near real time
  • Tiered alerts via web, email and Telegram
ad-monitor / overview
OverviewToday · 14:00 run
Google 158Meta 182
Urgent3
Alert12
Notice27
Healthy298
AccountPlatformSignalTier7-day
Client A — SearchGooglepacing 186%URGENT
Client B — Advantage+Meta0 conv / 24hURGENT
Client C — YouTubeGoogle3 disapprovedALERT
Client D — CatalogMetafrequency 6.2NOTICE
Client E — BrandGoogleend date passedNOTICE
Ad Monitor · web consoleTiered by risk, not by volume

Weekly Ad Insight System

Insight clients can act onFinding · in use since 2025

Problem

Every week, hours of manual work went into analysing ad performance — then condensing it into insight clients could understand and act on.

My role

Led the team in defining requirements and use cases, and built a cross-media analysis tool that turns raw ad data into insight in one pass — fast, and at volume.

Result

  • 150+ clients covered · 50+ reports a week produced by the tool
  • 80% of output ready to deliver; only parts need a human edit
  • A global consumer-electronics client with many campaign flights: weekly insight 3 hours → 30 minutes
Finding · Weekly insight · Client F · W38● Client-ready
Creative refresh paid off on Meta; Search is carrying brand demand.
  • MetaCPA down 18% week on week after the refresh. Move 10% of prospecting budget to the two winning concepts.
  • GoogleBrand search volume up 22% — the offline launch is landing here. Keep brand budgets uncapped this week.
  • LINEReach steady, CTR below benchmark. Test a shorter headline before adding budget.
CPA · Meta · 6 weeksNT$
−18%week on week
W33W34W35W36W37W38

GenAI Monthly Report

Automated analysis & insightIn production

Problem

Wrap-up decks needed analysis written page by page — slow, and hard to standardise.

My role

Driver: requirements, use-case design and the solution path for the hard technical problems; engineered by the team.

Result

  • −50% wrap-up time
  • ~2 h saved per client per month
  • Live for 3 of 4 client tiers
P.4 · Performance overview

Original deck: charts only

P.4 · Performance overview
AI analysisCPA fell 12% month on month, driven mainly by wider retargeting audiences; next month, shift 15% of budget to the two best-performing creatives.
Reads each page of an existing deck and writes analysis that fits its context · illustrative

Copy Incubator

Content capacity & costLive since late 2024

Problem

Copy for 100+ clients leaned on outsourcing, with 2–3 day turnarounds on rush jobs.

My role

Architect: core architecture, most of the spec, and the exact character-count validation problem.

Result

  • 1.5× output, same headcount
  • −50–60% outsourcing spend
  • Rush turnaround 2–3 → 1–1.5 days
Brief
→
AI generation
→
Rule checks
  • ✓Exact length
  • ✓Banned terms
  • ✓Brand voice
  • ✓Platform specs
→
Ready to use

↺ Editor feedback keeps refining the rules

Creative generation with rule-based control · illustrative

AI Agent Optimisation

Human-in-the-loop optimisationIn development · early wins

Problem

Ad optimisation depends heavily on individual experience; brand requirements, past decisions and their reasons are scattered across people, hard to carry forward or repeat.

My role

Leading development: designed the full loop from analysis and recommendation to optimisation and record-keeping, plus the human sign-off step.

Result

  • Automated analyse → recommend → optimise
  • Records brand needs, decisions and reasons and background
  • Early success cases already in hand
01Analyse

Reads account data, flags anomalies and openings

02Recommend

Prioritises against the brand's goals

03Human sign-off

A colleague approves, adjusts or rejects

04Optimise

Executes, then tracks the result

Brand memory

Brand needs · decisions & reasons · background

↺ Every decision is recorded and feeds the next round of analysis

Human-in-the-loop optimisation · illustrative
04 — CareerFrom specialist to helmsman

Ten years, every layer of the job.

Each role added a layer: hands-on media, then accounts, then a team, then data and AI across a whole company.

2026.05 —

Associate Client Success DirectoradGeek · Taipei

Head of the data team: lead the company's data and AI product direction across ~140 active clients, setting goals, owning delivery and mobilising technical resources across departments.

2023.03 — 2026.04

Senior Client Success ManageradGeek

Moved deliberately from media operations into data, MarTech and AI; built the product line and the company-wide AI adoption programme.

2020.06 — 2023.02

Advertising Team LeadEFUN Digital Technology

Led a 10-person team across media strategy, service quality, talent development and MarTech projects — including Ads Data Hub with Google Taiwan and the AdHero platform. NT$300M+ cumulative spend across the EFUN years.

2018.03 — 2020.05

Account SupervisorEFUN Digital Technology

Ran optimisation and owned the entry for the 2019 Google Premier Partner Awards App Excellence winner (Hong Kong & Taiwan) — the only Taiwan agency to win; led the team to Google 2019 Digital Guru Awards: Branding Star and Performance Star.

2016.03 — 2018.02

Google Ads Account SpecialistEFUN Digital Technology

60+ clients and NT$180M cumulative spend across brand, gaming and e-commerce — Google Ads, Yahoo, Apple Search Ads and DV360.

Edward Chen speaking on stage
Fig. — Speaking & trainingFormer appointed instructor for Google Taiwan · Securities and Futures Institute · corporate training for leading brands
National Cheng Kung University · B.S. Electrical EngineeringLanguages · Mandarin (native) · English (professional working, TOEIC 910)
05 — PracticeHow I lead
i.

Start from the business problem.

Budget, growth, capacity or risk come first. The tool is chosen last.

ii.

Adoption is the only review that counts.

Rules come from the team's real day; full use with no training is how I know they were right.

iii.

Label estimates as estimates.

Every number carries its method. If the method changes, the number changes — publicly.

06 — ContactMandarin · English

Let’s talk.

EmailShow email address LinkedIn/in/edward-chen CVDownload PDF ↓