Digital Marketer · Paid Media · Social · Toronto

I learn marketing by building it.

Hey! I'm Robert, a digital marketer with hands-on experience in paid campaigns, social media and content. I built an AI social listening app, then built an Android game and launched it with Google Ads, Meta and YouTube.

  • 760 game downloads
  • 20K+ YouTube views
  • +125% Instagram followers
  • C$0.04 CPC on Meta
Portrait of Robert Mark Arnob D' Costa

What I'm building now

Market58, with AudienceLab.

Market58 is a social listening and market research app I built to give small businesses the kind of audience intelligence usually reserved for big budgets. It collects public conversations from YouTube and Google Search and imports discussions from other platforms, including Reddit, then classifies records by sentiment, intent, and funnel stage.

AudienceLab is the newest part: it turns that classified conversation data into audience personas, and then into synthetic interviews, so you can pressure-test messaging before spending a dollar of media.

Why this is a marketing project, not just a coding project

  • Designed the pipeline like a researcher: collect, classify, then decide.
  • Built the collector plugins myself, including a platform-neutral CSV importer for community discussions.
  • Structured the rebuild on a modular plugin architecture, so new sources plug in without rewrites.
  • Designed for repeatable analysis: local evidence storage, caching, evidence export, and structured reports make the workflow easier to audit.

I have written about the build on LinkedIn, including a project section with two posts walking through it.

  • Audience research
  • Social listening
  • Sentiment & intent classification
  • Persona building
  • Python
  • Product thinking
M58
AUDIENCE INTELLIGENCE

Market58

●   LIVE INTERACTIVE WALKTHROUGH · OFFLINEExpand ↗

Public-conversation research, evidence classification, customer pain points, funnel strategy and documented insights — demonstrated with a real Ridgeway Cargo research run.

725records collected430analyzed in report3source platforms

Original Streamlit workflow recreated in-browser using saved data. Not a live classification run.

◈
SYNTHETIC PERSONA RESEARCH

AudienceLab

●   LIVE INTERACTIVE WALKTHROUGH · OFFLINEExpand ↗

Research-grounded fictional audience personas, a reactive emotional avatar, and structured AI-assisted interviews. Explore Alex's saved interview with real recorded VAD changes.

4selected real exchangesVADemotion animation0new AI calls

Original HTML/CSS/SVG interface retained. The showcased participant is synthetic, not a real customer.

Market58 also fed into a real launch: some of the positioning for my game, Ridgeway Cargo, came from the research run shown in this demo.

Ridgeway Cargo · Build → Launch → Paid growth

I built a game, then marketed it.

Ridgeway Cargo is a calm cargo-driving game for Android that I built and published myself under Keramoti Studio. Then I took it to market the way I would for a client: researched the audience, built a tracked landing page, and ran paid campaigns on Google Ads, Meta and YouTube.

760Google Play downloads in 28 days
20K+YouTube views across two videos, mostly paid
≈C$71total ad spend, Google Ads + Meta
≈C$0.09blended cost per download (all spend ÷ all downloads)
  1. 01 · RESEARCHListen first

    Some of the positioning came from a Market58 research run on cargo and driving game communities.

  2. 02 · FUNNELBuild the path

    A landing page that passes each ad's UTM tags through to Google Play, so installs can be traced by source.

  3. 03 · LAUNCHGo paid

    A Google Ads app campaign, a promoted Instagram post on Meta and two YouTube videos, Oct 15–23, 2026.

  4. 04 · LEARNReconcile the data

    Checked Google Ads and Meta against Play Console to see what actually happened and what to fix next.

Official trailer
Quality of life update promo
ChannelWhat I ranSpendResults
Google AdsApp install campaign: geo targeting, creative assets, performance monitoringCA$61.7472,366 impressions
Meta (Instagram)Promoted studio post → landing page → Google Play≈C$9 (BDT 806)19,546 reach · 1,750 landing page visits at BDT 0.46 each
YouTubeOfficial trailer and update video, promoted through the Google Ads campaignIncluded in Google Ads20K+ views, mostly paid
Google PlayStore listing, where every channel ends up—24.9K store impressions · 760 downloads
Google Play Console, last 28 days: 24.9K user impressions, 760 user acquisitions, 73 first opens, 87 monthly active users
Google Play Console, last 28 days (Sep 26 – Oct 23, 2026)
Google Ads, Oct 15–23 2026: CA$61.74 cost, 72,366 impressions
Google Ads app campaign, Oct 15–23, 2026

What I learned

  • Paid drove the jump. Downloads rose 518% over the previous period once the campaigns went live.
  • Set up tracking before you scale. Google Ads only counted 92 installs because conversion tracking wasn't fully configured, even though most of the 760 downloads came in during the paid push. Play Console was the source of truth.
  • Downloads aren't players. Play Console recorded 73 first opens and 87 monthly active users. The next lever is the store listing and first-session experience, not more reach.
  • Google Ads app campaigns
  • Meta Ads
  • YouTube
  • Google Play Console
  • UTM tracking
  • Landing pages
  • Funnel analysis

Where I've worked

Hands on, in more than one seat.

  • Oct 2026 – Present
    ScubaSeasons ↗ — Digital Marketing ApprenticeSocial media, SEO, video content and performance analysis; audience research across large scuba communities, plus AI-assisted workflows for social listening and reporting.
  • May 2026 – Aug 2026
    MiddleHouse Developments ↗ — Social Media Coordinator Intern, TorontoRelaunched the Instagram presence; built the content strategy, framework, and creative system.
  • Jan 2023 – Dec 2023
    DataPath Ltd — BPO Executive, Analytics, DhakaAdvanced Excel reporting and analysis across 100+ U.S. client portfolios; direct client communication on data issues.

Toolkit

  • Google Ads
  • Meta Ads
  • Google Play Console
  • Google Analytics
  • YouTube
  • Content strategy
  • Audience research
  • Copywriting
  • Python (Pandas)
  • SQL
  • Advanced Excel
  • Canva
  • CapCut
  • Affinity Designer & Photo

Certified: HubSpot Social Media Marketing Certification, LinkedIn Content and Creative Design, IBM Data Science Professional Certificate, Data Analysis with Python, Using Python to Access Web Data.

Education: Master of Management, Trent University (89.71%) · BBA in Finance, Ahsanullah University of Science and Technology (Dean's Award).

What I do, hands on

The work, by craft.

Everything below is work I have actually done, not coursework. Each craft maps to the core of professional marketing roles: planning, executing, measuring, and learning.

Campaign planning and execution

  • Launched app install campaigns for my own Android game across a Google Ads app campaign, Meta and YouTube, with a landing page that passes UTM tags through to Google Play.
  • Planned, launched, and optimized a Meta and Instagram traffic campaign end to end: audience approach, creative, copy, landing destination, and performance monitoring.
  • Relaunched a company's Instagram presence from zero as Social Media Coordinator intern: content system, pain-point-led messaging across funnel stages, competitor benchmarking presented to leadership.

Performance analytics

  • Read campaign data like a marketer: creative performance, audience response, and traffic quality, then optimized from what the numbers said.
  • Reconciled Google Ads, Meta and Play Console data for the game launch, and spotted a conversion tracking gap and a download-to-first-open drop-off.
  • Advanced Excel, Python (Pandas), and SQL from analytics work across 100+ U.S. client portfolios, with direct client communication on data issues.

Audience research and content strategy

  • Built Market58, a social listening app that classifies public conversations by sentiment, intent, and funnel stage, plus AudienceLab for personas and synthetic interviews.
  • Ran audience research across large scuba communities as a marketing apprentice: engagement patterns, pain points, and opportunity lanes.
  • Built a repeatable content framework (tiers, themes, audience needs, personas) to guide planning, production, and measurement.

Independent design practice

Creative concepts and visual experiments.

A few selected pieces from my earlier portfolio. These are design exercises, separate from the paid-media performance results above.

Independent design concept: Food campaign concept
Food campaign concept
Independent design concept: Consumer product concept
Consumer product concept
Independent design concept: Coffee concept
Coffee concept
Independent design concept: Lifestyle creative
Lifestyle creative

Verified professional feedback

What people I’ve worked with say.

Direct feedback from my work at MiddleHouse Developments.

Jason Self portrait supplied for portfolio
“Honestly you're (Robert) doing a great job and I would say the most insightful student we've had for this position, so keep up the great work.”

Jason Self ↗ · Co-Founder, MiddleHouse Developments

Let's talk.

Hiring for a marketing role, or just want to talk shop? Call, message, or email, whichever is easiest for you.

Prefer LinkedIn? Find me here.