Executive Summary
The UTM Campaign Builder helps campaign managers create consistent tracking links, configure advertising-platform templates, build promotional URLs, and process campaign ideas in bulk with AI assistance.
It turns repetitive campaign setup into a guided workflow with validation, recommendations, copy-ready outputs, saved history, and bulk processing. The product is a campaign launch-readiness and governance layer—not an ad-buying platform.
My role: Product strategy, requirements definition, workflow design, validation logic, AI guardrails, and portfolio narrative.
Naming note: The prototype uses the working name “Metabase.” This independent portfolio project is not affiliated with Metabase, Inc.
Current working prototype
- Single-campaign UTM generation
- Rule-based warnings for selected GA4 source and medium conventions
- Google Ads and Microsoft Ads tracking-template support
- Sitelink and custom-parameter helpers
- Promotional campaign URL generation
- AI-assisted bulk campaign processing
- CSV and Google Ads Editor exports
- Optional campaign history when PostgreSQL is configured and healthy
- GTM-compatible event tracking
- An illustrative performance preview—not a live media forecast
Business Problem
| Pain point | Operational reality |
|---|---|
| Inconsistent naming | Marketers use different capitalization, abbreviations, and separators. |
| Incorrect UTMs | Required values may be missing, reserved, spaced, or inconsistently cased. |
| GA4 misclassification | Source and medium combinations can route traffic into unintended channels. |
| Google Ads setup errors | Templates, ValueTrack parameters, suffixes, and sitelinks are easy to misconfigure. |
| Manual operations | High-volume naming, copy, and exports create repetitive work and copy/paste errors. |
| Low launch visibility | Teams cannot easily see whether a package is ready or still needs review. |
Business impact
- Unreliable channel and campaign reporting
- Slower campaign launches
- More rework across marketing, analytics, and media teams
- Inconsistent attribution across platforms
- Reduced confidence in performance data
- Greater risk of publishing assets that require formatting or policy review
Product Vision
Make every campaign measurable, reviewable, and ready to launch without requiring every marketer to be a tracking expert.
The product direction is a standard workspace for translating campaign intent into governed tracking assets and campaign packages prepared for final review.
Business Objectives
Primary
- Reduce time required to create a standards-aligned campaign tracking package.
- Improve consistency of UTM naming and parameter formatting.
- Reduce preventable Google Ads tracking and upload errors.
- Give marketers an easy review step before activation.
- Support both one-off and high-volume campaign production.
- Make campaign setup knowledge repeatable instead of expert-dependent.
Secondary
- Create a reusable internal source of campaign history.
- Build an extensible foundation for future ad-platform and analytics integrations.
- Demonstrate practical, human-reviewed AI assistance—not autonomous publishing.
Success Metrics
North-star outcome
Percentage of campaign packages that meet defined readiness checks on the first review.
These are recommended production metrics. The prototype contains GTM event utilities but not a complete reporting dashboard.
| Metric | Definition | Initial target |
|---|---|---|
| Campaign setup time | Opening the tool to copying a valid package | Reduce by 50% |
| First-pass validity | Campaigns passing validation without correction | 90%+ |
| Naming compliance | Saved campaigns following the agreed format | 95%+ |
| Launch rework rate | Campaigns returned for tracking or format fixes | Reduce by 40% |
| Bulk processing adoption | Eligible multi-campaign requests using bulk workflow | 30%+ |
| AI acceptance rate | AI output accepted with minor edits | 70%+ |
| Export success rate | Export attempts producing a usable file | 99% |
| Data quality incidents | Missing, malformed, or misclassified tracking data | Downward trend |
| Review-to-launch time | Package completion to media activation | Establish baseline |
Target Users and Personas
Campaign Manager
Needs: Fast setup, clear options, copy-ready URLs, and confidence that required fields are complete.
Workflow: Choose campaign type → enter destination → select channel → accept suggestions → copy or save.
Performance Marketing Specialist
Needs: Platform templates, ValueTrack parameters, sitelinks, custom parameters, and exports.
Workflow: Configure paid-media tracking without manually rebuilding platform syntax.
Marketing Operations or Analytics Lead
Needs: Enforced standards, validation feedback, history, auditability, and adoption metrics.
Workflow: Govern taxonomy and identify tracking issues before campaigns launch.
Content or Creative Strategist
Needs: Campaign briefs translated into names, ad copy, sitelinks, and structured exports.
Workflow: Scale campaign concepts across products and audiences.
Reviewer or Approver
Needs: Format and policy-review warnings, generated outputs, and a record of what was approved.
Workflow: Review drafts now; use status, comments, and version history in a future release.
Business Scope
In scope
- UTM creation for organic, paid, and promotional traffic
- Source and medium recommendations plus campaign naming suggestions
- GA4-oriented validation and warnings
- Google Ads and Microsoft Ads tracking guidance
- ValueTrack templates, final URL suffixes, sitelinks, and custom parameters
- Promotional campaign URL generation
- AI-assisted bulk campaign, ad-copy, and sitelink suggestions
- CSV and Google Ads Editor export
- Campaign history retrieval and basic GTM-compatible usage events
Current out of scope
- Buying or publishing ads directly
- Live synchronization with Google Ads, Microsoft Ads, Meta, LinkedIn, or StackAdapt
- Live campaign performance reporting or attribution modeling
- Budget pacing and bid optimization
- Identity, approval, or role management in the current prototype
- Guaranteed AI accuracy or autonomous publishing
- Replacing GA4, a CRM, an ad platform, or a marketing automation system
Business Process
Current state
- Receive a campaign brief.
- Manually choose source, medium, and campaign name.
- Construct UTM links and platform templates.
- Send the setup to analytics or media for review.
- Correct errors through email, chat, or spreadsheets.
- Launch the campaign.
- Discover some reporting issues later.
Proposed future state
- Open a single or bulk campaign workflow.
- Enter campaign context and destination URL.
- Receive platform-specific source, medium, and naming recommendations.
- Resolve formatting, taxonomy, and GA4 risks immediately.
- Generate URLs, templates, suffixes, sitelinks, copy, and exports.
- Review warnings and edit the output.
- Copy or export the approved package into the media workflow.
- Store the campaign record and usage events for future reference.
Value Proposition
Speed
Less manual URL and tracking-template construction.
Quality
Validation catches common errors before launch.
Enablement
Guidance makes expert practices accessible to more marketers.
Scale
Bulk processing turns repeated setup into a structured, reviewable workflow.
AI Product Strategy
AI co-pilot role
- Generate campaign names from the landing page, campaign type, and source.
- Generate Google Ads headlines and descriptions.
- Generate contextual sitelinks.
- Return a structured campaign package for review.
Human responsibility
- Product and offer accuracy
- Legal and regulatory review
- Brand voice
- Landing-page alignment
- Budget and targeting decisions
- Final platform activation
AI quality guardrails
- Validate URLs before processing.
- Restrict bulk input to 50 campaigns per request.
- Enforce Google Ads length limits in the response path.
- Show format, taxonomy, and policy-review warnings instead of hiding them.
- Start exported campaigns in a paused review state.
- Provide deterministic fallback content when AI processing fails.
- Never represent simulated performance as observed performance.
| Dimension | Evaluation question |
|---|---|
| Relevance | Does the output reflect the landing page and campaign type? |
| Compliance | Are character limits and required fields respected? |
| Brand safety | Does the copy avoid unsupported claims or risky language? |
| Naming quality | Does the name follow taxonomy and remain readable? |
| Edit distance | How much manual editing is required before approval? |
| Consistency | Do similar inputs produce comparable structures? |
| Failure behavior | Does the tool fail clearly and preserve the user's work? |
Stakeholders and Responsibilities
| Stakeholder | Responsibility |
|---|---|
| Marketing Operations | Own taxonomy, process, adoption, and governance |
| Performance Marketing | Validate platform templates and exports |
| Analytics | Own GA4 channel definitions and reporting quality |
| Brand / Legal | Review AI copy and promotional claims |
| Product Manager | Prioritize roadmap, define success, and manage tradeoffs |
| Engineering | Maintain the application, data, integrations, and reliability |
| Campaign users | Provide accurate inputs and review outputs |
Risks and Mitigations
| Risk | Impact | Mitigation |
|---|---|---|
| AI creates inaccurate copy | Brand or policy risk | Draft labels, human review, evaluations, and claim checks |
| Users trust simulated metrics | Poor planning decisions | Label the preview as illustrative until real data is connected |
| Taxonomy rules become stale | Reporting inconsistency | Version rules, assign an owner, and review changes |
| Platform syntax changes | Broken templates or imports | Version templates and maintain regression tests |
| No authentication | Campaign-data exposure | Add SSO, roles, and tenant isolation before production |
| Weak audit trail | Unclear asset history | Add versions, approvals, and an activity log |
| Database unavailable | Save and history failure | Add health checks, retries, error states, and monitoring |
| Large AI requests | Cost or latency spikes | Use limits, queues, progress states, caching, and budgets |
| Unsafe destination URLs | User or brand risk | Add domain policies, reputation checks, and allowlists |
Product Roadmap
Phase 1 · Current
Prototype foundation
- Guided UTM builder and validation
- Google Ads tracking helper
- Promotional campaign builder
- Illustrative performance preview
- Bulk AI workflow and CSV exports
- Campaign history
Phase 2
Production readiness
- Database monitoring and resilient error states
- Authentication, roles, and approval statuses
- Version history and audit log
- Automated unit, API, and end-to-end tests
- Real job progress
- AI validation, draft labels, and prompt/version tracking
Phase 3
Connected operations
- Google Ads and Microsoft Ads APIs
- GA4 validation and campaign status sync
- Domain and URL health checks
- Shared taxonomy administration
- Approval workflows and comments
- Workspace adoption reporting
Phase 4
Intelligence layer
- Learn from accepted edits and rejected outputs
- Recommend conventions from historical performance
- Detect duplicate or conflicting campaigns
- Compare planned and actual performance
- Explain campaign-quality scores
Product Decisions and Tradeoffs
UTM Campaign Builder: Designing a governed campaign operations workflow that combines deterministic validation with human-reviewed AI assistance.
ProblemCampaign tracking work was fragmented and error-prone.
InsightThe user needed guided creation and governance—not another analytics dashboard.
SolutionCombine deterministic validation with human-reviewed AI generation.
Product decisionKeep publishing out of scope and make outputs copy-ready.
OutcomeOne workflow supports single and high-volume campaign production.
LearningAI is most useful when bounded by rules, platform constraints, and review.
Supported claims
- Built a working full-stack prototype.
- Translated GA4 and Google Ads rules into an interactive validation workflow.
- Designed a human-in-the-loop bulk generation experience.
- Separated deterministic validation logic from probabilistic AI generation.
- Created structured campaign output and export workflows.
Do not claim without measurement
- Guaranteed attribution accuracy
- Real campaign performance improvement
- Production-grade AI accuracy
- Live platform synchronization
- Actual conversion lift from the simulated preview
Business Acceptance Criteria
- 01A marketer can create a valid campaign package without manually constructing the URL.
- 02The tool explains why a source, medium, and campaign combination may be misclassified.
- 03Google Ads users can copy a tracking template and understand where to place it.
- 04Promotional campaign metadata appears in generated outputs.
- 05A user can process multiple campaigns and export the result.
- 06AI failures do not erase submitted campaign inputs.
- 07Generated suggestions are clearly differentiated from approved campaign assets.
- 08Saved campaign history is available when the database is configured and healthy.