Data Orchestration & Analytics Infrastructure

This project was not about replacing one analytics tool with another. It was about rebuilding trust in the company's data.

Different departments were working hard, yet leadership regularly received conflicting reports because each team defined, tracked, and attributed metrics in its own way. Small inconsistencies in tracking had accumulated over several years until they became an operational problem.

I led the effort to standardize measurement across teams, redesign the tracking framework, automate reporting, and establish a common analytics language. The result was an infrastructure that supported faster decision making, reduced manual work, and created a reliable foundation for future experimentation, personalization, and product analytics.

The retirement of Universal Analytics forced organizations around the world to rethink how they collected and interpreted digital data. For 5Paisa, the deadline exposed problems that had existed long before the migration was announced.

Historical tracking configurations, inconsistent naming conventions, duplicate tags, and different attribution rules had gradually created multiple versions of the truth. Teams were spending valuable time debating numbers instead of solving business problems.

I believed that migrating to GA4 without addressing these underlying issues would simply move inaccurate data into a new platform. The migration therefore became the starting point for a much broader transformation of the company's analytics ecosystem.

The migration to GA4 was the trigger, not the problem. The real issue was that the organization had gradually lost confidence in its own data. The same campaign could produce different numbers depending on which team prepared the report.

Marketing focused on acquisition. SEO measured organic performance. Branding tracked campaign visibility. Product looked at engagement. Sales relied on CRM reports. Each team had built its own reporting process over time, and every process made sense in isolation. The problem appeared when leadership tried to combine them into a single business view.

Before recommending any solution, I wanted to understand how data actually flowed through the organization instead of how everyone believed it worked.

I started by documenting the complete reporting journey. Every major data source, reporting tool, manual spreadsheet, and stakeholder was mapped before discussing implementation.

Area	Observation
Tracking	Legacy tags and inconsistent implementation across properties.
Reporting	Heavy dependence on manual spreadsheets.
Attribution	Different definitions used by different teams.
Leadership	No single source of truth for business reporting.
Automation	Limited reuse across recurring reports.

This exercise helped separate technical issues from process issues. Both needed to be solved together if the migration was going to create lasting value.

After reviewing the existing implementation, I found that inaccurate reporting was rarely caused by one major mistake. It was the combined effect of many small decisions made over several years.

• Tracking standards changed as different teams owned the platform.
• Historical tags remained active long after they were needed.
• Campaign naming conventions were inconsistent.
• Business metrics were interpreted differently across departments.
• Manual reporting increased the possibility of human error.

Solving only the technical migration would not remove these issues. The organization also needed common definitions, shared ownership, and repeatable reporting processes.

I decided that the migration should solve today's problems without creating new ones a year later. Instead of treating GA4 as the final destination, I treated it as the foundation of a cleaner analytics ecosystem.

The solution focused on four parallel tracks. Standardize tracking. Define a common attribution model. Automate recurring reports. Build an architecture that future teams could maintain without starting from scratch.

One of the biggest sources of disagreement came from attribution. Different teams credited the same lead to different channels because they followed different rules. Before implementing anything, I worked towards a common language that every team could use.

Focus Area	Objective
Channel naming	Create consistent campaign taxonomy.
Traffic sources	Remove duplicate classifications.
KPI definitions	Ensure every team reported the same metric.
Ownership	Clearly define responsibility for every channel.

Once everyone agreed on the rules, conversations shifted from questioning numbers to improving business performance.

Reporting consumed valuable time across multiple teams. Similar spreadsheets were created every day even though most of the data already existed inside the analytics stack.

I defined an automated reporting approach using GA4, Google Tag Manager, BigQuery, Python, and Looker. Instead of rebuilding reports repeatedly, teams could access updated dashboards with the latest information already available.

Automation was not only about saving time. It reduced manual errors, improved reporting consistency, and allowed teams to spend more time interpreting insights instead of preparing spreadsheets.

• Build once and make it reusable.
• Keep implementation simple enough for future teams to maintain.
• Reduce manual effort wherever automation is practical.
• Document every major tracking decision.
• Measure quality before measuring quantity.

Once the framework was agreed upon, implementation became a phased exercise instead of a single migration. Existing tags were reviewed, unnecessary configurations were removed, and new tracking standards were introduced before additional reports were built.

I worked closely with engineering and external partners to validate every important event. This reduced the risk of inaccurate reporting reaching business teams after launch.

Alongside the larger analytics initiative, I also contributed to data instrumentation for the Google SPOT section. The same tracking standards, naming conventions, and reporting principles were applied so that new products entered the ecosystem with clean data from the beginning.

Dashboards built for this initiative gave stakeholders immediate visibility into performance without creating a separate reporting process.

This project required alignment across product, engineering, acquisition, SEO, branding, marketing, analytics, and sales. Every group depended on the same data but used it differently.

My role was not only to define the product requirements but also to help teams agree on common definitions, review implementation decisions, and ensure the final reporting reflected the same business reality for everyone.

Goal	Expected Outcome
Reporting accuracy	One trusted source of business metrics.
Operational efficiency	Replace repetitive manual reporting with automation.
Cross team alignment	Consistent attribution across departments.
Scalability	Framework reusable for future products and campaigns.

The biggest outcome of this initiative was confidence. Teams no longer spent meetings arguing about which report was correct. They could focus on understanding why numbers changed and what action should be taken next.

Standardized attribution reduced reporting conflicts across business functions. Automated reporting removed more than two hours of manual work every day and gave stakeholders faster access to reliable information. The same framework also became the foundation for future experimentation, personalization, and product analytics initiatives.

Outcome	Business Value
Unified reporting	Single source of truth across teams.
Automation	Saved more than two hours of manual reporting each day.
Data quality	Consistent attribution and KPI definitions.
Scalability	Framework reused for future products and campaigns.

Treat migration as an opportunity

Rebuilding the analytics foundation during the GA4 migration solved problems that had existed for years instead of carrying them into a new platform.

Standardize before automating

Automating inconsistent reports would only have produced inaccurate results faster. Shared definitions came first. Automation followed.

Design for future teams

Documentation, reusable tracking patterns, and common naming conventions were just as important as the implementation itself. The system needed to remain understandable long after the project ended.

This project reinforced an important lesson for me. Analytics is not only about technology. It is also about creating shared understanding across an organization. Even the best dashboard has little value if different teams interpret the same metric differently.

Looking back, the most valuable outcome was not the migration itself. It was creating an analytics ecosystem that could support better product decisions for years to come.

Today I would extend the platform with automated data quality monitoring, anomaly detection, semantic metric definitions, and AI assisted root cause analysis so that unusual trends could be investigated much faster.

I would also introduce self service analytics that allows business teams to answer common questions without depending on manual report requests from product or analytics teams.