Building Data Frameworks That Bridge OKRs & Execution: Lessons from Trendyol Tech

Ideas are easy. Execution is everything — John Doerr

Introduction: Why Strategy Stalls

According to Harvard Business Review research and the findings of Kaplan and Norton, the creators of the Balanced Scorecard, there is a striking truth: 90% of companies worldwide fail to execute their strategies. In other words, the proportion of those who can translate those fantastic OKRs into operations is only 10%

But why? Because while strategy is often soaring in the clouds, engineering or sales teams are down on the ground, “dealing with the mud.” The bridge between them is usually missing or broken. If OKRs are nothing more than static figures on a spreadsheet, they cannot influence the daily decisions teams make.

This is exactly where “Data Frameworks” come into play — turning strategy into the fuel that powers operations.

Why Do We Need a Data Framework ?

Companies often fall into three fundamental traps while striving to become “data-driven”:

• Metric Overload, Meaning Deficit: To measure everything is to measure nothing. An abundance of data obscures focus and creates paralysis by analysis.
• Erosion of Trust: If data exists but no one trusts it, that data is nothing more than a liability. When teams encounter conflicting data points, they inevitably revert to following their gut instincts.
• Walking in the Dark (Missing Data): When critical information is missing at the moment of decision-making, you are condemned to rely solely on intuition.

The cost of these mistakes is high: Internal misalignment, sluggish decision-making processes, and roadmaps shaped by personal opinions rather than data-backed insights.

Hero of the Solution: Product Operations (Product Ops)

Who will resolve this chaos? Product Ops is uniquely positioned to tackle this challenge because of:

• A Broad Perspective: They possess a high-level view of the organization. They observe how data flows across the entire company and how different teams interact with one another.
• The Power of Influence: They can bridge the gap between leadership and execution teams. They have the necessary influence to redesign processes and convince teams to adopt a new data order.

Building a Robust Data Framework in 4 Steps

To transform data from a mere report into a decision-support mechanism, we must follow this process:

1. Define What to Measure

Instead of asking “Which metric should we look at?”, ask “Which business question are we trying to answer?”. Start with a clear question, such as: “What is the most automatable workload currently consuming the IT team’s time?”

2. Design the Data Collection and Processing Layer

This is the “technical kitchen” of the operation. Data should be pulled automatically via APIs or system logs rather than through manual spreadsheets. Treat this system like a product and start with an MVP (Minimum Viable Product).

3. Establish Data Quality and Trust

If no one trusts the data, the system fails. Establish joint QA (Quality Assurance) processes with data teams and create transparent dashboards. Organize “open door” sessions to address team inquiries directly.

4. Operationalize Insights

This is the stage where you “reap the rewards” of the system; data now translates into OKRs and roadmaps.

• Organizational Rhythm: Turn data review into a routine habit.
• Course Correction: Develop the agility to pivot plans based on data-driven insights.
• Strategic Impact: Utilize core insights to shape the company’s primary objectives.

Case Study: How Trendyol’s Tech Resolution Team Conquered Recurring Password Issues

To bridge this gap, we rely on a practical 4-step framework — Define, Design, Control, and Operationalize — and let’s break down exactly how it works in action through a real-world case study from Trendyol’s Tech Resolution team.

Mapping the Case Study to the 4-Step Model:

1. Define What to Measure (Metrics)

• Instead of just asking “Why are there so many tickets?”, we got granular with our Slack Channels and Ticket Flows. We measured:
• What percentage of the support tickets hitting the Tech Resolution team were purely “password resets”?
• Which specific systems (Active Directory, VPN or internal apps) showed the highest concentration of issues?
• The frequency of password loss per user and the current Mean Time to Resolve
• How often do users experience this?

2. Design the Data Collection and Processing Layer

We moved away from manual triage by leveraging our internal engineering power.

• Data: Live ticket data from specific Slack channels, Jira Ticket Managemenet and ITSM Service Management (for ticket) was continuously extracted.
• Processing: Using our internal automation platform, Raven, we routed these incoming tickets to an LLM. The AI didn’t just read them; it automatically categorized the Slack issues, built a structured dataset, and performed sentiment analysis to prioritize urgent operational blockers.

3. Ensure Data Quality and Trust (Control)

At Trendyol scale, you cannot rely on a “set it and forget it” AI. The Tech Resolution team performed weekly audits to verify the AI bot’s categorizations against our newly established Slack Issue Dataset. Any discrepancies were meticulously reviewed and used to fine-tune the LLM prompts. This feedback loop ensured our AI categorization remained a reliable, single “source of truth.”

4. Operationalize Insights (OKR & Roadmap)

• OKR: We set an ambitious target: “Reduce manual password ticket volume by 60% in Q2.”
• Action: The AI-categorized data clearly identified specific UX flaws and architectural gaps in certain internal applications. Instead of just resetting passwords faster, we integrated these insights into the Tech Enterprise Tools roadmap to completely redesign the login and Self-Service Password Reset (SSPR) experience.
• The Result (Actual Metrics): By combining Raven’s AI automation with root-cause UX improvements, we didn’t just hit our goal — we exceeded it. We achieved a 65% actual reduction in manual password tickets within the first quarter. More importantly, the MTTR for access issues dropped from hours to mere seconds, successfully shifting our team’s focus from repetitive manual support to sustainable engineering.

Conclusion

Bridging the gap between strategy and operations is not merely a technical challenge; it is a cultural transformation. As we saw with the Trendyol Tech Resolution team, successfully resolving an operational bottleneck requires a systematic approach. By defining the right metrics, designing an automated AI data pipeline, controlling accuracy through human oversight, and operationalizing those insights into a concrete roadmap, the team didn’t just patch a problem — they permanently reduced manual password tickets by 65% and slashed resolution times to mere seconds.

Ultimately, data does not exist simply for reporting. It exists to eliminate repetitive work, build better products, and create a real, tangible impact on your organization.

Remember: Data does not exist simply for reporting; it exists to build better products and to create a real, tangible impact.

Ready to make an impact? Want to build intelligent data frameworks, solve hyper-scale engineering challenges, and shape the future of e-commerce? Join our team! 🚀 **Explore open roles at Trendyol Tech Careers**