The Telephone Game
The Problem
Adding layers of people between idea and execution. Business person → Business Analyst → Project Manager → Data Developer → UI Developer → back to business. Six people. Six handoffs. Each handoff loses fidelity. Each translation adds cost. Each person added increases project time exponentially. A simple "show me last quarter's sales" becomes a 6-week project through the assembly line. Or worse: you're still stuck in 2007 where Business → BA → Sourcing → ETL → Modeler → BI Developer means a 2-month cycle for a dashboard. The manufacturing analogy is perfect: imagine Toyota running a production line where each station speaks a different language and has to translate work orders. You'd never ship a car. That's your BI process.
The Principle
Flatten the org. Get the person with the idea as close to the tools as possible. Give business analysts the same BI toolkit as developers. Train them. Trust them. Stop the assembly line of handoffs. The magic happens when someone who understands the business can build their own answer in 2 hours instead of waiting 6 weeks for IT. This is the self-serve wave: everyone on the same platform, fewer handoffs, faster insights. Think manufacturing: cross-functional cells where one team owns the whole widget, not 6 teams passing pieces. Reduce handoffs from 6 to 1 (or zero). Time-to-answer drops 90%. Fidelity stays perfect because no translation happens.
3 Steps to Take Action
- 1
**Map Your Handoffs** For your most common request types, draw the flow: - Who touches it? - How many handoffs? - Time at each step? Example: "New dashboard request" might be: Business → BA (3 days) → IT intake (5 days) → Data eng (10 days) → BI dev (14 days) → QA (7 days) → Deploy (3 days) = 42 days, 6 handoffs If >3 handoffs, you have a telephone game problem.
- 2
**Enable Self-Service** Give business analysts/power users direct access to tools: - Pre-built data models (so they don't need data eng) - BI platform training (Tableau, Power BI, etc.) - SQL basics (so they can query) - Templates and best practices Start with 5 power users. Train them. Let them build. Goal: Business users build 50%+ of routine reports themselves.
- 3
**Shift IT to Enablers** IT's new role: - Build/maintain data models and pipelines - Govern data quality - Create templates and training - Review complex work - Build what requires engineering NOT: Build every single report. Measure: - % of reports built by business users (target: 60%+) - Average handoffs per request (target: <2) - Cycle time (target: 80% reduction) Bonus: IT stops being a bottleneck and starts being strategic.
In Practice
A major bank had the classic 2007 problem: Business request → BA documents → Sourcing pulls data → ETL team loads it → Data modeler builds semantic layer → BI dev builds report → back to business. Average cycle time: 47 days. They shifted to self-serve: gave BAs access to Tableau + pre-built data models. Trained them on SQL basics. First month: BAs panicked. "We're not developers!" By month 3: BAs building 70% of reports themselves. IT focused on data quality + models. Handoffs dropped from 5 to 1 (just "approve the data model"). Cycle time: 47 days → 7 days (85% reduction). Fidelity improved because BAs who knew the business built exactly what was needed. IT stopped being a bottleneck and became enablers. The telephone game ended when you gave everyone the same phone.
