Data, Databases, and Dashboards — Lesson 5
Analytics, Dashboards, and Data-Driven Decisions
Learning Objectives
- 1Design analytics events that connect to business decisions.
- 2Build dashboards that answer specific questions rather than displaying noise.
- 3Establish backup, recovery, and data retention practices.
Analytics events that matter
Analytics events record meaningful actions: page viewed, form submitted, account created, checkout started, subscription canceled, support ticket opened. Good events are tied to decisions the business needs to make — they answer questions like "where do people drop off?" and "which channels produce paying customers?"
The most common analytics mistake is tracking everything without a plan. Hundreds of events create a data swamp where nobody can find the signal. Start with the five to ten events that matter most for your current business decisions. Add more as you identify new questions that need answers.
Every important event should have a clear name, a definition of exactly when it fires, and an owner who is responsible for ensuring it works correctly. Inconsistent event naming — "click_buy" versus "purchase_click" versus "buy_button_clicked" for the same action — makes analysis impossible.
Dashboards that drive action
A useful dashboard answers a specific question: Are leads increasing? Where are customers dropping off? Which campaign produces revenue? Which support issue is growing? If a dashboard does not connect to a decision, it is entertainment, not a business tool.
Design dashboards by starting with the decision, not the data. Ask: What question does this dashboard answer? Who looks at it? How often? What would they do differently based on what they see? This approach prevents the common mistake of building dashboards that display everything available rather than what matters.
Every metric on a dashboard needs a definition (exactly what it measures), a source (where the data comes from), an owner (who is responsible for accuracy), and a threshold (at what point someone should take action). Without these, metrics become decorations.
Backups, recovery, and retention
A backup is a copy of data stored separately for recovery purposes. Backups protect against accidental deletion, database corruption, ransomware, failed updates, and human error. They are your insurance policy, and like all insurance, they only matter when you need them.
The critical question about backups is not whether they exist but whether they can be restored. A backup you cannot restore from is worthless. Test your restore process before you need it. Know how long restoration takes, how much data might be lost between the last backup and the failure, and who has the authority and knowledge to perform a restore.
Data retention defines how long information is kept before deletion or archiving. Keeping everything forever increases storage costs, privacy liability, and the scope of any security breach. Define retention policies based on business need, legal requirements, and privacy obligations. Sometimes the most responsible data decision is to delete data you no longer need.
Case Study
The dashboard that changed the budget
Situation
An e-commerce company was allocating marketing budget based on traffic volume — the channel that drove the most visitors got the most money. When they built a dashboard that tracked revenue attribution instead of just traffic, they discovered that their highest-traffic channel produced almost no revenue, while a low-traffic email channel produced 40% of repeat purchases.
Analysis
The traffic dashboard was answering the wrong question. The right question was not "where does traffic come from?" but "where does revenue come from?" Shifting budget from the high-traffic channel to the high-revenue channel improved ROI by 60%.
Takeaway
The most important dashboards connect to revenue and decisions, not to the biggest numbers. Volume metrics without quality metrics can actively mislead.
Reflection Questions
- 1. What are the five most important questions your business needs data to answer? Do your current dashboards answer them?
- 2. When was the last time your organization tested restoring from a backup?
Key Takeaways
- ✓Analytics events should connect to business decisions, not just track everything.
- ✓Design dashboards around decisions, not around available data.
- ✓Every metric needs a definition, source, owner, and action threshold.
- ✓Backups must be tested — an unverified backup is not a backup.