Who this is for
This subskill is for Product Analysts who build dashboards and self-serve reports. If you need to let stakeholders slice metrics and zoom from summary to detail, mastering filters and drilldowns is essential.
Prerequisites
- Basic familiarity with a BI tool (e.g., Power BI, Tableau, Looker, or similar)
- Understand dimensions vs. measures, and basic aggregations (sum, count, avg)
- Have access to a simple dataset (orders, users, events) to practice
Why this matters
Real product decisions often depend on slicing a metric by segment and then drilling into outliers. Examples:
- Investigate a drop in conversion: filter by device, country, and experiment group.
- Understand feature adoption: drill from monthly active users to cohorts to user-level sessions.
- Root-cause churn spikes: filter by plan, region, and release version, then drill to problematic SKUs or features.
Concept explained simply
Filters limit what data is shown. Drilldowns change the level of detail (e.g., from Category to Product to SKU). Together, they help you answer both "what changed" and "where exactly did it change" without changing reports.
A simple mental model
Open the mental model
Think of a report as a series of sieves and lenses:
- Sieve (Filters): Decide which rows get through (e.g., last 30 days, Country = US).
- Lens (Drilldown): Decide how to view the surviving rows (e.g., roll-up by Category, then zoom to SKU).
- Context: Filters can apply at different levels. From widest to narrowest: Report-level > Page-level > Visual-level.
When charts look inconsistent, check which filter level is in effect. Conflicts often come from a visual-level filter overriding others.
Core ideas you will use daily
- Filter types: categorical (dimension), numeric (measure), relative date, top N, include/exclude, AND/OR logic.
- Filter scope: report/page/visual and cross-filtering between visuals.
- Hierarchies for drilldown: e.g., Category → Subcategory → Product → SKU.
- Drillthrough: jump from a summary to a dedicated detail page with the context carried over.
- Performance basics: fewer high-cardinality filters and thoughtful defaults improve load time and clarity.
Worked examples
Example 1: Diagnose a conversion drop
- Set a relative date filter to the last 28 days.
- Add a device type filter (Desktop, Mobile, Tablet).
- Add a country filter and select top 5 by sessions using a Top N filter based on Sessions measure.
- Observe whether the conversion drop is isolated to Mobile in 1–2 countries.
Why this works
Relative dates keep the dashboard fresh; Top N focuses attention on the most impactful segments, reducing noise.
Example 2: Feature adoption by cohort with drilldown
- Create a hierarchy: Cohort Month → Plan → Feature.
- Start at Cohort Month and review Feature Adoption Rate.
- Drilldown into a weak cohort to Plan, then to Feature to find underperformers.
- Apply a page filter to exclude cohorts with < 100 users to avoid tiny-sample noise.
Why this works
Hierarchies maintain context while diving deeper. A minimum-size filter avoids misleading rates.
Example 3: Revenue heatmap with drillthrough
- Visualize Revenue by Category × Country as a heatmap.
- Enable drillthrough to a detail page showing SKU-level revenue, returns, and margin for the selected Category and Country.
- Click a low-revenue cell and drillthrough to inspect SKU mix and return rates.
Why this works
Drillthrough gives a focused analysis space without cluttering the summary page.
How to build clean filters and drilldowns
- Pick filter defaults that match the primary question (e.g., last 28 days, key markets).
- Use report-level filters for must-have context (e.g., exclude test users), page-level for analysis context, visual-level only for special cases.
- Create intuitive hierarchies with consistent naming and ordering.
- Enable cross-filtering/cross-highlighting between visuals when it supports the story; turn it off when it creates confusion.
- Label filters clearly with units and definitions (e.g., "Active Users (7-day)" vs "Active Users").
Exercises you can do now
These exercises mirror the tasks below. The quick test at the end is available to everyone; log in to save your progress.
Exercise 1: Segment conversion with focused filters (ex1)
- Add a relative date filter for the last 28 days.
- Add filters for Device, Country, and Traffic Source.
- Use a Top N filter to show only the top 5 countries by Sessions.
- Confirm that all visuals on the page respect these filters consistently.
Expected output
A dashboard showing Conversion Rate segmented by device and traffic source, scoped to last 28 days and top 5 countries by sessions.
Exercise 2: Build a hierarchy and drillthrough (ex2)
- Create a hierarchy: Category → Subcategory → SKU.
- On the summary page, chart Revenue by Category with drilldown enabled.
- Enable drillthrough to a detail page showing SKU performance (Revenue, Units, Return Rate).
- Verify that drilling from a weak category carries Category context to the detail page.
Expected output
A summary visual with drilldown plus a SKU detail page reached via drillthrough, filtered by the selected Category.
Self-check checklist
- Are report-, page-, and visual-level filters not contradicting each other?
- Do filter names include units or definitions where needed?
- Is cross-filtering helping rather than confusing?
- Can a stakeholder move from high-level trend to item-level details in 2–3 clicks?
Common mistakes and how to self-check
- Too many filters. Symptom: stakeholders feel lost. Fix: keep only decision-driving filters; use defaults.
- Conflicting scopes. Symptom: visuals disagree. Fix: audit report/page/visual filters; remove hidden visual-level filters.
- Unstable Top N. Symptom: items jump in/out daily. Fix: pin time window and consider adding a minimum threshold.
- Drilldowns without hierarchy logic. Symptom: confusing order. Fix: define natural levels (e.g., Country → Region → City).
- High-cardinality slicers. Symptom: slow and hard to search. Fix: filter earlier by parent level or add search.
Mini challenge
Your PM asks: "Which mobile countries had the steepest drop in checkout completion last week, and which products were most affected?" Build a page with:
- Relative date = last 7 days vs previous 7 baseline
- Device = Mobile only (report-level)
- Top N countries by sessions (N=10)
- Drilldown Category → SKU and drillthrough to a SKU detail view
Tip
Use a difference or percent change measure to rank countries, then drillthrough to SKU detail for the worst country.
Learning path
- Master filter scopes and relative date filtering.
- Build clear hierarchies and enable drilldown.
- Add drillthrough pages for investigations.
- Tune performance: simplify slicers, reduce cardinality, set smart defaults.
- Practice on real stakeholder questions weekly.
Practical projects
- Growth Dashboard: Filters for device, region, acquisition; drilldown from weekly actives to feature usage.
- Commerce Performance: Category → Subcategory → SKU hierarchy with drillthrough to returns analysis.
- Experiment Review: Treatment vs control filter, date slider, and drillthrough to user-level event timelines.
Next steps
- Refactor an existing busy dashboard: remove two non-essential filters and add a single, meaningful drillthrough.
- Create a short guide for stakeholders explaining the filter defaults and drill path.
- Take the quick test below to check your understanding.