'The Numbers Don't Add Up'
Metrics didn’t match manual counts, destroying confidence in all analytics.
Business Impact:
Users defaulted to manual tracking, ignoring the value of our dashboard
Original
Redesigned
Context: Why This Mattered
Serving as our platform’s home page, our dashboard was supposed to prove “Hey, look what our platform can do for you!” In reality, engagement was effectively zero: a single recorded interaction across an entire quarter. This created some serious issues:
We promised 'powerful analytics', delivered 'meh.'
New users land on dashboard, see nothing useful, bounce
Can't prove value when nobody engages with the data
What We Discovered: The Root Causes
3000+
Sessions Analysed
15+
User Interviews & Feedback Reviewed
User research and data analytics identified four critical root causes for the near-zero engagement
'Answering the Wrong Questions'
Dashboard didn’t answer questions recruiters were asking (e.g., “apps per job” vs. the needed “apps per source”).
Business Impact:
Users fail to see the values they are paying for, risking churn.
'Interesting, but Now What?'
Data was “interesting” but not actionable. Users saw a number and didn’t know the next step.
Business Impact:
Every metric was a dead end. Users couldn’t connect insights to workflows, so the dashboard never entered their daily routine.
'Can't Dig Deeper'
No filters meant power users couldn’t answer specific business questions.
Business Impact:
Power users churned to competitors with better analytics.
Strategic Trade-offs
Through cross-functional discussions with Product, Engineering, and Customer Success, we aligned on success criteria and assessed the risks of a major re-launch.
Rebuild vs. Patch
With effectively zero engagement, we had nothing to lose. Convinced leadership to invest 3 months instead of shipping quick fixes.
Trust vs. Bells and Whistles
Leadership wanted AI features. We successfully pivoted the roadmap to data transparency first. Can’t upsell features if users don’t trust your data.
Filters vs. Custom Dashboards
Deprioritized “drag-and-drop” widgets in favor of session-persistent filters to solve the core pain point faster with less dev overhead.
Design Direction
The redesign focused on three strategic pillars:
Transparency
All metrics must be explainable to build trust and confidence
Customization
Enable recruiters to filter and segment data according to their specific needs
Actionability
Analytics must connect to workflows to create a complete user flow
Design Solution
Information Architecture
Ranked modules by actual session frequency and a manual audit of custom report requests.
“Must-have” metrics moved to the top, secondary data was grouped below to reduce cognitive load.
Building Trust
Introduced metric definitions via tooltips.
Design rationale
When users said “the numbers don’t match what I counted,” the issue was often that we count these numbers differently. Making these definitions explicit and always accessible restored confidence in the data.
The Filter System
Enabled users to segment data by time, company, department, and more. Automatically saving active preferences in the browser session to reapply them upon their next visit.
Design rationale
Filters directly addressed the “Can’t dig deeper” pain point and became our highest-impact feature. The filter design prioritized speed and clarity: persistent placement, instant application, and consistent with the rest of the platform.
Micro-interaction
Added skeleton states to provide immediate feedback during backend queries, improving perceived performance.
Strategic Visual Refresh
Included visualizations to make insights scannable & signify the big re-release
Design rationale
Besides the functional purpose of making data patterns more scannable, the stark visual change was a psychological reset for users who had spent months ignoring the old dashboard.
Contextual Actions
Connected KPIs to relevant workflows for deeper analysis
Design rationale
This solved the “Interesting, but now what?” problem. Contextual links transformed passive data consumption into active workflow engagement, making the dashboard a launchpad rather than a dead end.
Scalable Framework
Built reusable components aligned with the design system for scalability
Design rationale
Every chart type, tooltip pattern, and filter component was built as a reusable design system element, so as to ease future dashboard expansion while maintaining consistency.
Impact & Validation
How We Measured
Every metric here runs on the same constant cohort, start to finish:
- Only accounts active the full period count; beta testers and churned accounts are excluded from both windows.
- The old dashboard was interactive too (links, clickable elements) — engagement was near-zero despite that, not because there was nothing to click.
- 12.2025 (mid-month launch) was excluded from the analysis window to avoid skewing the average with a partial month.
- 05–06.2026 (DACH seasonal hiring slowdown) show in the chart but are excluded from the analysis. Folding them into the analysis window would have doubled the reported average (7.1% → ~15%).
- No other dashboard changes, campaigns, or onboarding updates shipped between launch and April.
~0%
Baseline Engagement
1 interaction across 2000+ views despite existing links, 09.2025 – 11.2025
7.1%
Post-Launch Click Rate
Sustained avg., 01.2026 – 04.2026
415+
Interactions in 4 Months
vs. 1 in the entire prior quarter
Click Rate Total Interactions
Click rate and total interactions, 09.2025 – 06.2026, constant cohort.
0
Analytics Support Requests Since Launch
vs. ~20 data-accuracy complaints and custom-report requests logged H2 2024 – 12.2025.
7.5%
Filter Adoption
7.5% of sessions use filters, validating workflow optimization.
7×
Scroll Rate Increase
Users are exploring the full depth of data, not just the “fold.”
Interactions Drove the Engagement Increase
The sustained click rate was driven by the three core behaviors we designed for: high-level filtering, chart drill-downs, and shortcuts to hiring artifacts. Users went from one recorded interaction per quarter to roughly a hundred per month, finally seeing data and immediately acting on it.
Visual refresh restored user interests
The new visual / visualizations successfully sparked users’ curiosity. By making the dashboard visually distinct from the “broken” version, we encouraged users to explore the entire page.
Organisational Impacts
Standardised Dashboard Framework
My dashboard components are now the org-wide standard, used across all product squads.
Data Culture Shift
Introduced quarterly data review and influenced leadership to promote metrics in future product briefs.
Key Takeaways
- Data without transparency is worthless. Users needed to understand how metrics were calculated before they would believe what they showed.
- Users relied on filtering and segmenting data to make sense of it. A “snapshot” obstructed this sense-making loop.
- Connecting analytics to workflows transformed the dashboard from “interesting” to “essential” in users’ daily routines.
- Resisting AI predictions and advanced features in favor of getting basics right first was the right call. Trust couldn’t be built on a shaky foundation.
- Impact numbers deserve the same rigor as the product. The raw post-launch averages looked far more dramatic, until cohort and denominator checks surfaced churn and seasonal effects inflating them. The figures reported here are the ones that survived that interrogation.
