Rebuilding Healthify’s reporting suite
Research, product design, data visualization, process improvement, UX writingHealthify (acquired by WellSky) was a startup that addressed health inequities by connecting people to social services. Case managers can find resources for their clients, create referrals, and check on progress made.
Our reporting product needed to be completely rebuilt. I designed new dashboards, spreadsheets, guides, and processes.
The challenge
Healthify was at risk of customer churn because of a low-quality reporting product. Customers received a mix of templates, ad hoc reports, and one-off requests, delivered by multiple internal teams. Content was a problem too: confusing metrics, data discrepancies, and redundant information.What I did
I created a framework for product offerings with a product manager, made prototypes to hand off to the lead data scientist, organized review sessions, and wrote customer guides. I also designed new processes for the Product Support team and trained them.Impact
Healthify regained customers’ trust and retained those at risk of churn. We now can regularly deliver higher quality, standardized reports with reliable data. Plus, account managers can start using reports in their quarterly customer calls.Product Support can set up reports for new customers in under an hour, as opposed to multiple teams taking a few days to coordinate.
Who are reports for?
How do we tell a data-driven story to customers, who have unique goals and needs? It wasn’t feasible to deliver highly targeted performance reports monthly.Instead, standardized templates work for most use cases. The product manager, account managers, and I narrowed it down to 3 primary user groups. Each group would get their own template.
Executive sponsors use the Insights Report to evaluate ongoing partnerships with social service organizations and track their communities’ needs. This report shows long-term trends and community demographics.
End-user supervisors check their teams’ productivity with the Operational Report. This report summarizes monthly user activity KPIs.
In-house or external data analysts use extracts for audits. These spreadsheets are the most granular — they have timestamps and user names behind all search, screening, and referral activity.
End-user supervisors check their teams’ productivity with the Operational Report. This report summarizes monthly user activity KPIs.
In-house or external data analysts use extracts for audits. These spreadsheets are the most granular — they have timestamps and user names behind all search, screening, and referral activity.
Crafting modular layouts
Another challenge was that customers have unique product configurations. For example, not all customers have access to screenings or referrals.The solution was to create modular sections based on major features (search, screenings, referrals). We could mix and match sections to create templates relevant to most customers.
A trade-off we accepted was that Looker can’t split reports into pages, so reports were long PDFs that weren’t easily printable. The alternative was to send them as a series of dashboards, but this would be too time-intensive to maintain.
Insights Report sections
Operational Report sections
Switch from data visualization to processes
Visual reports were on hold for a few months, so that engineers could implement new data models and data extracts. Meanwhile, I trained the Product Support team on how to use Looker. I helped them develop workflows for doing QA, delivering, and troubleshooting reports. This ensured that customers would see fewer bugs and data inconsistencies.
Validating designs
Account managers and product support specialists reviewed sample reports and gave helpful feedback:- It was hard to understand what conclusions customers were supposed to draw.
- Customers frequently have questions about some of the specialized terminology and calculations.
To address these concerns, I improved the written content:
- Chart titles are open-ended questions that the chart can answer
- Annotations are spread out throughout the reports. They define terminology and explain calculations.
- I wrote data dictionaries that define every metric. When customers receive reports, they’d also get a link to the relevant one.
Shipping the new reports
The new templates and data dictionaries addressed most customer needs — no more major product requests or emergencies.Post-script
The templates later accommodated a rebrand, new features, and new data models.
