One overview to assign, track and remind, so that 1000+ data points reach the right person

Role: Sole Product Designer, with 1 PM, 3 engineers & CTO.

Context

Mavue is an early-stage B2B SaaS company that helps small and mid sized manufacturers in Germany create their annual sustainability report. Throughout the year, they collect data on Environment, Social and Compliance topics, and different departments own different data. HR, for example, owns the HR numbers.

Problem

A company has to collect 1000+ data points, that are spread across many people who are not sustainability experts.

Especially challenging is the data entry:

• 350+ are numerical that have to be calculated (CO2 emissions)

• 650+ are written explanations (policies, action plans, risks)

Before, the data was collected in spreadsheets, and every department had its own. That meant no consistency, no way to know whether the data was up to date, and no way to be sure it was correct.

It was also often unclear who owns which data point, even inside one department. In HR, for example, one person may own payroll data like the gender pay gap, another training hours and a third sick days. So someone had to find the right owner for each data point, see who was late and remind them.

Constraints

• Early stage, few engineers: The product was built mainly to prove product market fit.

• Build fast: Existing components from a component library, plus a few new elements to keep it simple.

• Dense data, non expert users: 1000+ rows, each with an owner, a due date and a status, used by people who only enter their own numbers.

Solution & Key Decision

The overview lists every data point with its category, owner, progress, due date and status, with filters for overdue, not started, in progress and completed.

• Decision 1: a calm table, with actions on demand. Buttons on every row of a table this dense would be noise, so select, remind and open appear only on hover, as well as filters.

• Decision 2: bulk first. Selecting rows raises a floating bar to assign a data owner, change the due date or send a reminder for many data points at once. When assigning, people can choose to notify by email and leave a short note.

• Decision 3: a sanity check while typing. On the detail page, each value sits next to the previous year, so people can spot mistakes as they enter data. The empty state keeps the filters in view and offers a one click way back.

Key Screens

• Overview and hover actions: a dense list that stays calm.

• Rows selected and assign data owner: delegating many data points in a few clicks.

• Send reminder: following up without leaving the table.

• Data point detail: entering values with the previous year as a check.

• Empty state: finding the way back from no results.

Process

I started to gather the technical requirements and customers feedback with the internal team and benchmarking other data table solutions. I was looking at filters, sharing function and other key features. After that, I drafted 2-3 ideas and tested them as interactive prototypes with users and engineering. Based on the feedback, I finalized the Hi-fi design.

Outcome

• Faster: Many data points can be assigned in one step, instead of one by one.

• Less manual work: Reminders and assignments happen in one place, so nobody has to chase people by email.

• Clear ownership: Every data point has an owner, so everyone knows who is responsible.

• Better data quality: Values are shown next to the previous year while typing, so mistakes are easier to catch.

• Scalable: The same overview works for a few hundred or for 1000+ data points.

• One source of truth: One overview replaces a separate spreadsheet for every department.

What I would do differently today

AI tools were still very early when I worked on this. Today I would:

• Speed up the benchmark: Use Claude Projects to collect and compare how other data table tools handle filters, sharing and bulk actions, and to synthesize user feedback into themes, while I decide what matters.

• Explore more drafts: Use Claude Design and Figma Make to produce more than 2 to 3 variations quickly while staying consistent with the existing design, and test more of them with users.

• Test at real scale: Use Claude Code to build a prototype with 1000+ rows, so filters, hover actions and bulk selection are tested with realistic volume.

• Suggest owners: Explore AI that proposes a likely data owner for each data point, based on category and earlier assignments, with the person confirming each one.

Note: The screens shown were recreated, anonymized and adapted: data, copy and branding are invented to protect confidential product details.