GET /candidate/projects/data_quality_vision

{
  "title": "Data Quality: from cleanup to prevention",
  "summary": "Set a 3-5 year vision that moved HubSpot's Data Quality tools from cleaning up bad CRM data to catching it early, then shipped the first proactive features.",
  "details": {
    "company": "HubSpot, Data Quality (Operations Hub, now Data Hub)",
    "role": "Product Designer, then Senior Product Designer; sole designer on Data Quality",
    "team": "[PM], an associate PM (APM), [N frontend and backend engineers], Content Design",
    "stakeholders": "The Data Sync, Import, CRM and Workflows teams; Ops Hub sales and support",
    "methods": "Customer feedback analysis, alerts research, usability testing, vision workshops, UX audit of small fixes",
    "tools": "Figma, FigJam"
  },
  "overview": {
    "problem": "Customers found CRM data problems, such as duplicates and missing values, only after those problems had broken a report or a workflow. Fixing them was manual, and the tools only helped once the damage was done.",
    "solution": "Set a 3-5 year vision built on one principle: reactive, then proactive, then preventative. Shipped the first proactive features: alerts, a Missing Data tool, better duplicate management and in-line fixes.",
    "impact": [
      "The vision's principles still guide the Data Quality roadmap",
      "The Missing Data tool tested well with users [Add: test results]",
      "Coached an APM, gave engineers more ownership, and hired a replacement",
      "[Add: alert adoption or CSAT trend]"
    ]
  },
  "context": {
    "business": "Clean CRM data is the foundation for reporting, automation and, increasingly, AI across HubSpot. Data Quality was a growing part of Operations Hub, and data management was becoming a platform-wide priority.",
    "customer_problem": "Admins and ops teams found data problems only after the damage was done: a broken report, a duplicate contact emailed twice, a workflow that misfired. Fixing them was manual and repetitive.",
    "evidence": "Tracked feedback themes and internal questions in the #ops-hub-sales-help and #data-quality Slack channels; started a monthly readout of CSAT and customer feedback. [Add: top themes and CSAT figures]",
    "problem_statement": "How might we help customers catch and prevent data problems before they cause damage, rather than cleaning up afterward?"
  },
  "role": {
    "owned": "The UX vision and strategy for Data Quality, research and discovery (alerts research, anomaly detection exploration), end-to-end design of shipped features, and a running list of small UX fixes fit into each quarter.",
    "partnered_on": "A feedback capture and sharing process with the PM; coaching a new APM; Data Staging with Data Sync and Import; Data Quality insights in the CRM; a Data Quality activation project with CRM and Workflows.",
    "constraints": [
      "Bandwidth: also leading Data Enrichment in 2023 and 2024, so had to delegate",
      "Team maturity: few established processes and a brand-new APM",
      "Platform: Data Quality touches records owned by many teams, so changes needed alignment across them"
    ]
  },
  "process": {
    "listen_early_2023": "Joined from Data Sync; immersed in customer feedback and internal questions; started a monthly CSAT and feedback readout; streamlined one-pagers so the team shared one picture of customer problems.",
    "explore_2023": "Ran alerts research and an anomaly detection exploration; led discovery for duplicate alerts, a Data Quality email digest and automatic de-duplication.",
    "set_vision_2023": "Wrote \"Data Quality in 2026\", a 3-5 year vision built on three principles: reactive, then proactive, then preventative.",
    "ship_and_iterate_2023_2024": "Shipped the first proactive features, starting with alerts, then customer-requested fixes. Gave frontend engineers more ownership during high-fidelity mockups."
  },
  "decisions": [
    { "decision": "Where to start", "options": "Jump to prevention (AI anomaly detection) / one step at a time", "chose": "Proactive first (alerts), because customers needed to see problems early before trusting automatic prevention", "tradeoff": "The bigger AI vision had to wait" },
    { "decision": "Small fixes vs. new features", "options": "Only big features / room for small fixes", "chose": "A running list of small UX fixes fit into Q2 and Q3 alongside larger work", "tradeoff": "Slower progress on headline features" },
    { "decision": "Design-to-engineering handoff", "options": "Designer owns every high-fidelity screen / engineers co-own", "chose": "Frontend engineers co-own high-fidelity mockups; the team embraced it", "tradeoff": "Less direct control over pixel-level details" }
  ],
  "buy_in": "Used the vision deck and principles to show how each quarter's work led toward prevention; monthly feedback readouts kept customer evidence in front of everyone.",
  "solution": [
    "Proactive alerts: tell customers about data problems, such as new duplicates, as they happen (researched 2023, shipped 2024)",
    "Faster fixes where the problem is: in-line editing, exporting issues, better duplicate management, new reports, and a Missing Data tool for finding and filling empty fields",
    "A Data Quality component library and rewritten copy, so new features stayed consistent and clear"
  ],
  "impact": {
    "metrics": [
      "Strategy: 3-5 year vision and principles still guide the roadmap",
      "Features shipped: alerts, in-line edits, exporting issues, better duplicate management, new reports",
      "Usability: Missing Data tool tested well [Add: task success]",
      "Adoption / CSAT: [Add: alert usage, duplicates merged, or CSAT trend]",
      "Team: coached the APM; hired a replacement; engineers took more ownership of mockups"
    ],
    "quote": "Manager wrote that the work moved the team \"from a reactive to a proactive stance\" and that \"systems thinking and platform approach have elevated the work and the team.\""
  },
  "reflection": {
    "do_differently": "Tailor the message to each audience earlier; shorter, more visual stories would have built support faster with engineers, PMs and leadership.",
    "learned": "A principle as simple as reactive, proactive, preventative outlasts any single design and lets a team keep making decisions after you move on.",
    "next": "Hired a replacement who carried the vision forward; the principle fed into Data Studio's approach to cleaning data."
  }
}