From Data Sprawl to a Single Source of Truth

01Listen before you design

Listen|Step 1

Listen Before You Design

Start with people and their workarounds. Interviews reveal the habits, incentives and gaps behind fragmented information.

Start with people

Most data initiatives start with the data: an inventory of systems, a list of KPIs, a request for requirements. Start with people instead.

Fragmented data is rarely a technology problem at its root. It grows out of how people work, which tools they trust, what they’ve built to get around gaps, and what the organization rewards. None of that shows up on a system diagram. You find it by asking.

How to apply it

01 / 04
  1. Map

    Map who to talk to.

    List every function that creates, interprets or acts on operational data: operations leadership, production, maintenance and reliability, integrity, health and safety, projects, and technology. Include frontline people as well as managers, and always include IT.

  2. Ask

    Ask about work, not wish lists.

    Don’t ask what dashboard people want. Ask them to walk you through a recent decision: which tools they opened, which numbers they checked by hand, where they keep their own records, and why.

  3. Evidence

    Treat workarounds as evidence.

    The spreadsheet someone trusts more than the official system is the most useful thing you’ll find. Note every one.

  4. Patterns

    Look for the pattern behind the symptoms.

    Ask what in your culture, incentives or habits lets fragmentation grow. That is what will have to change, and no tool will change it for you.

In the field

With an oil and gas company, we held 26 conversations over two months across eight parts of the business before designing anything. The workarounds were everywhere, and so was the frustration.

Conversations
26
Months
02
Business units
08

It does get overwhelming... there’s so many different streams, sometimes it’s hard to maintain a focus on what we need to be focused on.

From the interviews

Fragmentation sources

The most important finding wasn’t technical. The company had a strong independent spirit: teams were trusted to solve their own problems their own way. It was a genuine strength. It also meant every team built its own reports, dashboards and naming conventions, and the result was many competing versions of the truth.

That’s kind of on us, we should have never got to the place we’re in now. We should have had standards out there that kind of dictated naming conventions and things.

From the interviews

The eight sources of fragmentation, sorted into three groups.

Information & Trust

  • Source data quality
  • Replication and duplication
  • Missing shared standards
  • No trusted central data store

Tools & Workflows

  • No system retirement policy
  • Dashboard accessibility

People & Support

  • Data literacy and training
  • IT coordination and capacity