Field story 5 min

# The Report Was Obsolete Before It Was Finished

One project's work was spread across seven separate databases. Producing its status took about four days by hand and was obsolete on arrival. The problem was not the report or the people. It was the data architecture underneath both.

The condition

A capable portfolio manager spends days stitching status together by hand, and the result is out of date before anyone reads it.

## Six projects and no answer

On a Thursday afternoon, a portfolio manager in a global IT organization asked a senior leader for fifteen minutes. She brought her laptop. She said she could not explain the problem in words; it had to be seen.

She was responsible for six projects. All six were tracked in the organization's delivery tool. The teams were using it. The work was in the system. And she could not say how any of the six projects were performing. She had been trying for three weeks.

The leader knew who was sitting beside her. This was someone she had hired specifically to bring discipline to the portfolio, one of the most capable people in the organization. If this person could not assemble a project view, the problem was not the person.

## One project, seven databases

She opened the first project. From a portfolio point of view it was one project. Inside the tool, its work lived across two separate organizational spaces: five projects in the first and two more in the second. Seven separate databases for one piece of work.

Each looked different: different types of work items, different fields, different names for the same things. One tracked scope in large planning chunks; another tracked it at a much finer level. Two used a hierarchy unlike anything else, and their builders were unavailable to explain it.

She had tried exporting everything into a spreadsheet. The exports came out in seven formats. Field names did not match. Statuses did not match: one team used Active, another In Progress, another Committed, and nobody could say whether those meant the same thing. The parent and child relationships were configured differently, so when she tried to roll the work up, the hierarchies would not line up. She had been building lookup tables by hand to translate one team's statuses into another's, and every time she finished a reconciliation, the underlying data had already moved.

The leader asked how long a single status would take, end to end. About four days of manual work for one project, she said, and it would be obsolete by the time she finished. Six projects in parallel was not possible.

## Walls built into the tool

The leader called us that afternoon, and the next morning we were on a call with both of them. We asked how many organizational spaces and projects existed, when the tool had been adopted, and what guidance came with it. The answers were imprecise. Nobody had a current count. Nobody remembered there being formal guidance. We asked for a week to map the entire landscape.

A week later we put the map on the screen. The portfolio manager spoke first: she had thought it was bad, and had not known it was that bad. The organization was running hundreds of sealed organizational spaces holding more than a thousand separate project databases. Each was walled off from the others, with no built-in way to see across.

This was not a setting anyone could change. The boundaries were architectural. The tool had been adopted without an enterprise data architecture, and its default unit of separation was the project, so every team that needed a place for its work created a new one. Every initiative that crossed two teams either lived in a third project or was duplicated in both. Over the years the projects multiplied and the walls multiplied with them. The tool did include a structural feature that lets a single project hold the work of many teams while keeping each team's work distinct. It had never been introduced, so teams used the only container they could see.

## One project, many teams

Two weeks later we showed the portfolio manager, the senior leader, and three of her peers a prototype: one project, with every team's work inside it, organized as a tree. Product lines at the top, sub-products beneath them, teams beneath those, individual workstreams at the leaves. Each team could see only its own work when it wanted to. A portfolio manager could filter to any combination of teams and see everything contributing to a project. One view, one database, one set of field definitions, one set of statuses.

She clicked through it. The work she had spent three weeks failing to reconcile was already assembled. The hierarchies aligned because they were the same hierarchy. The statuses meant the same thing because they came from the same configured set. The roll-up happened on its own. She looked up and said she could do her job with this.

The path was designed so it would not freeze the organization or start a political fight. The target structure would be built and validated before anything moved. Active work would migrate on a rolling basis as each team adopted the new structure, not in one big move. Legacy projects would stay in place for historical reference instead of being deleted or forced.

## Fixing the data instead of the report

A better dashboard built on the same seven databases would have been just as wrong, only faster. The fix had to happen underneath, at the boundary between teams, where the data was split before any report could see it.

In the pilot that followed, the delivery tool was rebuilt so every kind of work follows the same structure, from yearly initiatives down to daily tasks. Five reports that had been assembled by hand were replaced by three dashboards that update themselves, and leaders and teams began working from the same numbers with the same definitions. The habits around how work gets entered were still settling when the pilot closed.

At the end of the prototype session, the portfolio manager closed her laptop. When she first walked into that office, she said, she thought she was admitting she could not do her job. What she had actually been reporting was that the setup everyone had been asked to adopt made her job structurally impossible. The teams had done everything asked of them. The problem was never the work, and never the teams. It was the architecture underneath them, and nobody had known they were building on it.

When the data cannot cross the boundary between teams, no report can carry the truth across it for them.

Know someone running into this? Send it to them.

[Share on LinkedIn](https://www.linkedin.com/sharing/share-offsite/?url=http%3A%2F%2Fagileon.test%3A64380%2Finsights%2Fthe-report-was-obsolete-before-it-was-finished%3Futm_source%3Dlinkedin%26utm_medium%3Dshare%26utm_campaign%3Dinsight-share%26utm_content%3Dthe-report-was-obsolete-before-it-was-finished)

Email

### Share With Your Team

Choose who it is for. You send it yourself from your own email; Agileon never writes to your colleagues.

Your Message

Open in Email

Share

[Share on LinkedIn](https://www.linkedin.com/sharing/share-offsite/?url=http%3A%2F%2Fagileon.test%3A64380%2Finsights%2Fthe-report-was-obsolete-before-it-was-finished%3Futm_source%3Dlinkedin%26utm_medium%3Dshare%26utm_campaign%3Dinsight-share%26utm_content%3Dthe-report-was-obsolete-before-it-was-finished)

Email

Components

[Bridge](https://www.agileon.com/insights?component=Bridge)[Signal](https://www.agileon.com/insights?component=Signal)

Industry
Global IT

Seen in practice
[The work that fixes this Operating Model Design](https://www.agileon.com/professional-services/operating-model-design)

Keep reading
[Flow Global IT The Engine Nobody Builds](https://www.agileon.com/insights/the-engine-nobody-builds)[Flow Global IT We Stopped Making Documents and Started Building Machines](https://www.agileon.com/insights/documents-versus-machines)[Bridge Inside Agileon's AIC Experience](https://www.agileon.com/insights/inside-agileon-aic-experience)[Surface The Seven Components of Work](https://www.agileon.com/insights/the-seven-components-of-work)
