Dashboard Sprawl

Dashboard sprawl is what happens when a revenue org accumulates overlapping dashboards faster than it retires them. The symptom is easy to spot. Two people open two dashboards, both labeled pipeline, and get different numbers, and the meeting turns into a debate about which one is correct.

The count itself is not the problem. Forty dashboards with one shared metric layer is fine. Six dashboards with six private definitions of qualified pipeline is not.

How sprawl starts

Nobody sets out to build twelve pipeline views. Sprawl comes from cloning. A regional leader wants the standard view filtered to their territory, copies it, and now there are two. The copy is edited over the next quarter as that leader's questions change. The original is edited by someone else. Six months later the two disagree and neither owner knows why.

The second source is the one-off that never got deleted. A board question in Q1 produced a custom view, the board meeting ended, and the view stayed in the shared folder looking exactly as official as everything around it.

The cost

Sprawl converts review time into reconciliation time. That is the direct cost. The indirect cost is worse, because a team that cannot agree on the current number cannot argue productively about the forecast.

It also hides staleness. ORM's stated mechanism for forecast failure is that the model runs on assumptions the business or market already moved past. Dashboards fail the same way. A view built around last year's stage definitions keeps rendering, keeps looking current, and quietly reports against a funnel that no longer exists. Nobody catches it because nobody owns it.

Sprawl also masks the composition problems that matter. A dashboard showing 4x coverage looks reassuring on every one of the twelve copies, and none of them show whether that coverage sits in the wrong segment or rests on aged opportunities.

Cleaning it up

Start with usage data rather than opinion. Every BI tool records last-viewed dates. Archive anything untouched in ninety days, and expect that to clear a large share of the inventory with no objections.

For the survivors, apply one rule: every dashboard maps to a recurring meeting or decision. A forecast call needs a forecast view, and a pipeline review needs a pipeline health view. A dashboard that maps to nothing is a saved query someone published by accident.

Then move metric definitions out of dashboard filter logic and into a shared semantic layer, so forecast accuracy means one thing across every view instead of being reimplemented per dashboard. That single change prevents the drift that made cleanup necessary. It also makes the underlying sales forecasting work auditable, because the number on the screen traces back to a definition rather than to whichever filter its author picked.

Consolidation holds only if someone owns the shared folder. Without an owner, sprawl restarts the week after cleanup ends.