Skip to content

Reporting · 7 min read

Reading the dashboard

Every headline figure and panel explained, including how realisation rate and return on cost are worked out.


What it is for

The dashboard is the management-review pack. It is computed live from the register each time you open it, so there is nothing to refresh and no report to run.

Dashboard
The ImproveDesk dashboard showing six headline tiles — open, overdue, realised benefit, return on cost, median cycle time and reviews due — above breakdown panels by status, Deming stage, category, source, priority and ageing of open work.

The six headline figures

TileWhat it counts
Open Improvements at Proposed, Triaged, Approved, In progress, Implemented or On hold. The smaller figure beneath is the total in the register, so "11 of 14" means three are finished or rejected. Note that Verified does not count as open — the work is done and evidenced, it is simply not closed yet.
Overdue Open improvements past their target date. The line beneath reports how many open records have no owner.
Realised benefit Total realised value across verified and closed records only. Beneath it: the forecast across the whole register, and the realisation rate.
Return on cost Realised benefit divided by total cost estimate, with the total invested beneath.
Median cycle time Days from raised to closed, across records that have closed. A median rather than a mean, so one abandoned three-year item does not distort it. Shows a dash until something closes.
Reviews due Improvements whose benefit review date has passed.

Return on cost reads low early on, and that is honest. The numerator counts only work that has been verified or closed; the denominator counts the cost of everything you have committed to. A programme half way through a large improvement will show a figure below 1×. It rises as work completes.

Realisation rate above 100%

Not an error. It means completed improvements beat their forecasts. The comparison is realised benefit against the forecast of those same verified and closed records, so it is a fair like-for-like — adding an ambitious forecast to new work cannot move it.

The six breakdowns

Each panel counts the register a different way.

  • By status — where everything sits in the lifecycle.
  • By Deming stage — the same data rolled into Plan, Do, Check and Act. A register heavy in Plan is one that proposes more than it delivers.
  • By category — your own configured categories.
  • By source — where improvements come from. The panel that answers "how much of this is audit-driven?".
  • By priority — Critical through Low.
  • Ageing of open work — open improvements bucketed 0–30, 31–60, 61–90, 91–180 and over 180 days since being raised. The right-hand buckets are the interesting ones.

Raised against closed

A twelve-month table of how many improvements were raised and how many closed each month, with a balance bar.

Raised against closed
A table titled 'Raised against closed, last 12 months' with a row per month showing raised and closed counts and a balance bar.

This is the flow question: is the register keeping up with itself? Consistently raising more than you close means the backlog is growing, however healthy the individual records look.

Needs attention

The improvements the daily sweep has flagged, with the reason for each — overdue, stale, review due or unowned.

Needs attention
A 'Needs attention' table listing improvements with their reference, title, a reason such as Review due, Overdue, Stale or Unowned, and a number of days.

Each record appears once, under the most pressing reason, in this order:

  1. Overdue — past its target date.
  2. Review due — its benefit review date has passed.
  3. Stale — not updated within the project's stale threshold.
  4. Unowned — open with no owner.

The Days column is the record's age, not how late it is. For an overdue item it shows days since the improvement was raised, not days past target — so a record raised last week and already overdue shows a small number. The rows are sorted by that age too. Treat the column as "how long has this been with us", and read the actual target date on the record itself.

Trend

A history of open count, overdue count, realised benefit and median cycle time, captured weekly.

Trend
A trend table with a row per weekly capture showing the date, open count, overdue count, realised benefit and median cycle time.

This is the only part of the dashboard that is not computed live — it is written by a job that runs once a week. A brand-new register shows a message rather than a table until that job has run at least twice, which is expected rather than a fault.

Using it in a management review

A workable running order, and roughly what each part answers:

  1. Open and overdue — is the register under control?
  2. Needs attention — what specifically needs a decision today?
  3. Raised against closed — are we keeping pace?
  4. Realised benefit and return on cost — is this worth doing at all?
  5. By source — where is the demand coming from, and is that the right mix?
  6. Trend — are the answers above better or worse than last quarter?

For a record of the discussion, export the register at the same time — see Exporting to CSV.