A transformation program's executive dashboard should show the few measures a sponsor can act on: benefits realization, schedule confidence, budget burn, adoption, risk exposure, decision cycle time, and open blocker age. Each measure needs a plain formula, a baseline, a threshold the sponsor has agreed to, and a named owner. The sections below give each formula, a worked example written with [placeholders], and the decision that follows when the number crosses its threshold.
Which KPIs belong on an executive transformation dashboard?
A measure earns a place on the executive dashboard only when a sponsor decision depends on that measure and the team can agree its formula in a team meeting. Measures that are easy to collect but change no decision stay in the team's working notes until they pass the checks below.
Check each candidate against four questions:
- Which sponsor decision changes if this number moves?
- Can the team state the formula and data source in a meeting?
- Is there a record from before the change started that can serve as a baseline?
- Who will own the number and bring it to the sponsor?
A measure that fails a question stays in the team's working notes until it passes. Give each measure a short line with its question, formula, data source, and owner.
How do you calculate benefits realization?
Benefits realization is the measured benefit to date, minus its baseline, divided by the benefit the plan expected for the same date, minus that baseline, expressed as a percentage. The lead calculates the figure only after the sponsor agrees the measure and the baseline is on record.
Formula: (benefit measured to date minus baseline) divided by (benefit planned for the same date minus baseline), expressed as a percentage. This is an editorial definition, not a standard taken from a listed source, so use it only after the sponsor agrees it.
Worked example, as steps the lead fills in with placeholders:
- Write the plan's claim: the plan expects [benefit measure] by [date].
- Record the measured value: on [date], the team measured [actual value].
- Record the baseline, taken from records kept before the change started: [baseline value].
- Record the planned value for that date: [planned value].
- Calculate: ([actual value] minus [baseline value]) divided by ([planned value] minus [baseline value]), expressed as a percentage, which gives [realization percent].
Check the denominator before dividing. If the planned change, [planned value] minus [baseline value], is zero or below, do not calculate a percentage. Flag the case to the sponsor instead, because a plan that expects no change or a fall cannot show progress as a percentage. If no baseline is on record, report the measured value with a note that the baseline is missing, and ask the sponsor to agree a baseline date before the first review.
What the sponsor sees: when the realization figure falls below [threshold percent] for [period], the sponsor receives the benefit, its baseline, and a decision: a recovery plan from the team, or a changed target.
Set the threshold with the sponsor in the meeting, choosing it from the benefit schedule the plan already holds. If the benefit is still the right measure, the sponsor asks the team for a recovery plan. If it is not, the sponsor changes the target before asking the team to work harder.
Ask the lead: What have we assumed about [benefit] that has not happened yet, and when will we know?
How do you calculate schedule confidence?
Schedule confidence is the share of milestones due in a period that are still forecast for their agreed dates, expressed as a percentage. A milestone with no agreed date is left out of the count, and a milestone whose forecast moves past its agreed date counts as missed.
Formula: milestones due in the period that are still forecast for their agreed date, divided by milestones due in the period, expressed as a percentage.
What the sponsor sees: when confidence falls below [threshold percent], the sponsor sees which milestone moved, what it is waiting on, and the decision that would bring it back to its agreed date.
Worked example, with placeholders: by [date], [due number] milestones are due. As of [date], [on-track number] of them are still forecast for their agreed dates. Schedule confidence is [on-track number] divided by [due number], expressed as a percentage, which gives [confidence percent].
For work items, The Burndown describes a cycle time histogram that can forecast likely delivery times, and its example chart reads that 75% of items are completed in 5 days or less. That is the source's illustration, not a program result. Build the same histogram from the team's own past cycle times before giving the sponsor a forecast date.
Agree a level with the sponsor, such as [threshold percent] confidence for the period. A milestone whose forecast moves past its agreed date goes to the sponsor as a decision, whatever the period's figure.
How do you track budget burn?
Budget burn is actual spend to date plus committed spend not yet invoiced, divided by planned spend to date for the same period, expressed as a percentage. Leaving out committed amounts makes the figure look healthier than the program is.
Formula: (actual spend to date plus committed spend not yet invoiced) divided by planned spend to date for the same period, expressed as a percentage.
What the sponsor sees: a burn above [threshold percent] of planned spend for [period], or a committed amount above [amount], goes to the sponsor as a decision.
Start with committed spend. Money the program has promised but not yet invoiced belongs on its own line in the burn figure, so the sponsor can see it apart from actual spend.
Worked example: the plan put [planned spend] against [work package] by [date]. Actual spend to that date is [actual spend]. Committed but not yet invoiced: [committed amount]. Burn is ([actual spend] plus [committed amount]) divided by [planned spend], expressed as a percentage, which gives [burn percent].
Compare spend with finished work in the same table row. List each work package, whether it is complete to date, and its spend, using this template:
| Work package | Complete to date | Planned spend | Actual spend | Committed, not invoiced | Burn percent |
|---|---|---|---|---|---|
| [work package] | [yes or no] | [planned spend] | [actual spend] | [committed amount] | [burn percent] |
A burn ahead of plan becomes a sponsor decision when the cause is scope or a dependency the team cannot clear alone. Otherwise the lead answers it with the spend plan.
Before the next sponsor meeting, the lead checks each row where spend is ahead of finished work and writes the cause beside it.
How do you measure adoption?
Adoption is the share of intended users who used the new process, tool, or practice in a period, expressed as a percentage. Fix the list of intended users first so the denominator does not change between reviews, and count only uses that happen in the system the program rules name.
Formula: intended users active in the period, divided by intended users on the list, expressed as a percentage.
What the sponsor sees: any team below [threshold percent] over [period] is raised with the sponsor, together with the old process that team still uses.
Break the figure down by team first. A program-wide rate can hide one group that is still working the old way.
Worked example: the list has [listed number] intended users in [team]. In [month], [active number] of them used the [new process]. Adoption is [active number] divided by [listed number], expressed as a percentage, which gives [adoption percent].
Agree a level with the sponsor, such as [threshold percent] for any team over [period]. If one group is low, the sponsor's help is usually a clear instruction, a removed blocker, or a decision to retire the old process. A general reminder to everyone is less likely to reach the group that is working the old way.
Ask the lead: Which team is lowest, and what are they still doing the old way?
How do you measure risk exposure?
Risk exposure is the sum, over open risks, of each risk's likelihood rating multiplied by its impact rating, on a scale agreed before the first review. Report it alongside the open risks that have no owner or no next review date, because a sponsor cannot act on a risk nobody owns.
Formula: the sum, over open risks, of likelihood rating times impact rating. Report beside it the open risks with no owner or review date, divided by open risks, expressed as a percentage.
What the sponsor sees: a risk at or above [threshold score] goes to the sponsor with the decision needed and the date it is needed by.
| Risk | Likelihood rating | Impact rating | Score |
|---|---|---|---|
| [risk name] | [likelihood rating] | [impact rating] | [score] |
| [risk name] | [likelihood rating] | [impact rating] | [score] |
Each score is the likelihood rating multiplied by the impact rating. Exposure is the sum of the Score column, shown as [exposure total]. Open risks with no owner or review date: [number] out of [number] open.
Agree the scale with the sponsor before the first review, and keep it the same for the whole program.
Ask the lead: Which open risk has no owner, and who can name one by [date]?
How long does a sponsor decision take?
Decision cycle time is the number of days from the day a decision was asked to the day the sponsor answered. Report the median for decisions closed in the period, and list open decisions with their age, so the lead can see whether waiting on the sponsor is slowing the team.
Formula: for each closed decision, the answer date minus the date the ask was sent. Report the median of those values.
What the sponsor sees: an open decision older than [threshold days] comes to the sponsor with a request for a decision or an answer date.
Worked example: the ask about [decision] was sent on [sent date] and answered on [answer date]. Cycle time for this decision is [answer date] minus [sent date], in days. The median covers all decisions closed in [period] and comes to [median days].
Agree the median trigger with the sponsor, such as [threshold days].
If the median rises, the sponsor sets an answer date for each open decision and says what the team should do until then. If no one answers, the decision stays open and goes back to the sponsor at the next review.
Ask the lead: What decision is waiting, by what date do we need it, and what can the team do until then?
How do you track open blocker age?
Open blocker age is the number of days each unresolved blocker has been open, counted from the day it was raised to the report date. Sort the list oldest first and name the party each blocker waits on, because an old blocker can mean a decision nobody has asked for yet.
Formula: report date minus the date the blocker was raised, for each open blocker. Also count the open blockers older than the threshold the sponsor agreed.
What the sponsor sees: with the threshold agreed as [threshold days], any blocker older than it goes to the sponsor with the decision needed to clear it.
Worked example: blocker [description], raised on [raised date], waiting on [party]. Age on [report date] is [report date] minus [raised date], in days. Above the agreed threshold of [threshold days]: [yes or no].
When a blocker passes the threshold, the sponsor makes the decision, names who will make it by a date, or accepts the delay in writing. If none of these happens, the blocker stays open and returns to the sponsor at the next review.
How do you set baselines and thresholds before the first review?
Set baselines and thresholds by putting every measure in a table before the first sponsor review, with its formula, data source, owner, refresh schedule, baseline, and threshold. Take each baseline from records kept before the change started, and ask the sponsor to confirm each threshold in the meeting.
- Build the table with a row for each measure. Columns: formula, data source, owner, refresh schedule, baseline, threshold.
- Take each baseline from records kept before the change started. The Quality-AI case study sets its first phase as identifying, baselining and measuring KPI metrics. It is a vendor's case, so borrow its order and not its results.
- Name an owner for each row, even when a tool calculates the number.
- Write each threshold as the point where the lead brings the item to the sponsor.
- Send the table before the first review with the note below.
Template for the note:
Subject: Please confirm the program measures for [program name] by [date]
[Sponsor name], the table attached lists the measures we will report at each review. For each one, please confirm the threshold that should bring it to you as a decision. I have marked the owners. If anything is missing or the thresholds need adjusting, reply by [date] and we will update the table before [meeting name].
What mistakes make the dashboard misleading?
Common mistakes include too many measures, a healthy-looking figure beside a stalled task, definitions that change between reviews, and no named owner. Cut any measure that drives no decision, freeze each definition for the whole program, and check the page against each one before the sponsor sees it.
The execviva guide lists dashboard overload, vanity metrics, inconsistent definitions, and unclear ownership as pitfalls. The other points are editorial judgment, not findings from a study.
- Too many measures. Cut any measure that drives no decision.
- A healthy figure beside a stalled task. Read the measures together. Benefits realization can look fine while schedule confidence falls, and low adoption can explain a benefit that has not arrived yet. Put related measures side by side in the review.
- Definitions that move. Freeze each definition for the whole program. If one must change, say so in the next status report and show the old and new values side by side.
- No owner. Every row in the table names a person.
Before the next sponsor meeting, send the definitions table and ask the sponsor to confirm which thresholds should trigger a decision.