First, the key terms.

PV
Planned value: budgeted value of work scheduled by a date.
EV
Earned value: budgeted value of work actually completed.
AC
Actual cost incurred for the measured work.

The idea, explained

A risk is an uncertain event or condition, while an issue has already occurred. Record causes, consequences, owners, responses and triggers, not just colored ratings. During delivery, progress must be tied to accepted work rather than optimistic percentage estimates. Earned value combines budgeted work and actual cost under a defined baseline. It supplements—not replaces—quality, benefits and model-risk measures. Closure is a controlled transition to ownership and support, not simply the date a team stops coding.

STEP 1

Plan responses and ownership

Choose avoidance, mitigation, transfer or acceptance where appropriate, and identify residual risks. A risk register needs review dates and triggers. Data-access failure, label quality and deployment harm deserve attention alongside schedule risks.

STEP 2

Interpret progress consistently

Cost performance index is EV divided by AC; schedule performance index is EV divided by PV. Values below one can flag cost or progress shortfalls under the baseline. A schedule index is not a direct number of calendar days late.

STEP 3

Close with a usable handover

Confirm acceptance criteria, support ownership, data and access arrangements, remaining defects, contractual obligations and benefit measurement. Capture lessons about failed assumptions so later teams can act on them. An owner must accept any remaining operational risk.

Follow a worked example.

A fictional pilot has PV $10,000, EV $8,000 and AC $12,000 at the reporting date. Its model also fails a required subgroup quality check. These are separate kinds of evidence that a status report must not average away.

Walk it through, one step at a time

  1. CPI = 8,000 ÷ 12,000 ≈ 0.67: each dollar spent has earned about $0.67 of budgeted work value under this baseline.
  2. SPI = 8,000 ÷ 10,000 = 0.8: less budgeted work is complete than planned. Neither ratio establishes that the model is safe.
  3. Investigate causes and response options. Do not close as successful until acceptance and risk criteria are addressed and the receiving owner can operate the system.

Where people get stuck

Completing work that fails acceptance criteria can inflate apparent progress; earned value depends on credible definitions of completion.

Now try it yourself.

Pause before opening each answer. Explain your reasoning in your own words.

1. Check the core idea

Does spending less than the planned budget prove favorable performance?

Reveal the answer & reasoning

No. Less may have been delivered, quality may be unacceptable or important costs deferred. Compare completed scope, actual cost, timing and outcomes.

2. Apply what you learned

A model is ready but no team accepts monitoring responsibility. Is handover complete?

Reveal the answer & reasoning

No. Operational ownership is part of the deliverable. Closing without it leaves failures and user impacts unmanaged.