Leadership tools · Version 1.0 · 4 November 2026

SPARK · The funding test for indicators

Bruno Hounkpati, Founding Partner and CEO

SPARK is a five-criterion test that decides whether a lead indicator deserves a budget line: signal strength, pattern cohesion, anchored alignment, resilience and knowledge velocity. Score each criterion 1 to 5 and fund the measure only when no score is below 3 and the average is 4 or more.

What it answers

Does this indicator deserve a budget line?

When to use it

Before funding any lead indicator, including each LACAS indicator.
When a dashboard keeps growing and nobody acts on half of it.
When a measure starts to look too good. Pressure to hit a number weakens what the number tells you (Goodhart, 1975).

The five criteria

	Criterion	What it means	The question to ask
S	Signal strength	Predicts results 6 to 18 months out.	Over the past two or three years, did this indicator move before results did?
P	Pattern cohesion	Moves with industry and economic shifts.	When the sector or the economy turned, did this indicator register it?
A	Anchored alignment	Fits the strategy and stakeholder priorities.	Which strategic priority does it serve, and who owns it?
R	Resilience	Hard to game or distort.	Could a team improve the number without improving the business?
K	Knowledge velocity	Teaches you fast enough to adjust.	Does it update before the decision it informs is due?

Scoring anchors

	1	3	5
S	No link seen with later results	A link, but irregular or less than 6 months ahead	Moves 6 to 18 months ahead, checked on your own history
P	Blind to outside change	Registers it late	Registers it as it happens
A	No named priority, no owner	Loosely linked to the strategy	Tied to a named priority, owned by someone with the decision rights
R	Easy to move without real change	Can be moved with some effort	Moves only with real work, cross-checked against a second source
K	Yearly, or after the decision	Quarterly	Monthly or faster, inside the decision window

Scores of 2 and 4 sit between the anchors.

The rule

Score each criterion 1 to 5. Fund a measure only when no score is below 3 and the average is 4 or more. Otherwise fix it or drop it.

The SPARK rule as a decision path.

Why both conditions. A strong average can hide one fatal weakness. A measure that is easy to game misleads you however well it predicts. The floor stops one high score from buying back a flaw. The average stops a measure that is merely passable everywhere.

Two conditions, both required. One weak score cannot be bought back by strong ones.

Show the underlying numbers

Measure	S	P	A	R	K	Average	Decision
Share of sales from products launched in the past three years	5	4	4	2	5	4.0	Fix

Worked examples (illustrative scores)

Five illustrative measures against the floor at 3.

Show the underlying numbers

Measure	S	P	A	R	K	Average	Decision
AI spend	1	2	2	1	1	1.4	Drop
Key decisions built on AI forecasts, checked against results	5	4	4	4	5	4.4	Fund
Succession plan on file	1	2	3	1	1	1.6	Drop
Key roles with a successor who has already done the job	5	3	5	4	4	4.2	Fund
Share of sales from products launched in the past three years	5	4	4	2	5	4.0	Fix

Measure	S	P	A	R	K	Average	Decision
AI spend	1	2	2	1	1	1.4	Drop
Key decisions built on AI forecasts, checked against results	5	4	4	4	5	4.4	Fund
Succession plan on file	1	2	3	1	1	1.6	Drop
Key roles with a successor who has already done the job	5	3	5	4	4	4.2	Fund
Share of sales from products launched in the past three years	5	4	4	2	5	4.0	Fix

The last line. The average passes. The floor does not. Relabelling old products lifts the number without any new product. Fix it: define what counts as new before the year starts, have finance check it, then score again.

Common mistakes

Scoring alone. Score with the owner of the measure and with one person who does not own it. A gap of 2 points or more is the conversation to have.
Scoring the intention. Score the measure as it is collected today, not as it will be once the new system is live.
Letting the average decide. One score below 3 means fix or drop, whatever the average.
Letting green measures skip the test. An offset bought, a policy signed or a report filed scores low on resilience. Emissions per tonne, checked against metered energy, can pass.
Testing once a year. A measure marked SPARK focus is tested again the next quarter.
Keeping a dropped measure “for information”. If it informs no decision, stop collecting it.

One-page template

One row per measure: measure, owner, S, P, A, R, K, lowest score, average, fund, fix or drop, fix needed, next test date.

Download the template (PDF)

What it builds on

Kusek and Rist (2004), Ten Steps to a Results-Based Monitoring and Evaluation System, World Bank, step 3. The “CREAM” of good performance indicators, a checklist for whether an indicator is usable.
McChesney, Covey and Huling (2012), The 4 Disciplines of Execution. Lead measures.
Goodhart (1975). Pressure to control a number weakens what it tells you.

What SPARK adds. Criteria built for lead indicators that must earn funding (prediction horizon, link to the outside economy, resistance to gaming, learning speed) and a rule under which no single strength can buy back a weakness.

Frequently asked questions

What does SPARK stand for?

Signal strength, Pattern cohesion, Anchored alignment, Resilience and Knowledge velocity.

What is the SPARK rule?

Score each criterion 1 to 5. Fund a measure only when no score is below 3 and the average is 4 or more. Otherwise fix it or drop it.

Why is the average not enough?

A strong average can hide one fatal weakness, such as a measure that is easy to game.

How does SPARK differ from SMART and CREAM?

SMART (Doran, 1981) checks how an objective is written. CREAM (Kusek and Rist, 2004) checks whether an indicator is usable. SPARK decides whether a lead indicator earns a budget line.