Qualitative KPIs in Public Affairs: From Vague to Verifiable

By Paul Shotton, Advocacy Strategy

Leadership has started asking public affairs teams a sharper question. Not just “what did you achieve?”, but “how do you know the work is doing anything?” It is a fair question, and a hard one — because advocacy rarely offers clean proof. The political process, as we put it in our training, is more like gears than a railway. Outcomes arrive through many hands, and you can seldom show that your action alone caused the result.

That is why the most useful measures in public affairs describe the “how” — the quality of the work and the movement it produces — rather than only the “what” of the final outcome. And the measures that capture this best are usually qualitative. They are also the ones teams get wrong most often: left so vague they mean nothing, or quietly swapped for a number that is easy to count and tells you little. This piece is about writing the qualitative ones well.

Why the qualitative indicators are the hard ones

In advocacy you rarely prove causation. Several organisations are working the same file, positions move for reasons you cannot see, and the decisive conversation often happens in a room you were not in. That uncertainty makes counting tempting, because counts feel solid. But a count can be a vanity metric — impressive on the surface and empty underneath. A thousand impressions without a discussion in committee is not a success, if a discussion in committee was the point.

The indicators that actually matter capture meaning rather than motion: whether a policy position has been recognised, whether a piece of legislative text has shifted, whether a stakeholder relationship has been repositioned. These are narrative outcomes, effortful to articulate and to assess — which is precisely why they get dodged, and precisely why they belong at the top of any scorecard.

Three ways teams get them wrong

The first is the vague qualitative indicator. When we reviewed one organisation’s KPIs, the weakest read “national procedures are considered more predictable” and “improved alignment between countries.” Each names a direction, but offers no evidence, no baseline, and no way to tell at year end whether it happened. The second is the vanity metric — meetings held, papers filed, likes and impressions — counts that look like progress but are not tied to any outcome. The third is the indicator with no baseline and no means of verification: even a well-phrased outcome is unusable if you cannot say where it started, or where the proof will come from.

How to write a good one

Start from the outcome, not the activity. In our Advocacy Academy training we use a simple progression. “Hold ten stakeholder meetings” is an activity count. “Hold meetings with ten of our fifteen key identified stakeholders” is better — it adds quality and a target. “Hold meetings with ten of fifteen key stakeholders such that our issue is actively raised cross-party in committee” is a real KPI, because it carries the outcome the meetings are meant to produce. The discipline is to keep adding the clause that answers “so that what?” until the indicator points at meaning.

Stage	Indicator
Activity count	“Hold ten stakeholder meetings.”
Better: quality + target	“Hold meetings with ten of our fifteen key identified stakeholders.”
Strong: adds the outcome	“Hold meetings with ten of fifteen key stakeholders such that our issue is actively raised cross-party in committee.”

Then name the means of verification — the single most useful column most scorecards leave out. Where will the evidence come from: a committee report, a Council text, a regulator’s communication, a panel ruling? An indicator without a means of verification is a vibe; with one, it is an indicator. Give it a baseline, a milestone, and a target as well — where you are now, where you expect to be at a checkpoint, and what success looks like at the end.

For genuinely soft outcomes — how a meeting went, whether a relationship is moving — build a small rating framework rather than pretending precision. Code each instance consistently across the team: contributed, neutral, or went nowhere. Tteams can diligently log every stakeholder meeting yet have no way to tell one that moved their file from one that achieved nothing; a simple, agreed coding scheme fixes that without inventing false numbers.

Finally, test the indicator. We ask whether it is valid, objective, reliable, accessible, useful, and owned — does it represent the objective, is it unambiguous, is it consistent over time, can it be collected at reasonable cost, is it useful for a decision, and do the people who must act on it agree it makes sense. Keep two to four indicators per outcome. Qualitative indicators are expensive to draft and to assess, so spend them only where meaning lives — at the goal and objective layers.

The CBAM file, worked through

Take a fictional manufacturer, whose finished product depends on a material covered by the EU’s Carbon Border Adjustment Mechanism. At the goal layer, a qualitative indicator might read: downstream-manufacturer competitiveness is named as a consideration in the Commission’s CBAM extension proposal — verified in the published proposal text, against a baseline where it is absent today. At the objective layer: the covered input for finished products is treated proportionately in the negotiated scope — verified in the committee report or Council general approach, with one supporting count, that the proportionate-treatment ask is reflected in at least one committee or Council text.

At the strategy layer, one indicator per pillar: the covered-input concern is acknowledged by the rapporteur or shadows in scope discussion, verified in committee debate and draft amendments. Notice the contrast. “Make CBAM fairer for downstream manufacturers” is an aspiration. “Eight meetings held with DG TAXUD” is motion. “The covered-input concern is acknowledged by the rapporteur in scope discussion, evidenced in committee amendments, up from no recognition today” is an indicator you can defend to a board.

A closing test

A golden rule is to write every indicator as though it might be made public. It keeps the language honest and the evidence real. Artificial intelligence can help with the mechanics — drafting candidate indicators from a priority list, clustering narrative signals, keeping a scorecard current — but it cannot decide what counts as evidence or supply the means of verification. That judgement stays with you.

So the test for any public affairs KPI is simple. Can you say what would count as success, where the proof will come from, and where you stand against it today? If you can, you have an indicator. If you cannot, you have an aspiration — and no dashboard will rescue it.