From policy gap to intervention logic

From policy gap analysis to intervention logic and SMART KPIs

main tellingdepth: standard3 min markdown

D3.1, the PoliRural+ RAP Methodology, asks that the measures in a regional action plan be complete, avoid overlaps with ongoing initiatives and build on what exists. Getting there takes four steps.

flowchart LR
  A[Inventory of existing policies] --> B[Gaps: finance, legislation, training]
  B --> C[2-5 challenges, 2-10 measures each]
  C --> D[Intervention logic per measure]
  D --> E[SMART impact KPIs and a monitoring committee]

1. Inventory what already exists

Start by listing the policies, strategies and plans on the theme: those in force, in preparation, or expected soon. D3.1 gives a 2014 example from Bosnia and Herzegovina: more than 40 current policies referred to skill needs, and many of the measures they proposed had never been carried out. Reading such documents shows who the stakeholders are, which agencies are involved and where money might come from. D3.1 also recommends talking to local officials, which it calls “Management by Talking Around”.

2. Look for the missing measures

Action plans are often incomplete. D3.1 names three gaps to check:

• Finance: no single EU fund pays 100 %, so plans combine CAP, ERDF, RRF, the Cohesion Fund, the Just Transition Fund and others, plus EIB loans or green bonds. Technical assistance such as the EIB’s ELENA programme can help.
• Legislation: money is in place, but a law blocks the measure, or no law enables it. Energy communities are a known case.
• Training: officials need to understand concepts such as the circular economy or the bio-economy before they can support them.

3. From challenges to measures

The vision work yields 2 to 5 challenges. Each becomes 2 to 10 measures. Over time the work moves from beneficiaries to the actors who will implement the measures, because “only the ‘actors’ are able to make those measures happen”.

4. Intervention logic and KPIs

The intervention logic is a narrative that links a measure to the rest of the mix and to the challenge it addresses. It can be written as a logframe. D3.1 illustrates it with training:

Indicators should be SMART: specific, measurable, achievable, relevant and time-bound. D3.1 advises putting most effort into impact KPIs, because attention tends to stop at outputs once contracts are checked. Impacts often take 3 to 10 years to appear. A good plan also sets up a monitoring committee, led by organisations able to hold local government to account, and a budget for interim evaluations and mid-course corrections.

The next block applies these four steps to Vidzeme.

Worked example: Vidzeme turns a monitoring gap into a measure plan

another tellingexample world: latviadepth: standard3 min markdown

The Vidzeme Planning Region (VPR) in Latvia leads PoliRural+ pilot 8. Its starting point is a plain gap: no systematic territorial monitoring of the VPR Development Programme. Below, the region’s policy brief (D7.3, Annex 8, 11 June 2025) is read through the four steps of the gap-to-intervention-logic method.

Step 1: what already exists

The brief’s strategic alignment table works as the policy inventory. It lists Latvia’s Regional Policy Guidelines 2021–2027, the VPR Sustainable Development Strategy 2030, the VPR Development Programme 2022–2027, the regional Smart Specialisation Strategy (seven high-value clusters), Latvia’s CAP Strategic Plan 2023–2027, the Green Deal, the Digital Decade 2030, Cohesion Policy, Horizon Europe and Digital Europe. The plan builds on the existing programme. It does not create a new one.

Step 2: the gaps, by type

Step 3: measures

• Standardise monitoring: unified data-collection protocols, mandatory annual reporting cycles, formal data-sharing agreements
• Launch VPR PATHWATCH, a centralised, AI-driven monitoring system with geospatial and mobile access
• Create permanent specialist posts and train municipal staff in data literacy and AI
• Bring citizens in through a mobile app, for accountability
• Position VPR as a national model and build a Baltic monitoring network

PATHWATCH is a recommendation in the brief, not a running system.

Step 4: targets

The brief sets three targets for 2028: integrated multi-level data governance, an AI-driven geospatial monitoring platform that moves “from retrospective reporting to predictive analysis”, and regional analytics capacity. These are time-bound, so they meet the T in SMART. To make them fully SMART, each still needs a measurable indicator and a baseline. An illustrative chain (constructed here, not taken from the brief): training sessions (measure) → municipal staff trained (output) → municipalities reporting on the unified protocol each year (outcome) → regional programme decisions made on current data (impact).

The evidence came partly from PoliRural+ open calls. InnoMatrix carried out the EDQMS_VR outreach study on economic development quality monitoring, and NovoFutura ran D2D, a two-phase hackathon on regional development data. Chapter 4 calls this path data to decisions.

Check yourself: Your Region's First Flagship Measure

You lead a regional development agency. Your gap analysis shows that development data is scattered across municipalities and ministries, indicators do not match, and nobody can say whether the regional development programme is working. The board wants one flagship in the new action plan. What do you put first?

A) “Build the platform” Commission an AI-driven geospatial monitoring dashboard now. A visible tool will pull the data together.

B) “Count what we do” List the activities (training sessions, data workshops, platform launch) and report how many were delivered each year.

C) “A complete, measurable mix” Build on the existing programme. Add data-sharing agreements, annual reporting and staff training next to the platform, with a financing mix for each. Draft them with the agencies that will run them, write an intervention logic for each, and track SMART impact KPIs through a monitoring committee.

D) “Use the fund we know” Apply to the local LEADER group for the whole programme, because the team has used it before.

More than one option has a defensible core. Choose one, then read why it works or where it breaks.