1. Introduction
This document sets out the general methodology used to produce the Welsh Devolution Activity Monthly Review Report. The monthly review synthesises weekly open-source intelligence (OSINT) monitoring into a single strategic-level product. Its purpose is to track, contextualise, and assess the evolving online discourse surrounding Welsh devolution, independence, and related constitutional questions.
The report is designed around qualitative insight rather than raw metrics. Although quantitative data (follower counts, view counts, engagement figures) is captured during weekly monitoring, the monthly review foregrounds narrative analysis, audience behaviour, and influence mapping to give decision-makers a richer understanding of who is shaping the debate, how, and why.
The methodology is structured around three core analytical sections, each of which addresses a distinct intelligence question. These sections are: the Narrative Timeline, the Persona and Algorithm Analysis, and the Influencer Network and Evidence Board.
2. Report Production Cycle
The monthly review follows a two-stage production cycle that moves from weekly tactical reporting to monthly strategic synthesis.
2.1Weekly Monitoring
Throughout the reporting month, weekly devolution activity analysis reports are produced. These capture platform-specific data across Instagram, X, Facebook, YouTube, TikTok, and other relevant platforms. Each weekly report records key events, debate intensity, cross-platform analytics, recurring myths and strategic implications for TCE.
2.2Monthly Synthesis
At the end of the reporting period, the weekly reports are reviewed in their entirety. All numbers, insights, and qualitative observations are cross-referenced to identify trends, recurring themes, and inflection points that may not be apparent at the weekly level. The monthly review then distils these findings into three analytical sections, described in detail below.
3. Section 1: Narrative Timeline
3.1Purpose
The Narrative Timeline provides a chronological account of the key events and developments that shaped online devolution discourse during the reporting month. It is intended to give readers a clear picture of what happened, when, and how events connected to one another.
3.2 Data
Source material is drawn from the weekly reports, which capture activity across multiple social media platforms, news outlets, political communications, and community forums. Events are not limited to those formally planned; in fact, a key observation is that many significant moments in the devolution debate appear organically and at short notice, including pop-up rallies, viral posts, and spontaneous commentary responding to breaking political news.
3.3 Analytical Approach
Events are plotted chronologically and grouped into thematic clusters where appropriate (for example, responses to a party conference, reactions to a government announcement, or a viral content cycle). The analysis assesses the following dimensions for each event or cluster:
- Nature and origin: Was the event planned, reactive, or organic?
- Platform distribution: Where did the conversation take place and how did it spread across platforms?
- Audience response: How did different audience segments react, and did the event shift sentiment or mobilise new voices?
- Political context: How does the event relate to the broader political landscape, including party conferences, elections, policy announcements, and legislative developments?
3.4 Election and Conference Coverage
The monthly review pays particular attention to political party conferences and election-related events, as these serve as focal points for pro-devolution, anti-devolution, and neutral discourse. The analysis notes the degree to which online activity around these events is orchestrated versus organic, the speed at which narratives form, and the extent to which party messaging is amplified, challenged, or ignored by online audiences.
3.5 Output
The final output for this section is a written timeline that maps the month's key moments, highlights causal connections between events, and provides analytical commentary on the overall direction of the narrative arc.
4. Section2: Persona and Algorithm Analysis
4.1Purpose
This section examines how different types of online users experience the devolution debate differently, based on who they are, what they engage with, and how platform algorithms curate content for them. It answers the question: does everyone see the same debate, or are different audiences living in fundamentally different information environments?
4.2 Persona Construction
A set of defined personas (Wales personas previously created) is maintained to represent distinct audience segments within the Welsh devolution landscape. Personas are constructed based on observable behavioural and demographic characteristics and may include profiles such as: a politically engaged Welsh nationalist, a unionist with UK-wide media consumption habits, a younger Welsh-language user on TikTok, a rural community member engaging primarily through Facebook groups, and so on.
The persona list is provided separately and may be updated between reporting periods as the audience landscape evolves. Each persona is not a single real individual but a composite profile representing a recognisable audience type.
4.3 Algorithm Differentiation
For each persona, the analysis examines how platform algorithms shape the content they are likely to see. This is achieved by reviewing the types of content surfaced to users matching each persona's profile, including recommended videos, suggested posts, trending topics, and search results. The key question is how algorithms differ between personas, and the extent to which algorithmic curation reinforces existing views, introduces users to new perspectives, or channels them toward more extreme content.
Factors considered in the algorithm analysis include:
- Content recommendations: What does each platform suggest to users who match this persona's engagement history?
- Echo chamber indicators: Is the persona being served a narrow or broad range of perspectives?
- Cross-platform journeys: Does a persona who begins on one platform get funnelled to another (for example, from a Facebook group to a YouTube channel)?
- Linguistic and geographic signals: How does language preference (Welsh vs. English) or geographic location affect the content served?
4.4 Matching Personas to Online Activity
Each persona is matched against the observed online activity captured in the weekly reports. This matching process identifies which personas are most active, which are growing or declining in visibility, and which are most susceptible to influence from key actors. It also highlights where personas overlap; for instance, a rural Facebook user who also consumes YouTube farming content that carries a devolution message.
4.5 Output
The output for this section is a comparative analysis of the personas, showing how each experiences the devolution conversation differently. It highlights the algorithmic mechanisms that create divergent realities and identifies potential vulnerabilities or opportunities in how different audiences are reached.
5. Section 3: Influencer Network and Evidence Board
5.1Purpose
This section identifies the key individuals, organisations, and accounts that are driving the devolution debate online. A recurring finding is that only a handful of people are genuinely influencing the pro-devolution, anti-devolution, and neutral conversation in Wales. The purpose of this analysis is to determine who is really pulling the strings, identifying the true origins of debate rather than its surface-level amplifiers.
5.2 Influence Identification
Influential actors are identified through a combination of quantitative reach metrics and qualitative assessment of their role in shaping narratives. The analysis distinguishes between several types of influence:
- Originators: Accounts that consistently introduce new arguments, framings, or calls to action that are subsequently picked up by others.
- Amplifiers: Accounts with large followings that spread content created by others, giving it reach without necessarily originating the ideas.
- Connectors: Accounts or individuals who bridge different communities; for example, linking a Welsh-language activist community with an English-language media audience.
- Authenticity anchors: Public figures or content creators whose perceived authenticity lends credibility to the broader movement (for example, a well-known farmer whose lifestyle content carries a devolution message).
5.3 Network Mapping
Relationships between influential actors are mapped to reveal coordination, mutual amplification, and information flow. The network map is structured in tiers:
- Tier 1 - Core influencers: The small number of accounts that originate the majority of significant narratives.
- Tier 2 - Secondary amplifiers: Accounts with meaningful reach that consistently amplify Tier 1 content.
- Tier 3 - Media and institutional amplifiers: News outlets, opinion platforms, and institutional accounts that give the debate mainstream visibility.
Connections are categorised as coordinated (evidence of direct collaboration or cross-promotion) or organic (parallel activity without apparent coordination). The distinction between pro-devolution, anti-devolution, and neutral ecosystems is maintained throughout.
5.4 Evidence Board Methodology
The evidence board is a visual analytical tool modelled on a detective's investigation board. It is designed to help connect the dots between actors, events, and narratives that might otherwise appear unrelated. The board plots:
- Key actors and their roles (originator, amplifier, connector, authenticity anchor).
- Connections between actors, showing the nature and direction of information flow.
- Key events from the Narrative Timeline that are linked to specific actors.
- Platform presence for each actor, showing where they are most active and influential.
- Growth and engagement trends drawn from the weekly data.
The evidence board is intended to provide an at-a-glance strategic view of the influence landscape, making it easy to identify central nodes, isolated actors, emerging players, and declining voices.
5.5 Output
The output for this section is a tiered network map accompanied by the evidence board visual, with written analysis explaining the key findings, notable changes from the previous month, and any emerging actors or shifting alliances.
6. Cross-Section Integration
Although each section addresses a distinct analytical question, all of them are designed to be read as an integrated whole. The Narrative Timeline establishes what happened; the Persona and Algorithm Analysis reveals who saw it and how; and the Influencer Network and Evidence Board identifies who made it happen. Together, they provide a comprehensive picture of the devolution discourse ecosystem.
Cross-references between sections are made explicit where appropriate. For example, if a key event in the timeline was initiated by a Tier 1 influencer identified in the evidence board, and was algorithmically amplified to a specific persona group, this chain of cause and effect is drawn out in the analysis.
7. Data Sources and Limitations
7.1 Sources
The report draws on open-source data from the following platforms and sources:
- Social media platforms: Instagram, X, Facebook, YouTube, TikTok, and other relevant platforms as they emerge.
- News media: Welsh and UK-wide news outlets, including online editions and opinion sections.
- Political communications: Official party statements, press releases, conference coverage, and election materials.
- Community forums and groups: Public Facebook groups, Reddit communities, and other online gathering spaces relevant to Welsh politics.
7.2 Limitations
The methodology operates within several known constraints:
Data collection is limited to publicly available content and does not include private messaging, closed groups, or content behind paywalls unless otherwise noted. Algorithmic analysis is inferential; it is based on observed content surfacing patterns rather than direct access to platform recommendation engines. Persona matching relies on composite profiles rather than tracking specific real individuals. Finally, the dynamic nature of online discourse means that certain activity may emerge between weekly reporting intervals and is incorporated into the monthly synthesis retrospectively.