Most comms dashboards fail not because the data is wrong but because nobody looks at them. A well-designed dashboard gets opened every morning by the CCO and drives the weekly leadership conversation. A poorly designed one gets glanced at once and forgotten. The difference is design discipline.
Start With Three Questions, Not Thirty Metrics
Before building anything, write down the three questions your most senior stakeholder needs answered each week. For a UK corporate comms team, those questions are typically:
1. Are we in the news, and is the tone positive or negative? 2. How are we performing against our competitors? 3. Is there anything emerging that we need to act on?
Every element on the dashboard should map to one of these questions. If a chart does not help answer any of them, remove it. A dashboard with five to eight well-chosen metrics will be used. One with 25 metrics is a report, and reports get filed, not read.
The Five Metrics That Work
Based on what actually gets used in UK comms teams running Meltwater, Cision, or Signal AI dashboards, these are the metrics that consistently drive decisions:
1. Coverage Volume Trend (With Context)
A simple line chart showing daily or weekly mention volume, with key events annotated. The annotations are what make this useful -- without them, a spike is just a number. Mark campaign launches, regulatory announcements, competitor news, and crisis events directly on the chart.
2. Share of Voice vs Competitors
A stacked bar or area chart comparing your brand against three to five named competitors over a rolling 90-day period. Use quality-weighted share of voice where possible -- a mention in the FT or BBC should count more than a syndicated press release on a content farm.
3. Sentiment Trend
A weekly sentiment trend showing the ratio of positive, neutral, and negative coverage. Avoid showing a single "sentiment score" -- it obscures the detail. Instead, show the trend line for negative coverage separately, since that is what leadership actually watches.
Flag any week where negative sentiment exceeds 25% of total coverage. In most UK corporate programmes, the baseline for negative coverage sits between 8% and 15%, so anything above 25% warrants investigation.
4. Message Pull-Through
Track whether your key messages are appearing in coverage. Define three to five core messages and measure how often they appear in Tier 1 and Tier 2 outlets. Report this as a percentage: "42% of coverage in Tier 1 outlets included our net-zero commitment messaging this month."
5. One Business Proxy
Connect media activity to a business signal. This could be website traffic from earned media referrals (track via UTM parameters), inbound enquiry volume, recruitment application rates, or investor sentiment scores. You will not get a clean attribution chain, but directional correlation is enough to keep PR measurement credible.
Layout Principles
Above the Fold
The top third of the dashboard should contain the three numbers your CCO cares about most. Use large-format cards:
- Total coverage this period (with % change vs previous)
- Share of voice rank (e.g., "2nd of 5 competitors")
- Sentiment flag (green/amber/red)
These should be visible without scrolling and readable in under five seconds.
Middle Section
Trend charts go here. Two or three time-series charts covering volume, sentiment, and share of voice. Keep the time axis consistent across all charts so patterns are easy to spot.
Bottom Section
Detail tables or drill-downs for the comms team. Top stories by reach, journalist bylines, outlet breakdown. This section is for practitioners, not executives.
Common Mistake: The Data Dump Dashboard
A UK energy company built a Brandwatch dashboard with 31 widgets covering every social platform, sentiment by sub-topic, influencer scores, hashtag clouds, emoji analysis, and geographic heat maps. The CCO opened it once, could not find the answer to "are we in trouble this week," and never opened it again. The comms team rebuilt it with five widgets and a one-paragraph narrative summary at the top. The CCO now checks it every Monday morning.
Add a Written Narrative
The single most effective addition to any dashboard is a three-to-five sentence written summary at the top explaining what changed, why, and what action is recommended. Automated dashboards cannot do this -- it requires a human analyst to interpret the data and write the insight.
Format:
This week: Coverage volume up 35% driven by FCA announcement on consumer duty. Negative sentiment spiked to 22% on Tuesday following Guardian investigation into branch closures. Competitor [X] gained share of voice with proactive ESG announcement. Recommended action: prepare reactive lines on branch closure programme ahead of expected follow-up coverage.
Tool-Specific Tips
- Meltwater: Use the custom dashboard builder and schedule PDF delivery to the exec team. Set up saved searches as dashboard widgets rather than trying to build everything from the global feed.
- Cision: The CisionOne dashboard supports custom layouts. Export to PowerPoint if your leadership prefers slide decks.
- Signal AI: Strong on narrative analysis -- use the "emerging narratives" widget if available, as it surfaces rising topics before they spike.
- Brandwatch: Best for social-heavy dashboards. Use the "signals" feature for anomaly detection rather than building manual threshold alerts.
- Google Data Studio / Looker: If your platform supports API export, consider building dashboards in a BI tool for more layout control and cross-source integration.
Review Cadence
Review the dashboard design quarterly. Remove any metric that has not been referenced in a leadership conversation for three months. Add new metrics only when there is a specific question they answer. Dashboard creep is the main reason dashboards stop being used.