11.12 · Research / Client dashboard concept
The client dashboard, made visible.
The specs already commit to a client-facing dashboard: 11.09 defines the summary-only client view, 11.10 places it as one of the two products on the shared data layer. This page turns those commitments into four working concepts, each pushing a different design principle to the front, all running the same sample data so they compare honestly. Nothing here is a live client read.
What the spec already commits to
The prototype invents no scope. Every element below traces to a written commitment in 11.09 or 11.10.
- One system, two views, one permission toggle. The internal Admin and Analyst dashboard is the working tool; the client view is the same system with a controlled, summary-only surface. Internal-first, not internal-only.
- Score. The daily Media Impact Score per product, with trend. How well the media performed, not how much was spent.
- Why. Every score ships with its decomposition: daypart mix, audience alignment, network strength, digital channel performance.
- Attribution. Lag-aware outcomes by channel: linear credited across multiple days, digital same-day.
- Narrative. An editable, generated plain-English draft in the voice a client actually reads.
- Send. One action: Teams internally, email to the client. Reporting and client communication become one motion.
- The honesty check. The baseline read is visible on the client surface, not buried: when media does not beat momentum, the dashboard says so.
- Internal extras behind the toggle. Benchmarking percentiles, anomaly flags with smoothing notes, component explanations, and exports appear only in the internal view.
Design principles
From the brand's own rules, applied to product surface for the first time.
- The client sees clarity, never the machinery. One score, one reason, one recommendation, one action per screen.
- Score and decision first. The biggest thing on the page is the read, not a chart gallery. Charts exist to explain the score, never to decorate.
- The narrative is the product. A client who reads only the paragraph should know what happened, why, and what Diray recommends next.
- Honesty is a feature with UI. The baseline panel has equal visual rank with the score. Trust is the differentiator; it gets pixels.
- One brand system. Ink, Broadcast Red, hairlines, mono labels: the dashboard looks like Diray, not like a BI tool with a logo pasted on.
Four concepts, one spec
Same data, same commitments, four different answers to "what is this product?" Each concept pushes one design principle to the front. The point of reviewing four is to choose a center of gravity, then steal the best moves from the others.
| Concept | The bet it makes | The room it wins |
|---|---|---|
| A · The Scorecard | The dashboard is a daily instrument panel: score, why, attribution, narrative, send, in cards. The safe, familiar shape. | Daily operator use. The vendor conversation. |
| B · The Daily Read | The narrative IS the product. A 40-second editorial brief where every number is tappable evidence. No chart gallery at all. | The CMO's inbox. The weekly client read. |
| C · The Decision Desk | Judgment is the product. A feed of keep / fix / reallocate calls, each with its receipt, organized around the weekly loop. | The weekly client call. Proving "operators answer for it." |
| D · The Signal Room | The read as broadcast. Dark, TV-native, one glance from across the room: dial, channel signal, the proof. | The QBR wall. New-business demos. |
Concept A · The Scorecard
The baseline concept. Interactive: flip the Client view / Internal toggle to see the permission model work. Send buttons are inert by design.
Concept B · The Daily Read
The editorial bet: if the narrative is the product, make the product a publication. One headline, one recommendation, one action. Tap any dotted number and its evidence opens inline; the charts exist only when summoned. This is the concept a CMO forwards.
Concept C · The Decision Desk
The judgment bet: clients are not buying charts, they are buying calls. Every card is a decision with its receipt: what we did, why, and the numbers behind it, organized around the Mon–Fri loop. Flip to Internal to see draft decisions waiting for the Friday call, invisible to clients until called.
Concept D · The Signal Room
The broadcast bet: the dashboard as a control room, one glance from across the room. Now fully animated on load (the arc draws, the score counts, the bars rise) and interactive on hover: every channel opens its read, the dial opens the why, every proof card carries one more line, every skyline day tells its score. Built for the QBR wall and the new-business demo.
The demo script
Five moves, in order, when presenting to the team, leadership, or the vendor. They run in any concept; the parenthetical tells you where each move lands hardest. Open the concept full screen first.
- Start on the score. "This is what a client sees first: how well yesterday's media performed against the outcome they care about. A number and a trend, not a wall of charts."
- Point at the honesty panel. "Every read ships with the baseline check. When media does not beat momentum, this panel says so. No competitor shows a client this."
- Walk the why. "The score always explains itself: what moved it and by how much. This is the interpretability requirement from the build brief, on screen."
- Read the narrative, then press send. "The paragraph is generated, editable, and one click from the client's inbox. Reporting and communication become one motion."
- Flip the toggle. "Same system, internal view: benchmarking, anomaly flags, exports. One permission toggle instead of two products. This is the two-dashboards spec from 11.09, working."
What this is not yet
- Not the vendor build. 11.09 defines what the vendor ships; this is the experience target to hold that build against.
- Not real data. Every number is sample. Wiring to live MIS output is a build decision, not a design one.
- Open decisions for the group: per-client theming, information density defaults, mobile behavior, and where planning tools eventually surface. Bring these to the same discussion as the V1 scope review on 11.09.