Skip to content
Academy
Marketing Academy · Field Work●Analytics & Attribution
All projects·From: Dashboards (Looker, etc.)
MiniBuild the Asset· 30 minutes

Build the Two-Dashboard Split: Executive Scorecard and Operator Deep Dive Spec

Zendesk

Objective: Given Zendesk's list of 12 candidate metrics, sort each into the Executive Scorecard or Operator Deep Dive, then write the decision sentence for the 5-7 that make the cut.

You're the growth marketer at Zendesk. Leadership wants one exec-level dashboard and the paid media team wants a separate operator dashboard, and right now everything lives in one 20-tile mess.

Split 12 candidate metrics into the two dashboards the lesson defines, cut anything without a decision sentence, and spec the final layout.

Before you start

What you'll need

Free path (everything below is enough to finish)

Google Sheets
FreeSort metrics and draft decision sentences

Free, fast for a spec document

Looker Studio(optional)
FreeReference the layout hierarchy when finalizing the spec

Free tool the lesson recommends for actually building the dashboard next

The process

2 steps

Step 01 of 02

The Two Types of Dashboards You Need

The lesson splits every marketing dashboard into two audiences: an Executive Scorecard (CMO/VP, daily, no filters) and an Operator Deep Dive (campaign managers, filterable, hourly for paid).

Given Zendesk's 12 candidate metrics (blended CAC, ROAS by channel, CPL vs. target, pipeline created, top/bottom 10 creatives, spend pacing by campaign, budget spent vs. allocated, conversion rate by funnel stage, brand search volume, NPS, support-ticket-sourced signups, MQL-to-SQL rate), which 5 belong on the Executive Scorecard?

Google Sheets— A two-column sheet, one column per dashboard, one row per metric.

Procedure

  1. List all 12 metrics in column A
  2. Mark each as Exec, Operator, or Both
  3. Move any metric with no clear owner to a 'cut' list
Sample output
EXEC (5): Blended CAC, ROAS by channel, Pipeline created vs. target, Budget spent vs. allocated, MQL-to-SQL rate
OPERATOR (6): CPL vs. target, Top/bottom 10 creatives, Spend pacing by campaign, Conversion rate by funnel stage, ROAS by channel (filtered), Support-ticket-sourced signups
CUT (1): Brand search volume (no clear decision owner this quarter)

Healthy

Exec dashboard stays under 7 tiles with no filters needed; Operator dashboard is fully filterable by channel, date, and creative.

Unhealthy

The same 12-tile mess gets duplicated onto two pages instead of actually being split by audience and decision.

What this means

A metric belongs on exactly one dashboard by default; the same number appearing on both (like ROAS) is fine only when both audiences act on it differently.

So what do I do about it?

SymptomActionEffort
A metric doesn't clearly belong to either audienceCut it or find its owner before adding it to either dashboard5 min
YouYou can do this yourself, no engineering access required.

Step 02 of 02

Step 1: Write the Decision First

Every chart needs a finished sentence: 'If [metric] goes [above/below] [threshold], we will [specific action].' If you can't finish it, cut the chart.

Write the decision sentence for the 5 Exec Scorecard metrics from Step 1. Which one can't you finish?

Google Sheets— Add a 'Decision Sentence' column next to the Exec metric list.

Procedure

  1. Write one decision sentence per Exec metric
  2. Flag any metric where the sentence trails off into 'we will look into it'
  3. Cut or rewrite that metric before finalizing the spec
Sample output
Blended CAC: 'If CAC exceeds $110, we pause broad-audience campaigns and shift budget to retargeting.'
MQL-to-SQL rate: 'If MQL-to-SQL drops below 20%, marketing and sales run a joint lead-quality review.'
Pipeline created vs. target: (no finished sentence, cut from v1, revisit after Q1)

Healthy

Every surviving metric has a specific, finishable action attached, not a vague 'we'll investigate.'

Unhealthy

A metric ships on the final dashboard with a decision sentence that just says 'we will look into it.'

What this means

A metric that can't finish the sentence isn't ready for the dashboard yet, even if it's directionally useful.

So what do I do about it?

SymptomActionEffort
A KPI can't get a finished decision sentenceCut it from v1 and revisit once there's a real threshold to act on5 min
YouYou can do this yourself, no engineering access required.

Final deliverable

A two-dashboard spec: Exec Scorecard and Operator Deep Dive metric lists, each with a written decision sentence, ready to hand to whoever builds the actual Looker Studio report.

See a reference example
Sample output
MapmyIndia, dashboard spec (excerpt)

EXEC SCORECARD
  Blended CAC — 'If CAC exceeds ₹1,800, pause low-intent display campaigns.'
  Enterprise pipeline created — 'If pipeline falls 15% behind target, reallocate budget to the enterprise SDR-supported channel.'

OPERATOR DEEP DIVE
  CPL by campaign — filterable by state and product line

Success criteria

You're done when you can:

  • Sorts all 12 metrics into Exec, Operator, or Cut with a stated reason
  • Writes a complete decision sentence for at least 4 of 5 Exec metrics
  • Final Exec list stays within the lesson's 5-7 metric limit