lesson
35 min

A Dashboard That Tells the Truth

Build a one-page dashboard that follows the machine from signal to result and shows its health.

What you will be able to do

Design a one-page GTM dashboard with funnel numbers and health numbers, and calculate its rates with a short program.

One page, two kinds of numbers

A dashboard is one page that shows the numbers you need to make decisions. For a GTM machine, it needs two kinds:

Funnel numbers follow organizations through the machine:

  1. Signals found
  2. Enriched
  3. Scored above the threshold
  4. Approved by the reviewer
  5. Contacted
  6. Replied
  7. Meetings held
  8. Results (partners, funders, customers)

Health numbers tell you whether the machine is safe to keep running:

  • Bounce rate
  • Opt-out rate
  • Spam complaints
  • Reviewer disagreements with the AI score

Funnel numbers tell you if it works. Health numbers tell you if it is doing harm. You need both.

Rates, not just counts

"We contacted 40 organizations" means little alone. "10 of 40 replied, a 25 percent reply rate" lets you compare one month to the next. Each step's rate is that step divided by the step before it.

The code below calculates every rate from a made-up month of counts. Change the numbers and run it again.

If you are on phone data, note that the first run downloads the Python runtime (several MB). You can read the code here first:

funnel = [
    ("Signals found", 120),
    ("Enriched", 96),
    ("Scored above threshold", 40),
    ("Approved by reviewer", 32),
    ("Contacted", 32),
    ("Replied", 8),
    ("Meetings held", 5),
    ("Results", 2),
]

print("SAMPLE DATA - not real results")
previous = None
for step, count in funnel:
    if previous is None:
        print(step + ":", count)
    else:
        rate = round(count / previous * 100)
        print(step + ":", count, "(" + str(rate) + "% of previous step)")
    previous = count

contacted = 32
bounces = 1
opt_outs = 2
print("Bounce rate:", round(bounces / contacted * 100, 1), "%")
print("Opt-out rate:", round(opt_outs / contacted * 100, 1), "%")
Run the code with the sample month. Then change the numbers to test your own made-up month.
Python 3
Ctrl+Enter to run | Tab to indent | Python runs in your browser via Pyodide

Where the weak step is

Look for the biggest drop. If many organizations are approved but few reply, look at your first line and your door. If many reply but few meetings happen, look at how fast a person follows up. The dashboard does not fix anything. It shows you where to look.

Keep it honest

  • Show the bad numbers too. A dashboard that hides opt-outs is lying.
  • Never show personal data on a shared dashboard. Show counts and organizations, not private contact details.
  • Label sample or test data clearly, so nobody mistakes practice numbers for real results.
  • Start simple. A shared spreadsheet with one chart is a real dashboard.

Quiz: Check your understanding

  1. 01Why does a GTM dashboard need health numbers like opt-outs and complaints?

    Choose one.

  2. 02Many organizations are approved, but very few reply. Where should you look first?

    Choose one.

Go deeper: Business Intelligence for Real Operators: Turning CRM & ERP Data into Decisions

Ask. Answer. Build together.