Skip to content
Academy
Marketing Academy · Field Work●Paid Ads
All projects·From: Connected TV & Streaming Ads
CoreSimulation· 40 minutes

Prove It Without a Click: Running a CTV Lift Study Under Pressure

Peloton Interactive

Objective: Navigate a simulated 6-week CTV campaign where last-click attribution shows almost nothing, deciding when to trust an incrementality lift study over the CFO's instinct to pull the budget.

You're the performance marketing lead at a Peloton-style connected fitness brand running your first CTV campaign, $40,000 over 6 weeks on Hulu and Roku, with a geo-lift test structured from day one.

Defend the channel through the weeks where last-click attribution looks broken, using the lift-study design correctly instead of reverting to last-click to make the CFO comfortable.

Before you start

What you'll need

Free path (everything below is enough to finish)

Google Analytics 4
FreeTrack view-through visits and site conversions as a directional secondary signal alongside the lift study

Free, already instrumented on most sites, useful context even though it's not the primary measurement method

Looker Studio
FreeBuild the weekly test-vs-holdout geo-lift comparison the CFO conversation actually depends on

Free, pulls straight from GA4 and Sheets, and produces the comparison chart described in the lesson's Step 4

The process

Simulation

Week 2 of 6

Week 2

Two weeks in. Google Analytics last-click reports show only 12 conversions attributed to "CTV" out of $13,300 spent so far. The CFO asks why you're not pulling the plug.

Sample output
Last-click view, Weeks 1-2
  CTV-attributed conversions: 12
  Cost per last-click conversion: $1,108

Geo-lift test (test markets vs. holdout markets), Weeks 1-2
  Test market trial signups: 640
  Holdout market trial signups: 512
  Lift so far: +25% (early, not yet significant)
Spend to date:$13,300 of $40,000
Budget remaining:$26,700

Last-click says CTV is nearly worthless. The geo-lift test, the actual measurement method the campaign was designed around, shows an early +25% signal. What do you tell the CFO?

Run this for real insteadOptional

The simulation above is a complete project on its own, free, no account or spend required. This is only for learners who want to test the same decisions against a real live account.

Minimum spend:$5,000 minimum to structure a readable geo-lift split across at least 2 test and 2 holdout markets
Minimum duration:28 days
Setup steps
  1. Pick matched test and holdout DMAs of similar size and baseline conversion rate
  2. Confirm the chosen self-serve platform or DSP supports geo-level exclusion targeting
  3. Set the frequency cap once at the DSP level before launch, not per-platform
Check-in schedule: Check test-vs-holdout delta weekly; do not change spend levels in test markets mid-flight

Final deliverable

A week-by-week defense log showing what last-click said, what the geo-lift test said, and which decision was made at each pressure point.

See a reference example
Sample output
FirstCry CTV flight log (excerpt)

Week 2: Last-click shows 8 conversions on ₹9.4L spend. Geo-lift early read: +18% trial signups, test vs. holdout, not yet significant. Holding the plan.
Week 4: Lift reaches significance at +23%. CFO approves remaining spend.
Week 5: Held frequency cap at 3/week despite a request to raise it; final lift held at +20% with stable brand sentiment.

Success criteria

You're done when you can:

  • Chooses the geo-lift study over last-click or a mid-flight budget cut at the week-2 pressure point
  • Holds the frequency cap at week 5 rather than trading sentiment for a rushed final push