Access & Data

Your ad accounts, data warehouse and conversion events connected and validated.

Timing:

Weeks 1⁠–⁠4

1.1

Kick-off and ad platform access

In the kick-off we align on attribution, channels and the metric the pilot will be judged on. Your marketing team then gives Churney access to your ad accounts and, if you use one, your MMP.

When:

Week 1

Your team:

Marketing

1.2

Sync your data warehouse

Your data engineer shares historical and incremental data warehouse views with Churney and we hold a data integration meeting to align on how and what data is shared.

When:

Weeks 1⁠–⁠2

Your team:

Data

1.3

Create your conversion events

Your team creates the baseline conversion events in Meta or Google (or your MMP) and confirms the event and external IDs we match on. This is the step that most often delays a pilot, so we start it on day one.

When:

Weeks 1⁠–⁠4

Your team:

Marketing, Developer

1.4

Tag your data for Churney’s model

Together, using our data tagger tool, your data engineer helps us map and tag critical events in your data so we align on the business logic behind it.

When:

Week 3

Your team:

Data

1.5

Data validation meeting

We share reporting and aggregate statistics to validate our understanding of your data and identify business objectives. You confirm the revenue and active user figures look correct.

When:

Week 4

Your team:

Data

Model, Sizing & Test Design

Churney builds a predictive LTV model on your data, sizes the opportunity and agrees the test design with you.

Timing:

Weeks 5⁠–⁠6

2.1

We build your model

Our data science team trains your prediction model on your data and recommends the best campaign type (max conversions or max value) for the experiment.

When:

Week 5

Your team:

Churney

Milestone #1

2.2

Opportunity sizing meeting

We present the initial predictions and the expected impact on your business, and align on campaign setup and key metrics.

When:

Week 6

Your team:

Marketing, Data

2.3

Sign off the test design

Before anything goes live we agree in writing on the budget per arm, the run length, which weeks count as learning phase, the primary metric and the uplift that counts as success. The read-out is judged against exactly this.

When:

Week 6

Your team:

Marketing

Signal, Experiment & Read-out

We set up and check the signal, run the BAU vs. Churney experiment, and read out the result against what we agreed.

Timing:

Weeks 7⁠–⁠17

3.1

Set up the signal

We map your baseline events to measure signal loss and set up the campaign-specific Churney signal, sent to the ad platform directly or through your MMP. Your team creates the matching conversion event on the ad platform side.

When:

Week 7

Your team:

Churney, Marketing

3.2

Live signal check

The signal runs live for a week before launch while we check volume, match quality and value coverage, so the experiment starts on a signal we know works.

When:

Week 8

Your team:

Churney

Milestone #2

3.3

Launch the experiment

The BAU vs. Churney experiment goes live at the start of week 9, with the budgets agreed in step 2.3.

When:

Week 9

Your team:

Marketing

3.4

Learning phase

The ad platform learns the new signal while we check volume, audience and the model frequently. These two weeks are excluded from the evaluation.

When:

Weeks 9⁠–⁠10

Your team:

Churney

3.5

Measured period

Four clean weeks of data for the evaluation. We monitor signal volume and performance, and campaign settings stay unchanged unless both sides agree.

When:

Weeks 11⁠–⁠14

Your team:

Churney

Milestone #3

3.6

Read-out and decision

Cohorts mature for two weeks, then we present the result against the success criteria from step 2.3 and agree on the next step: scale up, extend or stop.

When:

Weeks 15⁠–⁠17

Your team:

Marketing, Leadership

From kick-off to a decision in four months.
Churney's pilot, week by week.

Modeling

Analyzing data and building predictive lifetime value models

Sharing

Providing access to campaigns, historical user and revenue data

Activating

Sending predictions to ad networks for optimization

Model, Sizing & Test Design

Signal, Experiment & Read-out

See whether your data predicts long-term value.

From kick-off to a decision in four months.Churney's pilot, week by week.

Modeling

Analyzing data and building predictive lifetime value models

Sharing

Providing access to campaigns, historical user and revenue data

Activating

Sending predictions to ad networks for optimization

Access & Data

Model, Sizing & Test Design

Milestone #1

Signal, Experiment & Read-out

Milestone #2

Milestone #3

See whether your data predicts long-term value.

From kick-off to a decision in four months.
Churney's pilot, week by week.