Experiment Cost Calculator


Use the Experiment Canvas to map out your experiment.

Use the Experiment Cheatsheet to see when to use this experiment.


Landing Page Experiment with a single ad

Currently this calculator supports the landing page experiment with a single ad. The purpose of this experiment is to test if the target customer is enticed by the value proposition in the ad to click through to the landing page.


Experiment Structure

The experiment consists of a facebook or google ad, where the user clicks the ad and gets redirected to a landing page. We measure the conversion of the ad to the landing page, and compare it against a benchmark conversion rate for the sector. We assume that the differences between the clickthrough rate for this ad and the benchmark tell us if the ad resonates with the target audience. We count clickthrough on the ad as conversions.


When is the experiment a success

Benchmark
5.00%
Expected conversion
7.50%

Where to find benchmark values:


Calculate the outcome for your experiment

Impressions
100
Conversions
50
Outcome
<5  
 
 >5
 

Outcome: validated

For 100 impressions, the benchmark would expect 5 conversions. You have scored 50 conversions, which is 45 conversions more than expected, exceeding the needed difference 0. This means you can declare this experiment validated.

Note: You have reached the minimum number of results needed to declare the experiment validated or invalidated. Continuing the experiment will not give you extra information. You can safely stop the experiment.


Cost Per Click (CPC)

To calculate the cost based on advertising, you need to know the Cost per Click (CPC) of your ad. The CPC rate depends per ad network and per segment. Below you can find some links that can help you find the CPC for your ads.


CPC:
$1.50

Where to find CPC rates:


Expected minimum cost: $0.00

This will be reached if your experiment performs extremely well, and gets ahead of the benchmark by 0 conversions immediately.


Expected maximum cost: $0.00

This will be reached if your experiment does not perform better than the benchmark, and the full 0 impressions are needed to be able to confidently declare a verdict.


This calculator was based on the excellent calculator by Evan Miller, adapted to use with a Landing Page experiment.

Please note that the outcomes presented by this calculator are intended only as a guideline, and no definite conclusions should be based only on these outcomes. The algorithm used gives only an approximation of real results, and discretion is advised in real-world situations.