Should I pause this ad, or wait for more data? This is a question that I’ve asked myself on countless occasions when running ad campaigns. Beyond the rules of thumb, are their statistically grounded methods to make this decision? This tool uses the Sequential Probability Ratio Test (SPRT) to give you an answer. It tells you whether an ad’s under or over performance is backed by sufficient statistical evidence, and whether you should pause or scale it.
What is a SPRT?
In simple terms, it’s a statistical method for decision-making that determines whether you have enough data to draw a reliable conclusion. It’s commonly used in fields like quality control and clinical trials to reduce the number of samples needed while still making accurate decisions. Here, the same approach is applied to advertising data so you can determine whether you have enough information to confidently decide if an ad should be paused.
How to use:
Click the Settings tab and enter the following info: (a) your target conversion rate; (b) the conversion rate that you would consider underperformance; (c) your average cost per click; and (d) your average conversion value. You can also adjust your error tolerance levels to set the risk of making an incorrect decision. Increasing the acceptable risk of a false positive or false negative means less data is required to reach a conclusion.
Following this, return to the Batch Calculator tab and enter the clicks and conversions for the ad in question. The tool will return one of three outcomes: (a) you don’t have enough data, (b) stop the ad, or (c) scale the ad.
