Collection of Sufficent Data for Well-Informed Decisions

Rules should only be triggered after enough data is collected to make an informed decision.

To answer the question of when there is sufficent data, the minimum number of clicks needed for one conversion is a good pointer.

Minimum Number of Clicks for 1 Conversion

The following formulas can be used determine an ideal threshold:

              Medium Reliability (65% Probability of Sale):

Number of clicks = 100 ÷ Conversion Rate (%)
Example: 100 / 4% = wait 25 Clicks

             High Reliability (90% Probability of Sale):

Number of clicks = 100 ÷ Conversion Rate (%) x 2
Example: 100 / 4% x 2 = wait 50 Clicks

Having a conversion rate of 4%, for example, does not mean the product is guaranteed to generate a sale after 25 clicks. To avoid excluding keywords prematurely, it is recommended to wait around twice as long, ie. using the second of those formulas.

Organic or PPC Conversion Rate as Calculation Basis

In order to calculate “X clicks” you must use an expected conversion rate. For this you can either use the organic CR of your product/ product group, or even better the average PPC CR of your product/ product group.

The following table provides a breakdown of the number of clicks needed (for both medium risk and low risk thresholds), according to the respective conversion rate:

Automatic Calculation of X Clicks

If the threshold value is set to "1X" clicks, Sellics will automatically calculate the corresponding value for the respective adgroup:

The advantage here is that one rule-set can be used for multiple (differing) adgroups.

Calculation Method

The dynamic threshold (1X) will be automatically calculated on the basis of the PPC conversion rate within the past 90 days of the respective adgroup's products.

If there are multiple products within an adgroup, Sellics weighs their differing conversion rates based on the amount of orders they generated in that adgroup.

If the product has no conversion data in that timeframe, the system will make use of the average CR in the adgroup, campaign or all products in the same marketplace in that order.

Did this answer your question?