Marketing A/B Testing
When developing marketing campaigns, you should always A/B Test, which is referred to as creating multiple campaigns for the same goal. By creating multiple versions of a controlled campaign, it could help your business determine the most effective campaign based on a side-by-side comparison.
Example.
In a study conducted by the Coca-Cola Company, taste was considered the most important factor for their declined sales between the 1970s and 1980s. In response, the New Coke was developed to enhance the flavor and generate sales. With the approval of more than half of the 200,000 blind product testers, the New Coke was introduced to replace the original flavored Coke— which was a disaster for the company. Because of the unfavorable results and backlash from customers, the New Coke was withdrawn from the market and the Coke Classic was introduced with the original ingredients. Regarding the Coca-Cola Company’s quantitative study and the New Coke, the test provided inconclusive results because of the lack of information given to the participants. Though the participants enjoyed the flavor of the New Coke, it did not outweigh their desire and loyalty of the original Coke. The New Coke’s example of relying on only one type of study displayed poor predictive validity. A customer’s like toward a product is not a good predictor of whether they would purchase the product over another product they also like. Had the Coke Company taken a segmented approach and kept both products on the store shelves, they could have had a better idea of which the users liked best and which would generate the most sales in a side-by-side or A/B comparison.
Most businesses who do A/B testing often fail to give their testing ad a chance at achieving its best possible results. Furthermore, without testing the ad properly, these individuals or companies may ultimately lose money, time, and potentially reject a successful ad that just never had the chance to prove itself successful due to the lack of owner knowledge. Having a control ad running throughout the testing of a new ad can save money and eliminate performance history as a variable in the test ad. By creating multiple copies of a control ad, you can avoid giving 50% of impressions to an unproven ad which may fail. Creating multiple copies of a control ad allows you to save on cost considering the control ad is always running. Instead of losing 50% of ad impressions to the testing ad, that percentage shrinks to 20% because of the additional control ad copies. In addition to this type of testing saving on cost and potentially accelerating the testing process, it also allows for you to create a valid testing environment by comparing the performance of the test ad against the copies of the control only.