This A/B testing might hurt your performance

One of the main keys to achieve better performance is to constant A/B testing. You are advised to test literally everything: word order, adjectives, full keyword vs. partial keyword, call to action, punctuation. While there’s no excuse for not A/B testing your PPC campaigns, some tests might actually hurt your performance. How is it possible?

Let’s assume that you test two exactly the same ad copies with different call to action: “Visit our store” or “Buy now”. The problem is you test the same concept – you encourage people to take the desired action. Some might argue that the used words make difference but – according to Jacob Nielsen studies – people don’t read, they scan. At the first glance, these ads copies don’t provide different values to the users. The results are very likely to be similar. But let’s assume that “Buy now” was the winner. What did you learn: that “Buy now” is the best call-to-action ever or it was better performing in this particular context? Should you always use the winning call-to-action it in your ads? Not really, since it might have performed better only with the ad copy you used. If you use different unique selling proposition, “Visit our store” might bring better results. That’s why this small change won’t really help you to achieve the long term improvement of your campaigns.

When you focus on minor changes of you ad copy, you are probably missing some opportunities. In the both ads you are testing the same unique selling propositions so you are targeting the same audience. You get clicks from the people who responded to your message. But with just one message, you don’t reach people for whom your ad doesn’t present enough value to click on it. If you test only two ads: in the first one you include minimum price 50€, in the second one price range from 50 to 100€, you receive clicks only from the price-responsive users, but what about those who focus on quality, fast delivery or great choice?

Time for a conclusion. Don’t waste your time. Instead of focusing on minor, not meaningful changes, try to test  completely different variables: price, great choice, discount, seasonal offer, exclusivity, free registration, quick delivery, social proof etc. This way you learn what your customers respond to and that’ll help you to improve long-term performance of your account.