As with "site name," the keywords "storage" and "furniture" were also evaluated for their advertising contribution. As a result, the ad management screen determined that "furniture" was not linked to conversions and therefore should be suspended. However, after the evaluation, it was determined that far from being suspended, it actually contributed more to advertising than "site name . "
An example of how attribution analysis was conducted to verify the accuracy of a hypothesis
Now, .
When we first started supporting Company A, we phone number database calculated its marginal ROAS and CPA, and then re-measured the effectiveness of its advertising, which revealed that the ROAS and CPA via search-based advertising had deteriorated significantly.
So, for example, when they looked at the advertising effectiveness of non-brand words such as "furniture" and "mail order" and brand words such as "site name" and "product name", they found that the brand word campaign on the ad management screen had 457 conversions with a CPA of 433 yen, while the non-brand word campaign had 81 conversions with a CPA of 15,869 yen.Company A's marginal CPA was calculated to be 12,000 yen, so simply judging from the numbers on the ad management screen, they would have to reduce advertising for non-brand words.
However, there was a hypothesis that users who came to our EC site from non-brand words would eventually come from brand words and make a purchase, so the result was profitable , and this was a topic of discussion within the company. Therefore, to find out the truth of this, we used attribution analysis for three months to reevaluate CV to TCV and CPA to TCPA using an equal evaluation model. As a result, the TCV and TCPA were as follows.
2016_nettan3_4These results revealed that there was no significant difference between brand and non-brand word conversions and TCV, or CPA and TCPA. This means that there are few cases where users who come into contact with non-brand words end up purchasing with brand words. Because Company A's marginal CPA was 12,000 yen, they boldly reduced the proportion of non-brand word advertising, which resulted in a significant improvement in ROAS.
In fact, there are many cases where companies have been able to improve their strategies by evaluating the contribution of their products to brand words, as Company A did, so we highly recommend that you verify whether your hypothesis is correct.