運用型広告(Standard Ads)|アルゴリズム

Use SmartNews AI to optimize your ads

SmartNews Ads utilizes three main types of data: user behavior, viewed content, and ad interaction history to analyze and use that information for effective ad delivery.

Our AI uses this data to maximize conversion and reach by predicting which users are most likely to click on your ad.

The number of conversions required for sufficient optimization is 10 per creative.

When submitting a new ad creative, your ad will be given a priority auction opportunity until a certain amount of results are accumulated.

In addition, as optimization progresses, creatives that are predicted to perform well will be enhanced for delivery, and creatives that are predicted to perform poorly will be restricted.

Therefore, you can expect performance improvement even if you don't frequently turn creative ON/OFF.

Concrete behavioral pattern and proven record of improving ad performance

Successful creatives see a significant increase in the number of impressions.

Recommended settings

Turn on as many creatives as possible (5 or more recommended per CPN)

  • When effective creatives lose steam, auction opportunities will increase for creatives that are predicted to be the next most effective and have not received many impressions. Therefore, we recommend that you do not set creatives to "off," even if there are creatives that are not getting many impressions on the first try.
  • For SKAdNetwork campaigns, please turn set creatives to "on" as much as possible in the allocated slots.
  • There is no limit to the number of creatives that can be set to "on."

Try not to turn set a creative to "off" until you have at least 10 conversions (CV)

  • vCTR prediction is fairly accurate at 20 clicks, and for CVR (conversion rate) prediction, learning is maximized at 10 conversions.
  • If a creative with a very high vCTR and an extremely low CVR is getting a lot of impressions, we recommend manually turning it off.
Was this article helpful?
1 out of 1 said it was helpful

Contact Customer Support

Sent!
thank you

Thank you for your feedback.
We will try to improve the article.

What kind of problems did you have?

「その他」の場合はコメントにて詳細をご記入ください

Contact Customer Support

はじめてガイド(概要)

×