3 Movies That Weaken Machine Learning and How Adobe Analytics Can Turn It Around

[#1]

Movies are great storytellers and best in giving lesson which are applicable to performance marketing too. Here are titles that digital marketing analysts resonate with when it comes to performance analysis and optimisations.

3 Movies That Weaken Machine Learning and How Adobe Analytics Can Turn it Around

1. Wrong Turn

When digital marketers or media managers neglect the learning phase, there are multiple loopholes and wrong turns made along the way. Digital Analysts role is crucial at this stage most especially during the first seven days of campaign launch. The learning phase is not a myth, ads in its infancy stage are not as intelligent as its advanced stage. Hence, to avoid making unnecessary modifications that will lead ads to the wrong way, maximise Adobe Analytics.

Adobe Analytics Dimensions to Use to Turn it around:

• Website Entries- You can eliminate ads with high drop off rate (Drop Off Rate= 100%- Website Entry Rate). Find out more about the drop off rate calculation.
• Bounce Rate- You can pause ads with high website bounce rate.
• Countries Web Entries and Bounce Rate- You can discontinue ads in locations with high drop off and/or high bounce rate.

2. A Very Long Engagement

When ads have been live for too long or served with too high volume, it leads to ad fatigue. It is not the audiences who experience fatigue, but the ad serving machine when it is no longer reinforced enough. When using Adobe Analytics to turn around the effect of this very long engagement, use Week dimension.

Adobe Analytics Dimensions to Use to Turn it around:

• Download Rate- If your ads’ main call to action upon landing on the website are file downloads, this dimension will be expected to decline. In the example below, it is noticeable how the website download rate started to drop from Week 10 and continuously on a downtrend until Week 16 of the campaign:

The principle is the same depending on your ads main call-to-action on the website such as video view or contact form submission rate. In such scenarios, Adobe Analytics tell you that your ads require optimisation or creative refresh.

• Average Time on Site- If the target audiences were expected to read more, explore, or consider your offer on the landing page, the average time on site will be good indicator whether your ads are still bringing good quality audiences to your site. In the example below, it is obvious that from week 14 onwards, web users spent lesser time on your site. But you do not wait till week 16 to give recommendations to the media managers, a bi-weekly ads optimization is suitable.

3. Fast & Furious

When ads run for too short duration, or media managers are too quick to make campaign edits, both lead to insufficient machine learning. From the understanding of the learning phase principle, ads need to gather enough data and insights for smarter ad serving, else, the ads will keep on learning. In such cases, Adobe Analytics data will also be erratic

Adobe Analytics Dimensions to Use to Turn it around:

• Website Entry Rate- if your ads’ media buying objective is to bring users to the website, the entry rate will be erratic. In cases like below, you can raise the alarm and alert your social team or media managers that the campaign isn’t getting enough traction when it comes to website traffic.

• Download Rate- If your ads’ main call to action upon landing on the website are file downloads, Adobe Analytics data will show how this dimension had never stabilised.

The key is to allow the ads serving machine to stabilize and use Adobe Analytics to support your recommendations for the best campaign delivery.

If you have other movies in mind to add to the list, please feel free to comment below!