I haven't updated an article in a long time, so let's read facebook officially todayDefine machine learning in your ads

I find it particularly interesting, as long asIf you follow these logics down-to-earth, you can do a good job of Facebook ads

============================

Let's read:

Focus 1, Facebook ads are the use of machine learning to make the most likely action (such as your goal is to buy, that is, the most likely to buy your things; for example, your goal is to add a shopping cart, that is, the most likely to add a shopping cart, rather than the person who buys, so the target setting of the ad is very important) to see your ad;

Focus 2, machine learning method in the case of obtaining sample data enough, you can ensure that your ads are very stable out of orders, roi is also very good, positive and negative feedback are maintained in the case, do not move it, a day 1000 dollars of advertising, roi is 2.5, an image ad can run for a month;

=======================

3. The focus here is on what we're going to do if we want to take advantage of Facebook's machine algorithms for a long time:

A, in the learning stage do not - unchanged should change (but if the roi is rotten, or off);

B, the time of the learning stage

Within the conversion window - for example, if you choose to see results within 1 day, you must see results within 1 day

The number of times the results are 50 times, this machine learning method can guide the ads to find the best audience;

Within the conversion window - for example, if you choose to see results within 7 days, you must see results within 10 days

The number of times the results are 50 times, this machine learning method can guide the ads to find the best audience;

======================

4. Conditions for the end and start of machine learning

Ad a set up $500 per day, 7 days conversion window, the current two-day purchase results occurred 80 times,

A, the ad does not move, then Facebook may feel that your audience it knows, will be in accordance with your daily budget of $500 to try to put on the same characteristics as the previous 80 single people;

B, advertising plus budget, from 500 to 1000, this time, Facebook received instructions is not 500 plans per day, will be based on 1000, the start of new machine learning methods;

The second point in the screenshot below actually means that your ad can't get anyone else

Facebook recommends that you step back - for other goals, such as dropping from buying to adding a shopping cart

My advice is to ----- off the ads

====================

In general, it's the Facebook ad machine algorithm

The basic logic of the algorithm (the audience most likely to do an action)

Conditions for starting over (any little modification)

Conditions for a stable end (within the conversion window).50 results）

The Facebook ad tutorial can be seen at the link below

https://www.facebook.com/business/help/

Recently in the finishing of ideas, next time and share with you about Amazon and Facebook promotion, my own experience ha!