Select one or more mobile apps. For example, an app on both iOS and Android platforms.
Click Continue.
From now on, the recommendation will be saved as a Draft in the list.
Currently, the recommendations are only supported for apps placed in the App Store and Google Play stores
2. Basic parameters
Specify the basic parameters:
Enter a recommendation name by which you can distinguish it from the general list.
Select a key metric to evaluate the recommendation results:
Revenue from offers — in the summary reports,
you will see how much more revenue from recommended offers than from regular offers.
Total revenue — n the final reports you will see how much more revenue from users who selected the recommended offers than from regular users.
Total revenue includes not only payments for accepted offers, but all subsequent user payments.
Specify one or more placements in the app where the recommended offers will be displayed.
These are the names of the locations you use in your application code.
For example, top_banner, player_profile.
You can get placement_id from your app developer.
This parameter should be passed to the MyTracker SDK along with the offer_triger,
purchase_offer, offer_card events.
See Connecting MyTracker Personalize.
Click Continue.
3. Segment
Select the user segment for which the recommendation will run.
For example, users who have reached the fifth level or users with less than 1 dollar in the last 30 days.
On the fourth step, divide the selected segment into control and recommendation groups:
Control group — users who will be shown default offers.
This group is needed to run the A/B test and evaluate the results of the recommendation.
Recommendation group — users who will be shown the recommendation offers.
The size of the selected segment and the split settings affect the time of the A/B test,
that is, the time to get the results of the recommendation.
The smaller the recommendation group, the longer the A/B test will take.
After you run the recommendation and get a significant result, you can change the split again.
5. Object
In step five, select an object to offer to the control and recommendation groups:
Item — if there is a single item in your offer.
For example, a t-shirt in a clothing store, a golden sword in a game.
Set — if your offer includes a set of items to choose from.
For example, a t-shirt, shoes or watch in a clothing store,
or in-game currency of different denominations of 100, 200 or 300 diamonds.
6. Offers
Add offers that will be shown to users:
one offer for a control group to check the recommendation result
and several recommended offers, from which Personalize will select the best one for each user.
Click Add.
For the control and recommendation groups, the addition forms are the same.
Fill out the form:
Payload — a unique offer identifier within the application.
The payload can contain letters, numbers, and special characters,
depending on the logic of your application.
For example, the payload for the offer to buy 2 t-shirts on promotion
is com.app.2tshirts_discount.
SKU — an identifier of the offer price in the app store.
The SKU is not unique and can be used for several offers at once.
It can contain letters, numbers, hyphens, and special characters,
depending on the logic of your application.
For example, the SKU for the $5 offer is com.app.5dollars.
You can get Payload and SKU from your app developer.
These parameters should be passed to the MyTracker SDK along with the code>offer_triger,
purchase_offer, offer_card events.
See Connecting MyTracker Personalize.
Old price — the offer price without discount.
In the interface, you can display the old price as a crossed-out value or not use it.
Discount price — the discount price.
This is the price at which the user buys the offered number of items.
For example, $5 instead of $10.
Old value— the quantity of items without discount.
In the interface, you can display the old value as a crossed-out quantity or not use it.
Discount value — The quantity of discount items.
This is the quantity that the user buys.
For example, buy two t-shirts instead of one.
If you selected Set as the recommended object,
you can add more rows with items using the Add button
Click Continue.
The control group offer will be labeled A,
and the offers for the recommendation group will be labeled B1, B2, and so on.
Repeat steps 1-3.
Make sure you add one sentence for the control group and several sentences for the recommendation group.
Click Test connection.
You will be redirected to the testing form.
To create several similar recommendations, use duplication.
To do this, you need to have the rights to edit recommendations or be
the owner or manager of the account.
You can fully duplicate a recommendation with all apps, or duplicate a recommendation only for one app.
As a result, a step-by-step wizard with pre-filled fields will open. The duplicate will be created as a Draft.
For all apps
Select a recommendation from the list and click .
You can duplicate a recommendation in any status, except Draft.
The duplicate form will be automatically filled except for the Placements field.
Specify the placements in the app where the recommended offers will be displayed, in the next steps of a recommendation duplication.
On the recommendation view page, click Duplicate recommendation.
A duplicate can be created for any status, except Draft or Testing.
The duplicate form will be automatically filled except for the fields Placements
and User segment.
Specify the placements in the app where the recommended offers will be displayed, and choose the user segment, in the next steps of a recommendation duplication.
