A 5 minute quick guide to how the Aampe Integration process works
How do we integrate with Aampe?
You will need to do a 2 step process
- Push Integration
- Data Integration
What is Aampe Push Integration?
Aampe sits on top of your existing communication stack. We don't replace it. Aampe will access credentials to your existing push provider or communication gateway and trigger push notifications by using those credentials.
What is Aampe Data Integration?
Aampe data integration enables us to access your first party events data which is fed to our models. Aampe stores your events data in the geographical GCP region you prefer and does not store any PII data
Do I need to share my entire events table with Aampe?
It is not a necessity, but yes you can. Our algorithms preprocess, cleanup and cluster your event data. We will work with you to only ingest the events that are necessary for our learning and not store any extra information in our BigQuery tables.
However if you prefer to share only certain events, hop on to the next question
Which events from our data stream should we share with Aampe?
Events data can be broadly grouped into the following categories in order of their importance for our models
- Conversion (checkout started, purchase completed, etc.) - Bottom of the funnel events
- Messaging (notification clicked, notification received etc.) - Clickstream
- Intent (product page viewed, item added to cart, item added to wishlist, etc.)
- Discovery (search query entered, product list filtered, product category selected, etc.)
We just require Bottom of the funnel events and Clickstream from you to get started with Aampe.
However the more events you can provide, the more detailed user profiles Aampe will be able to build. More detailed user profiles lead to better learning and higher messaging success rates.
What information do you need from events data?
We at the minimum require these data points related to the events
- Unique user id
- Event name
- Timestamp of event
- Unique ID for the event - Optional
Apart from this we suggest any metadata related to the events is shared with us. For example:
For add to cart events, it's useful to add the item reference (SKU), price, currency, and discount
For start checkout and order completed events, it's useful to add the list of items, total price, currency, and discount
For page view events, it's useful to include the page name, brand name, game name, or restaurant name where applicable
Do we need to share anything else?
We will also need to ingest attributes about all users.
- Unique user id
- Notification token for the user
- Any other metadata (like country, timezone etc.) - Optional
This data can be combined with the aforesaid events stream or sent as a separate stream, whichever is easiest for you. We will do all the processing to fit into our data models.
You can reference Data Models for more details
Do I need to follow the exact data model?
Not at all, as long as the basic necessary fields (as outlined in above answers) are included in the data stream, we will do all the necessary pre-processing to fit into our data model.
Now you can go back to the Integration for detailed steps
Still not clear with how the integration should work?
No worries, please drop us a note at [email protected] . We will reach to help clarify any doubts
Updated about 1 month ago