- Integration in just 5 minutes (No changes will be made to the site during this process)
We value how important it is for you to have all of the data you need in order to validate our methodology and results. Once the integration is complete, you will be empowered to verify the data we report using your own independent measurements
Here are three steps to integrate:
-
-
- You can start by going to: datamilk.app/signin
- Register your email and domain
- The system will generate your 2 lines of code (snippet and pixel) with your own Domain ID and instructions for you to complete the integration.
-
Example of snippet and pixel for reference:
<!-- DataMilk script -->
<script async src="https://datamilk.app/magic_ai.js?id=YOUR_DOMAIN_ID"></script>
<!-- end script -->
<!-- DataMilk pixel -->
<script src="https://datamilk.app/magic_pixel.js?id=YOUR_DOMAIN_ID>&amount=REPLACE_AMOUNT&orderid=REPLACE_ORDER_ID¤cy=REPLACE_CURRENCY&gp={{ gross_profit }}"></script>
<!-- end pixel -->
- AI model training (Data collection phase; no changes will be made to the site during this process)
-
-
Once you have integrated DataMilk onto your production site we begin the data collection phase, analyzing user behavior and training DataMilk AI models.
- The traffic ratio is split 50% control group (original) and 50% treatment group (Smart) during this phase with no Smart Components running.
-
The minimum requirement before we can launch Smart Components (changes to the site) is at least 2 weeks and 500,000 sessions.
- These requirements typically provide adequate time to collect enough data to train the DataMilk AI models and conduct an A/A test (control group vs control group) to achieve the expected result of neutral uplift.
-
Once you have integrated DataMilk onto your production site we begin the data collection phase, analyzing user behavior and training DataMilk AI models.
- Independent Verification Plan (Choose what method to use to independently verify DataMilk results for your internal team)
Option 1: Google Analytics- Use DataMilk's tagged sessions to see A/B results in your GA out of the box. For more information on how this works, please access this Help Center section: Independently verify DataMilk results with Google Analytics.
Note: If you prefer to use another analytics tool, please notify your Customer Success manager which one you prefer and we will happily get it set up for you.
Option 2: Control the A/B test group selection yourself - notify your Customer Success manager at DataMilk and use this article as reference.
Let us know what you have in mind, we love new ideas to help our customers have confidence in the results.
- Launch (Approval phase; start or pause smart components on your site)
-
- Once the models are trained, the Smart Components are queued in your DataMilk dashboard as, ‘Coming Soon’.
- You can preview Smart Components before they are launched, share feedback and request changes.
- You have control to launch or block Smart Components to your site.