How does Pointzi's matching work?¶
Once the Pointzi SDK is integrated, it sends screen visits and clicks to the cloud. The SDK does not send personal data to the cloud unless you include it in your tagging. This is done by your developer either inside the SDK or from your backend applications.
- You create an experiment here.
- You create your filters to match your users
- You design your Tips and let save the experiment.
- Enable your experiment
What the matching engine does¶
Once enabled, the engine will continuously test for matches. The Pointzi platform will only match one experiment to one device - this ensures you are not spamming your users. Of course you can run multiple experiments at one time. The experiment will run CONTINUOUSLY until it expires.
If you make a change to a running experiment, it will ONLY affect the matches and installs that occur AFTER the change. (However you can't edit any experiment that is part of an A/B experiment because it will pollute the results).
You can re-target users with an updated experiment by using the "re-run" option. HOWEVER users who have previously matched may have seen your experiments content previously AND will see it again.
We encourage using "re-run" while you are creating your experiment and testing on just a few users.
We discourage using "re-run" on production users for the reasons above.
Please refer the video below.
Here is an example, we will consider iOS only just to keep it simple.
Scenario: Assume there is a total of 20 devices/installs (all iOS) that have matched and then 10 additional devices/installs after we created the experiment. Assuming we created our filters to match all of these installs (e.g "user's who just installed on iOS")
- The first 20 devices will match when the experiment is started. Matching is fairly real-time.
- For each of the 10 additional devices, they will be matched when they are created (or qualify for the match)
All Pointzi experiments (campaigns) will trigger real-time for the selected audience until it expired. See Timing tab on your experiment.
So you can have a new user qualify for a tip by selecting a target audience that is filtering with a Tag value/condition without having to re-run or schedule the experiment.