How to get top offer per channel in 8.4 NBAD solution
By default, the top offer selection happens across the offers and channels combination. For instance, if a customer is eligible for 2 offers across 2 diff channels, then only 1 channel gets selected for processing.
In the generated strategies using NBA designer, the strategy TopOfferOrBundleForOutbound picks up the top offer based on the priority across all offers/channels combination.
How do we achieve the requirements where the same offer should be communicated via more than 1 channels say Direct Mail and Email? There is no extension strategy for this requirement and I believe updating the OOTB strategy TopOfferOrBundleForOutbound is not recommended.
If it is ok to update this strategy and change the limit, will that have any impact on the action flow instance execution? With a single action flow across multiple channels, the way to implement the action flow is with a decision share and take required steps based on the channel. If the channel is email send Email shape, if DM use generic shape to write to DB or a file and do the post processing based on the outcomes on each of the channels.
It's ok to update the TopOfferOrBundleForOutbound strategy and change the limits per channel.
The impact on the action flow execution is that if the same action is emitted on multiple channels, a separate instance of the action flow is initiated for each channel, so you will need to implement the action flow with a decision shape and take required steps based on the channel. However, since each channel will have its own flow instance, beware of merging the logic from multiple channel treatments into a single stream within the action flow - while this logic will still work OK, each channel will execute separately, so there's no notion of merging the channel processing into a single stream.
Note that in 8.5, action limits can be configured separately for each channel, and for each action context level (e.g. Subscriber, Account, etc.), without modifying any strategies. However, no change to how action flow instances are initiated.