Limited Availability (LA) Releases indicate that the capability has passed the Preview Release phase, and we intend to take it to a GA status – we just aren’t there yet. It comes with the following caveats:
It fully solves for a base (minimal) number of valuable use cases. However, it is premature in some GA hardened capabilities (e.g. UI errors/warnings, usability, Devops/revision management scripts, implementation & upgrade guides, etc.).
If a client intends to use it in production, we ask that this be discussed with the Customer Decision Hub product team to ensure we can support the intended use case and upgrade path which may involve manual steps, special work, etc.
Presently, we do support analysis of action performance in production VBD, but our long term vision is to shift reporting to the Business Operations Environment (BOE) so customers should start planning for this. In fact, we are targeting some new operational action performance reporting in the BOE for v8.7.
Most large clients move their Interaction History (IH) data on a daily basis to their inhouse DB and they can retain this data based on their corporate strategy/policies.
Action and Treatment creation should be taken care of by the NBA Execution forum. Once related parties analyze and scrutinize each new Action request, they would be able to control the growth of Action amount. That’s the power and responsibility of the forum.
Such activities are taken care of by Pega Cloud and Global Customer Service teams in Pega cloud. On prem implementations can create a separate tablespace for IH schema and apply partitioning on the Fact table based on their corporate needs.
Yes, propensity can be calculated separately for every treatment. For example, if a customer is most likely to interact with a push notification, an action which has a push notification treatment is most suitable for that customer. The Customer Decision Hub is configured to calculate the propensity for each treatment. This NBA-Designer feature will automatically determine the best treatment for the customer. On top of that custom Adaptive models can always be built to measure channel lift.
By default, there isn’t any time to live value for IH Summary data sets and it’s not required. These data sets aggregate interaction history in real-time to limit and refine the data that strategies process. For example, you can create a data set that groups all offers that a customer accepted within the last 30 days and use that data set in your strategy. The rule itself, last 30 days, filters IH responses in run-time automatically. Thus, there isn’t any need to define a retention policy.
Conduct a performance test of the environment before going live and verify if number of threads, partitions, core and memory utilizations are as expected.
Run multiple test cases to find the most optimum values for your environment / hardware.
The data in Business Operations Environment (BOE) should not be masked since BOE, by its definition, is an environment in production zone and must contain real customer data to run accurate simulations.
Yes. Scenario Planner is configured to run in the Business Operations Environment (BOE), supported by data that is moved from production to the BOE via the Data Migration pipeline.