Your business logic, packaged as an intelligent planner.
Custom Planners turns reviewed business logic into a reusable decision workflow. We design, build, test and benchmark planners for decisions your team currently handles by hand, including eligibility, timing, scoring, routing, suppression and promotion rules that do not fit a standard automation.
The starting point is the current operating decision, not a request to automate everything. A useful planner has defined inputs, an eligible audience, explainable reasons and an owner who can stop it. Its effect is compared with your own workflow rather than a generic promise about time saved or performance lift.
Updated
Who this is for
This service is for teams with a differentiated decision that generic campaign rules cannot express reliably. It fits when operators repeatedly combine several data sources, review exceptions by hand or maintain different copies of the same rule across tools. The business logic should be clear enough to discuss and important enough to measure.
If a standard Intelligent Planner covers the requirement, begin there. If the data and execution interfaces are the missing piece, Implementation Services may be needed before custom logic can be evaluated. The discovery discussion should separate a rule problem from an integration or measurement problem.
What you get
| Deliverable | What it contains | Purpose |
|---|---|---|
| Planner specification | Inputs, decisions, reasons and ownership | Make the intended behavior reviewable |
| Eligibility and stop rules | Permission boundaries, exclusions, limits and pause conditions | Prevent optimization past an agreed control |
| Baseline and measurement plan | Starting workflow, fixed cohort, metric sources and windows | Compare the pilot consistently |
| Deployment on the existing ESP | A scoped decision-to-execution integration and validation plan | Keep current channel tools in the workflow |
| Monthly review | Decision reasons, observed outcomes and proposed rule changes | Maintain oversight as conditions change |
Examples of custom planner logic
A high-LTV reactivation planner can focus a reviewed cohort; a VIP frequency planner can coordinate contact pressure. Lead-quality routing, content-interest selection, partner suppression and seasonal promotion are other examples already described in the service. Each still needs a precise definition of eligibility and evidence for its inputs.
For example, a content-interest rule can help choose a relevant topic, but it must not override an unsubscribe. A high-value score can guide selection only after consent and suppression checks. A partner exclusion needs an agreed source and scope so the same person is not reintroduced by another campaign or integration.
Build when generic rules are not enough
Use custom logic for a decision that is specific to the operating model. Write down the current manual steps, the exceptions and the reasons an operator would decline to act. This often reveals a smaller reliable change than replacing an entire workflow. The planner should package those reviewed decisions rather than hide them behind a score nobody can explain.
How it works
- Define logic: identify the decision, owner, constraints and expected effect. Include examples of eligible actions, exclusions and situations that require human review.
- Map data: connect the required profile, event and permission inputs. Live Marketing Status and Event-Centric Timeline provide current state and sequence where needed.
- Build the planner: implement the reviewed rule and its reasons. Connect the result through the Decision API or the scoped execution interface on the existing ESP.
- Run a pilot: use an agreed cohort and validate the action and suppression paths. Production activation follows review of the evidence, scope and rollback plan.
- Compare and optimize: evaluate outcomes against the starting workflow, inspect adverse signals and update rules under oversight. Expand only when the pilot is dependable.
Benchmark against the current workflow
Freeze the baseline before comparing planner-assisted execution. Specify the starting audience, denominator, observation period and source for each measure. Separate reduced operator effort from improved customer response: a faster decision can still be wrong, and a later purchase is not automatically incremental revenue.
Inspect representative decisions as well as aggregate results. The operating owner should understand why the planner selected a person, why another was excluded and which input would change the result. A green integration check alone does not establish that the business rule is correct.
Maintain and optimize with oversight
Review reasons and performance with the team, maintain guardrails and keep source definitions current. Changes to consent, event meaning or channel controls can invalidate an earlier assumption. Monthly review gives those changes a place to be assessed; an urgent suppression or stop condition should be handled when it occurs rather than waiting for the calendar.
What it costs
Scoped after discovery. Cost depends on the complexity of the decision, required data, integration work and operating responsibilities. The scope should distinguish the planner specification and build from ongoing review or related implementation services. There is no public fixed price or guaranteed performance improvement for a custom planner.
What we don't do
We do not automate unclear permission, bypass suppression to achieve a volume target or promise a standard lift. We do not replace human strategy or approve unreviewable logic simply because it is model-generated. Missing inputs, unexplained decisions and failed checks need an agreed safe response before release. The baseline and rollback path are part of the work.
FAQ
When should we choose a custom planner?
Choose custom logic when a meaningful business-specific decision cannot be expressed reliably by an existing planner, and when its inputs, exceptions and effect can be reviewed and measured.
Can it run on our existing ESP?
The service scopes deployment around the existing channel tool. Available interfaces and controls are checked during integration rather than assuming every provider supports the same execution path.
What if the required data is missing?
Identify the missing source or definition during discovery and mapping. Obtain the evidence or simplify the rule before treating the planner as ready for a pilot.
How do you benchmark performance?
Compare against the current workflow using the agreed baseline, fixed cohort, observation window and sources, retaining errors and adverse signals alongside positive outcomes.
Who controls later changes?
The operating team reviews reasons, results and proposed rule changes. Monthly review supports maintenance while immediate stop conditions and suppression updates remain operational responsibilities.
Read before you buy
Read playbook Step 6: cohort release, Step 7: stop conditions and the re-engagement examples. They show how a concrete workflow combines action, permission and review before scaling.