Every automation catalogued, every property owned, every rollup reconciled.
The work most Planhat instances actually need. We audit the automation estate, write the data dictionary, consolidate what should never have been separate, and build the hierarchy rollups so the number leadership sees is the number the account team can reproduce.
Estates grow by accretion. Nobody designed them.
Eighteen months of different people solving different problems with new automations and new properties, none of it written down. The instance works until it does not, and nobody dares touch it.
Thirty-six automations, twenty unique intents
Near-duplicates written by different owners fire on overlapping triggers. Some update the same property in different directions. Nobody can say what breaks if one is turned off.
Properties with no owner and no source
Fields populated by an integration, an automation and a human, in that order, with no rule about which wins. The value you see depends on what ran last.
Parent numbers that do not match child numbers
Usage and revenue at the target company level are computed differently from the organisation and project levels beneath them. Every leadership meeting starts with reconciliation.
Calculated metrics that drifted from their definition
The metric is still named what it was named a year ago. The formula has changed three times and the documentation has not.
A warehouse feed nobody shaped
BigQuery or Snowflake pushes tables into Planhat with lag nobody accounted for, so week-over-week movement is sometimes measuring the sync, not the customer.
Audit, dictionary, consolidate, roll up.
Modelled on the estate work we delivered for a construction-tech SaaS client with a three-level hierarchy and thirty-six active automations, and refined since.
Automation audit
Week 1Every active automation catalogued: trigger, action, properties touched, hierarchy level and intent. Duplicates, conflicts and orphans identified. This alone usually explains three or four mysteries the team has lived with for months.
Data dictionary
Weeks 1 to 2Every property on every level of the hierarchy, with ownership, source system and propagation direction. The document that finally answers which system is the source of truth for each field.
Consolidation with rollback
Weeks 2 to 3High-impact consolidations rebuilt and tested in production, each with a written rollback plan communicated before deployment. Nothing outside the agreed list is touched without written approval.
Rollups and cascades
Weeks 3 to 4Usage, revenue and targets rolled up from the lowest level to the top, with gap to target and week-over-week movement computed consistently at every level. Cascades for properties that flow down. Sync lag handled explicitly.
Estate map and handover
Week 4 to 5A living reference your team can maintain: the automation catalogue, the dictionary, the rollup logic and the change-control agreement. Walked through on a call, handed over in your tools.
Deliverables you can point to.
Every item is handed over in your tools, walked through on a call, and accepted against written criteria before the final invoice.
Every active automation with trigger, action, properties, level and intent, plus a keep, consolidate or retire decision on each.
Ownership, source system and propagation direction for every property at every level.
Rebuilt workflows, each with a written rollback plan and a before-and-after comparison.
Usage, revenue and targets from project to organisation to parent, with gap and week-over-week movement at each level.
The working agreement your team keeps using after we leave: written rollback, communicated before deployment, approval outside scope.
A single living reference in Notion or your preferred tool, kept current on retainer.
This is the right engagement if
- You have inherited an instance nobody documented
- Numbers at the parent level do not match the sum of the children
- You are about to hire a RevOps lead and want them to inherit a map, not a mystery
- You want a foundation before building health scores or dashboards on top
Look elsewhere if
- You want new automations added without understanding the existing ones
- There is no appetite for rollback discipline in production
- You are looking for hourly Planhat admin support
Specifics.
How do you consolidate safely without a sandbox?
Each consolidation is designed on paper, reviewed with your owner, then deployed with a written rollback plan that has been communicated in advance. We monitor the affected properties through a full cycle before calling it done.
Can rollups handle more than three levels?
Yes. Three levels is the most common enterprise shape, but the rollup logic is designed per level and extends. What matters is that gap-to-target and movement are computed the same way at every level.
What about integrations pushing data from a warehouse?
We account for sync lag explicitly so movement metrics measure the customer rather than the pipeline, and we document the failure modes so your team recognises them when they recur.
Is this a prerequisite for health scoring?
Almost always. A score built on properties with unclear ownership inherits every one of those problems. The audit and dictionary are the foundation the score sits on.
Often paired with
Health scoring
A model built from your churn history and validated on a holdout set.
Read more Revenue visibilityTeam cockpits and NRR reporting
Gap to target and week-over-week movement at every level, plus prioritisation.
Read more After the buildStewardship retainer
A monthly block of senior hours that keeps the estate and its documentation true.
Read moreSend us the automation count. We will tell you what it usually means.
Thirty minutes on the estate as it stands today. If an audit makes sense, the fixed-fee proposal follows within two business days.