Build an Allocation Model

This tutorial walks through building a complete cost allocation in PlaidCloud. By the end you’ll have a working model that spreads a cost pool across a target dimension using driver data — the foundation of activity-based costing, IT chargeback, and shared-service distribution.

Takes about an hour. Allocations are PlaidCloud’s most distinctive feature; this is the best way to understand the model.

What You’ll Build

Spread total IT department cost across business units, weighted by each unit’s user count.

IT cost table  →  ┐
                  ├→  Allocation step  →  results table
User counts    →  ┤      (driver-based)
                  ┘
                  ↑
        Business unit dimension

Prerequisites

• A PlaidCloud workspace (start a free trial if you don’t have one)
• A project containing or able to import:
  • Values to allocate — IT cost by month, total or by sub-category
  • Driver data — user count by business unit
  • A dimension — hierarchy of business units (units → divisions → company)
• Familiarity with Concepts — workspace, project, workflow, dimension, allocation

Note

If you don’t have the source data yet, you can prep small CSV files and follow Load, Transform, and Publish Data to import them first.

Step 1: Set up the Inputs

Confirm you have these three things in your project:

Object	What it holds
it_cost table	Total IT cost (e.g., one row per month with the amount)
users_by_bu table	One row per business unit with a user_count column
business_units dimension	Hierarchy of business units, with leaf nodes matching the unit names in users_by_bu

If anything’s missing, load it via a workflow before continuing.

Step 2: Create the Allocation Workflow

1. Open the project and switch to the Workflows tab.
2. Click New Workflow. Name it “IT Cost Allocation”.
3. Open the new workflow in the Workflow Explorer.

Step 3: Add the Allocation Step

1. Add a new step from the Allocation category. Choose Allocation Rules — the most flexible option.
2. Configure the source:
   • Source table: it_cost
   • Source amount column: the column with the dollar amount
3. Configure the driver:
   • Driver table: users_by_bu
   • Driver value column: user_count
   • Driver match column: the column with the business unit name
4. Configure the target:
   • Target dimension: business_units
   • Target level: choose which level of the hierarchy to allocate to (usually the leaf level)
5. Configure the output:
   • Result table: it_cost_allocated

Step 4: Run and Inspect

1. Run the step.
2. Open the Tables tab and click into it_cost_allocated.
3. Each row represents one slice of cost going to one business unit. Columns include:
   • The source amount (e.g., total IT cost for the month)
   • The target business unit
   • The driver value used (that unit’s user count)
   • The allocation rate (user count ÷ total user count across all units)
   • The allocated amount (source × rate)

Step 5: Verify the Numbers

Three checks every allocation should pass:

1. Total reconciliation — sum of allocated amounts equals the source total (within rounding tolerance)
2. Rate sum — allocation rates sum to 1.0 (= 100%) per source row
3. Coverage — every business unit in your dimension that should receive a slice actually got one

The Allocation results guide has a full checklist for verification.

If something’s off, the most common issues are:

• Missing driver data — a business unit in the dimension with no row in users_by_bu won’t receive an allocation
• Mismatched names — driver table says “Sales East” but dimension says “Sales-East” (different spacing/casing won’t match)
• Zero or negative drivers — produce zero or unexpected allocations

Step 6: Use the Results

The result table is just like any other PlaidCloud table — it can be:

• Joined with other tables in further workflow steps (e.g., add a fully-loaded cost column to the GL)
• Published for dashboards or external consumers
• Re-allocated as a source for the next round of allocations (for cascading models)

Variations to Try

Once the basic model works, common extensions:

• Multiple cost pools — replace the single it_cost row with one row per IT sub-category (compute, storage, licensing) and allocate each independently with different drivers
• Multi-period — partition the source by month and produce one allocation result per period
• Layered allocations — allocate divisional overhead to units, then unit costs to products, then product costs to customers. Each layer is its own allocation step
• Recursive allocations — when shared services consume each other (IT serves HR, HR serves IT). See Recursive allocations

What’s Next

• Allocations guide — every option and configuration choice
• Rule-based tagging — different allocation rules per source row
• Allocation step reference — every workflow step in the Allocation category
• Dimensions guide — building the hierarchies that allocations target