Guide · Architecture

Anaplan DISCO explained: the model architecture that keeps models fast

By Umang Chauhan · November 14, 2026 · 7 min read

DISCO splits a model into Data, Inputs, System, Calculations and Outputs. What each layer holds, how data should flow between them and how to spot a model that has broken the pattern.

DISCO

Ask ten Anaplan builders how to structure a model and you will hear a lot about DISCO. It is the simplest rule of thumb in the Planual tradition: give every module a single job, and let data flow one way through five layers. Models that follow it are easier to hand over, easier to test and usually faster.

The five layers

DISCO: one direction, five jobs

  1. Data

    Raw and cleansed source data, often in a data hub

  2. Inputs

    What planners type in or approve

  3. System

    Mappings, time, switches and settings

  4. Calculations

    The logic, in small purposeful modules

  5. Outputs

    Reports and pages people read

Data moves left to right. Nothing in a later layer should feed back into an earlier one.
  • Data holds imported data in the shape the source gave it. It is not edited by planners. See Anaplan data hub for when this layer should be its own model.
  • Inputs are for assumptions and decisions entered by users. Keep them apart from calculations so you can see, and protect, what people are allowed to change.
  • System modules hold mappings, switches, time settings and other plumbing. They are small, stable and shared.
  • Calculations is where the logic lives. Each module should do one thing and be named for it.
  • Outputs shape results for readers and App pages. They read from calculations and never calculate much themselves.

Why it helps

Faster reviews and safer changes

When inputs, logic and reports are mixed, a change to one needs you to understand all three. With DISCO you can answer “what can a planner edit?” by opening one set of modules, and “what feeds this report?” by following arrows in one direction.

Less waste

Output modules are shaped for people, often with many columns and pages, while calculations are best at the lowest grain they need. Keeping them separate stops report layout decisions from inflating the cells your logic carries. The arithmetic is in Anaplan sparsity and dimensionality.

How to spot a model that has drifted

  1. Open the largest modules by line item count. A module holding inputs, logic and report columns together is the usual offender.
  2. Look for line items in an output module that feed other modules. That is logic hiding in a report.
  3. Check whether the same mapping is rebuilt in several calculation modules instead of living once in System.
  4. Ask who may edit each module. If the answer is “it depends”, inputs and calculations are mixed.

Anaplan Toolkit's Modules report lists each module with its dimensions and line item count, and Impact shows which line items feed which, so the review above takes minutes instead of an afternoon in Blueprint.

Do this audit in minutes, not days

Anaplan Toolkit is a free Chrome and Firefox extension that reads the structure of the model you have open: modules, line items, lists, formulas, usage and storage. It talks only to Anaplan and sends nothing anywhere else.

Get Anaplan Toolkit

More guides

All Anaplan guides →