New: the 540 Learning Theory — the thinking behind how we build teams and products. Read it →
54Systems
CO
FinanceShipping now

Cost Distribution

Allocate every riyal to the right cost object — auditable, IFRS-aligned, automated.

Platform
ERPNext / Frappe v15 (kayes_cost_distribution)
Built for
Finance · Shared-services · Multi-entity
Compliance
IFRS · Saudi regulatory
The problem

What Cost Distribution removes

Allocating shared costs fairly — and proving it in an audit — defeats most spreadsheets. Cost Distribution allocates every riyal to the right cost object.

Capabilities

What it does

1

Multiple methods

Payroll / effort-based and shared-pool cost distribution, applied consistently.

2

Audit-ready

Every allocation carries its basis, aligned to IFRS and Saudi regulatory requirements.

3

Automated

Runs on a schedule against real ledger data.

Module map

The depth, at a glance

5 modules — every one of them a first-class part of the same system, not an integration.

Effort-based distributionShared-pool distributionCost objectsAllocation runsAudit trail
The foundation

Written against the core — not copied from it.

It allocates directly against the core ledger, so the result is postings in your accounts.

Cost Distribution treats the ERP not as an external system to integrate with, but as its own runtime. It stores its data in the same model, posts to the same ledger, runs on the same workflow engine, obeys the same permissions and shows up in the same reports. On that foundation it adds what the platform lacks: the domain's objects, rules, language and screens.

The two-gap dilemma

Buyers are usually made to choose which gap to live with

Every business runs on a few hard things: a ledger that must balance, inventory that must be right, approvals that hold up to audit, and a permission model that keeps the wrong people out. Most of the market asks you to give up one side or the other.

Generic ERP

Broad core, shallow domain

Very good at the universal hard things — a ledger that balances, approvals that survive an audit. Not good at the specific way your operation actually runs, and that last mile of domain fit is where the value is felt.

Vertical app, beside the ERP

Deep domain, integration tax

Models the domain beautifully, then reaches back into the ERP through connectors that drift, duplicate data and quietly disagree. You pay a standing tax: engineering, reconciliation, latency, and doubt about which number is right.

Cost Distribution

Deep domain + native core

Uses the platform's data model, ledger, workflow and permissions as its foundation, and adds the domain depth on top. One system. Nothing to sync, because there is only one copy.

A custom point solution can do both, but only by re-implementing the accounting, workflow, security and reporting a mature platform already ships — most of the budget goes to rebuilding the floor before anyone reaches the product.

The integration tax, made visible

The clearest way to see the value is what disappears

Beside the platform

Two copies of the data · drift between syncs · reconciliation that never ends · every new feature re-plumbed back to the core · connectors that break whenever either side changes.

Integration is not a line item. It's a standing liability.

On the platform

One source of truth · nothing to reconcile · auditability that comes from the platform rather than a fragile mapping · saved engineering moved to the domain — the only place a product actually wins.

The product and the core cannot disagree, because they share the same records.

How we build it

Five principles, and the path a deployment follows

The same rules apply to every product we ship — they are what keep an upgrade routine instead of a migration project.

  • 1
    Build on the core, not beside it. The ERP platform is the runtime. If a capability exists in the core, we reach for it before building our own.
  • 2
    Inherit, don't rebuild. Accounting, permissions, workflow, audit and reporting are solved. Every hour re-implementing them is an hour not spent on your domain.
  • 3
    Speak the platform's data model. The domain is modelled in the core's own terms, so the product contributes to one ontology rather than a private schema — and is ready for agents.
  • 4
    Design for composability. Each product is a clean building block that others can extend and combine.
  • 5
    Ride the roadmap. Every platform upgrade, connector and ecosystem tool arrives at no extra cost.

How a Cost Distribution deployment lands

  1. 1Discovery
  2. 2Feasibility & estimate
  3. 3Design
  4. 4Staged build on staging
  5. 5UAT
  6. 6Launch
  7. 7Support

We migrate through staging before production and script every schema change, so each stage has a tested path back.

The full thesis — the six advantages, the integration tax in detail and the five-level maturity model we build against — is set out on the ERP-based development page.

Allocate every riyal to the right cost object — auditable, IFRS-aligned, automated.

Cost Distribution · Finance
How an engagement is priced

Licence and implementation are quoted per deployment — we estimate in ranges, not points, and every figure carries its assumptions, method and risks. Ask for a feasibility and you get a decision-ready go / no-go before anyone signs anything.

Feasibility & estimation →
Related

More in Finance

EX
Finance

Executive Command Center

The whole business on one screen — for the people who run it.

Learn more →
Keep reading

How we think about building this

Not ready to talk? These explain the method behind the product.

Insights

Why we build enterprise software on ERPNext instead of from scratch

The discipline that keeps vertical products fast to ship and safe to upgrade.

Read · 1 min
Guides

NPHIES, explained: what Saudi hospitals actually need from an HIS

A practical look at native insurance connectivity and revenue-cycle flow.

Read · 1 min
Product updates

How an availability grid stops a hotel from overselling

Inside the concurrency-safe engine behind Hotel Pro.

Read · 1 min

See Cost Distribution on your data.

Book a 30-minute demo and we'll show it running against a realistic version of your operation — not a canned dataset.