> ## Content Index
> Fetch the complete content index at: https://abundance.alloconomy.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# When the Chart of Accounts Keeps Growing
- URL: https://abundance.alloconomy.com/alloconomy/when-the-chart-of-accounts-keeps-growing/
- Published: 2026-11-14T18:52:45.000Z
- Updated: 2026-11-15T15:40:57.000Z
- Description: Name the growth driver: values, valid combinations, posting volume and the work each change creates.
- Author: Sathya Narayanan
- Tags: Alloconomy, Data Engineering, Chart of Accounts, #series-detail, #Import 2026-11-14 10:44

Data Engineering · Five essays

## From Clay Tablets to OpenAI

1. Part 1 [Learn from an old problem](https://abundance.alloconomy.com/alloconomy/the-chart-of-accounts-is-a-data-model/) An old, widely encountered problem offers ideas for deciding what modern accounting postings should retain.
2. Part 2 [Decide what belongs on the posting](https://abundance.alloconomy.com/alloconomy/what-deserves-an-account/) Distinguish natural accounts, segments, values and combinations, then decide where an attribute belongs.
3. Part 3 [Test four attribute choices](https://abundance.alloconomy.com/alloconomy/five-businesses-one-accounting-boundary/) One attribute decision each for Uber, Tesla, SpaceX, and AI providers OpenAI and Anthropic.
4. Part 4 **Understand the growth**Current essay Name the growth driver: values, valid combinations, posting volume and the work each change creates.
5. Part 5 [Explain why a report changed](https://abundance.alloconomy.com/alloconomy/accounting-agents-need-a-memory/) Use four report changes to test posting history, source evidence, allocations and dynamic groupings.

A mission identifier can earn its place in a posting combination and still be implemented badly.

The previous essay assumed we needed ledger balances and posting controls by mission. The next question is how much accounting structure each new mission creates. Does it reuse existing accounts and responsibility centers? Does it inherit every possible combination? Does someone create a new natural account for each mission and cost type?

Those choices have different growth patterns. To compare them, we need to name what we are counting.

## Five quantities that grow differently

**Master values** are the individual accounts, companies, cost centers and projects. If their counts are A, E, D and P, their combined count is **A + E + D + P**.

**Possible combinations** are the tuples those values could form. If every value can pair with every other, there are **A × E × D × P**. Business rules usually allow a subset of that space.

**Used combinations** are the distinct tuples that actually occur in postings within a stated scope. They can be much fewer than the possible or permitted combinations. If N journal lines use at most U permitted tuples, the number of used combinations is at most **min(N, U)**.

**Posting volume** counts the journal lines themselves. A million entries can reuse the same small set of combinations. Conversely, an identifier that is unique on each entry can make the number of used combinations grow with posting volume.

**Maintenance work** includes validation rules, mappings, [reconciliations](https://abundance.alloconomy.com/alloconomy/what-are-we-reconciling/) and exceptions. A rule can cover many combinations, so the number of combinations alone does not tell us how many rules a person must maintain.

These counts belong on separate lines in a design review. Each answers a different question about the system we are building.

## Adding projects: name the driver

Take a deliberately small example. A business has **120 relevant combinations** of company, natural account and cost center. We will call that fixed base **B = 120**. It is adding a Project segment to these entries.

Suppose each project needs at most **6** of the base combinations. With **P projects**, the design permits at most **6P** project-bearing combinations. At 50 projects, that is at most **300**. One additional project adds at most six.

Now consider a configuration that permits every project with all 120 base combinations. It permits **120P** combinations: 6,000 at 50 projects, with 120 more for each additional project.

| Projects P | 6 per project (max.) | 120 per project |
| ---------- | -------------------- | --------------- |
| 10         | 60                   | 1,200           |
| 50         | 300                  | 6,000           |
| 100        | 600                  | 12,000          |

**Both patterns are linear in P while B stays fixed.** The difference is the multiplier and the validity of what is permitted. The broader design may be appropriate if projects really use that breadth. If they do not, it creates more opportunities for miscoding and potentially more review work.

The database may store compact rules rather than materialize every permitted tuple. To estimate storage and close workload, inspect the ERP’s actual behavior and the combinations used. To assess control quality, inspect what those rules allow.

## When growth becomes quadratic

Suppose both the number of base combinations and the number of projects grow with business scale n. If B is proportional to n and P is proportional to n, and every project can combine with every base tuple, **B × P grows as n²**.

If each project continues to use at most six base combinations, the permitted set stays bounded by **6P**, which is linear in n. The restriction is meaningful only if it reflects how projects actually operate.

Add model version as another independent segment with V values and the unrestricted space becomes **B × P × V**. Whether this is linear, quadratic or cubic in a scale variable depends on which of B, P and V grow with that variable. Adding a segment increases the potential combinations; it does not establish one universal growth rate.

The Tesla warranty example gives us another useful comparison. Let **M** be the number of vehicle models and suppose there are **30 eligible base combinations** for warranty entries. If every model uses all 30, a Vehicle Model segment permits **30M** model-bearing combinations: 180 for six models, 210 for seven. That is linear in model count. Selling more vehicles within the same models adds claims and postings without adding model values.

With vehicle model retained only in the warranty subledger, the ledger’s combination count stays constant as M grows, under the same fixed base. The subledger still needs model definitions, balances and [reconciliation](https://abundance.alloconomy.com/alloconomy/what-are-we-reconciling/). The work has a different home; its cost must be included in the comparison.

## What turns growth into work?

A new segment value needs a definition, an owner, valid-use rules and a retirement policy. A new used combination may need a budget, [reconciliation](https://abundance.alloconomy.com/alloconomy/what-are-we-reconciling/) or mapping, depending on the business. New posting lines need processing and checking. A downstream slice needs evidence, mapping logic and historical versions.

The practical cost depends on how widely a change spreads. If every new mission requires edits to twelve reports and three interfaces, the repeated work is driven by the number of missions and those fifteen dependencies. If the reports consume an existing mission master and a shared mapping, adding a mission may require much less repeated editing. Exceptions can still dominate the workload.

This is why I would ask for two estimates before adding a dimension: **what will one more value require, and what will one change to its meaning require?** Count the manual touchpoints as well as the rows.

## Choices worth paying for—and choices to question

In the SpaceX exercise, the mission segment supports required balances and an open-mission check. Its cost may be justified. Reuse the natural accounts for labor, materials and other relevant categories; define valid mission combinations around real activity. Creating a new “labor expense for Mission 51” natural account repeats the mission classification inside the economic category and expands every account-based mapping that consumes it.

In the Uber exercise, a management region can be derived from retained city data. Making each regional reorganization a change to posting combinations would add accounting maintenance for a grouping whose purpose is reporting. The downstream choice still pays for a dated mapping and reproducible reports.

In the Tesla exercise, a modest model catalog could make the 30M design entirely affordable if finance needs model-level control in the ledger. If the warranty subledger already owns that control, repeating it in the ledger adds another set of classifications to [reconcile](https://abundance.alloconomy.com/alloconomy/what-are-we-reconciling/). Creating a separate natural account such as “warranty liability—Model Y” also makes account-to-statement mappings track the model catalog. A shared warranty-liability account plus a controlled model segment keeps those meanings separately maintainable. It can still produce the same number of used posting combinations; the advantage is where definitions and mappings are maintained.

In the AI-provider exercise, model releases already require usage classification. Creating model-specific compute-expense accounts would also make the natural-account list grow with the release catalog. Retained model detail and versioned allocations can answer the cost question while keeping economic categories stable. This design is worthwhile only if those allocations [reconcile](https://abundance.alloconomy.com/alloconomy/what-are-we-reconciling/) to the expense pools they distribute.

## The trade-off I would actually evaluate

For a proposed segment, name the control or reporting benefit, estimate values and valid combinations over a realistic horizon, and identify the work triggered by each addition or change. Compare that with the cost of retaining detail and deriving the attribute elsewhere.

A larger chart can be the sensible choice when it makes a necessary control dependable. A smaller chart can be expensive when it pushes essential [reconciliation](https://abundance.alloconomy.com/alloconomy/what-are-we-reconciling/) into fragile spreadsheets. The objective is to put each kind of growth where the business can maintain it—and to make the reason for paying that cost explicit.

---

[← Previous: Four Business Models, One Accounting Boundary](https://abundance.alloconomy.com/alloconomy/five-businesses-one-accounting-boundary/) · [Next: Accounting Agents Need a Memory →](https://abundance.alloconomy.com/alloconomy/accounting-agents-need-a-memory/) · [Complete series guide](https://abundance.alloconomy.com/topics/chart-of-accounts/)