A file landed on my desk last month with 240 active nominal codes. Eleven were variations on "office costs". Two held the same broadband bill in different months, because the bookkeeper faced a coin-flip on every transaction and didn't always call it the same way twice. The owner thought a detailed chart of accounts meant a well-run one. It meant the opposite.
TL;DR: On most new client files the first structural thing I do is put the chart of accounts on a diet. A chart of accounts is a reporting tool, not a filing cabinet. Every extra nominal is one more coding decision and a P&L line too thin to mean anything. Detail belongs in tracking categories, not the account structure.
Why Do I Strip the Chart of Accounts Before Anything Else?
Because a sprawling chart is the quiet cause of half the coding errors I find. A cluttered structure is a silent driver of bank-feed miscategorisation, which flows straight into the reports an owner decides on (Intuit, 2025).
The test I apply to every existing nominal is one question: would I ever want this on its own line in a management report? If the answer is no, it shouldn't be an account. It should be a tracking category, a contact, or nothing at all. That's the same diagnostic instinct that has me reaching for Excel before a reporting tool on a new file. Most small businesses run comfortably on a lean list of roughly 30 to 60 accounts (AYMiller, 2025), and the 240-code file collapsed to 54 without losing a single number anyone reported on.
The Three Places Sprawl Always Hides
In my experience it hides in the same three spots on nearly every file.
The first is cost-of-sales over-segmentation: a separate code for each supplier or material type, so the gross margin you need in one glance is spread across fifteen rows. The second is a dozen near-identical overhead codes - "Software - Adobe", "Software - Microsoft", "Software - Canva" - where one "Software" account plus a contact breakdown does the same job without the daily guesswork.
The third is the dangerous one: the "sundry", "misc" and "general" accounts that start as a holding pen and quietly become the biggest number on the page. When I open a file and the largest expense line is called "general", I know the diagnostic work hasn't been done. That's the same instinct I bring to a five-minute bank-feed hygiene check on a new engagement.
Should Detail Live in Nominal Codes or Tracking Categories?
Tracking categories, almost always. Owners find this counter-intuitive, because cutting an account feels like losing the detail. You're not losing it. You're moving it somewhere that slices the report without fragmenting it, which is what makes a file legible to whichever cloud reporting tool you run downstream.
Xero gives you two active tracking categories, each holding up to around 100 options before reports start to slow (Xero Central, 2026), with contact-level reporting on top. QuickBooks does the same with classes and tags. The nominal answers "what kind of cost is this"; the tracking category answers "which part of the business, which project, which client". Two different questions, two different tools.
Where Is the Floor on Cutting Accounts?
There is a floor, and knowing where it sits is the actual skill. Three things earn their own line every time. Statutory presentation comes first - your accounts have to map cleanly to the formats Companies House and HMRC expect. VAT analysis comes second, because a code that mixes standard-rated and zero-rated supplies costs you more at return time than it ever saved you. And genuinely decision-relevant splits come third: if you manage two product ranges as separate businesses in your head, they can have separate revenue lines.
The software draws its own version of the floor. FreeAgent enforces one, locking its pre-set categories and constraining custom codes to fixed ranges - income to 001-049, cost of sales to 100-199, admin expenses to 200-399 (FreeAgent Support, 2026). Xero is looser, taking any unique code of up to 10 characters (Xero Central, 2026), which is exactly why a Xero file drifts toward sprawl and a FreeAgent one rarely does. The discipline is knowing where the floor sits, not maximising or minimising for its own sake.
Frequently Asked Questions
How many nominal accounts should a small business chart of accounts have?
There's no fixed number, but a lean list of around 30 to 60 accounts covers most small businesses. The test isn't the count, it's whether each account would ever appear on its own line in a report you actually read. If it wouldn't, it's clutter, and clutter is where coding errors breed.
Should I use separate nominal codes or tracking categories to analyse my costs?
Use nominal codes for the type of cost and tracking categories or contacts for everything else - the department, project, client, or supplier it relates to. In Xero you get two active tracking categories; QuickBooks uses classes and tags. This keeps reports dimensional without forcing a coding decision on every transaction.
Why does my accountant want to delete accounts from my chart of accounts?
Because a thinner chart almost always produces cleaner reports and fewer mistakes. Each extra account is another place the same cost can land in two different rows across two months. Collapse near-identical codes into one account, move the detail to a tracking category, and your monthly comparatives and year-end both get faster.
If your chart of accounts has crept past the point of being useful, get in touch. Happy to run the diet over your file and show you what your reports look like once the sprawl comes out.