How to Review a Spreadsheet

The ICAEW (https://www.icaew.com/) is asking for comments on this draft:
Call for comments - How to review a spreadsheet | ICAEW

Oh jeez, you have to register for access.

Well, if anybody here is an English and/or Welsh accountant, please let us know if there’s anything useful here.

1 Like

Update: the people at ICAEW has informed (spreadsheet risks group I’m in) that the page was set up incorrectly and was meant to be public.

They’re working on fixing the situation.

That darned operational risk.

2 Likes

Okay, it’s available now, here:
https://www.icaew.com/-/media/corporate/files/technical/technology/excel-community/spreadsheet-review-publication.ashx?la=en

It’s 17 pages long, some of those pages are essentially blank/fluff.

Might be interesting to compare/contrast against ASOP 56

I will pull out a few things:

Several categories of errors can occur within a spreadsheet, including:
Data integrity errors – the source information is incorrect, leading to the infamous problem of “garbage in, garbage out”.

Formula errors – the data is incorrectly used in calculations, providing flawed results. Some of these can be detected during development and application by employing quality control techniques that consist of built-in tests (such as a check/flag that indicates whether the balance sheet balances).

Process errors – the management of the spreadsheet is compromised through errors such as using an incorrect version.

Communication errors – the spreadsheet is misleading or its results aren’t presented or explained well.

Meta errors – these exist outside the scope of the spreadsheet’s cells. For example, where a spreadsheet is not the best tool to answer the question at hand, or that the question isn’t the right one to be asking. These should be considered early on to avoid wasting effort on a detailed review of a fundamentally flawed approach.

There are five types of reviews:
Structural review: checking that the spreadsheet is well laid out, properly built and includes appropriate checks.

Data review: checking the data and assumptions are correctly sourced (e.g. the correct tax rates and thresholds are being used).

Coding review: checking the validity of calculations (e.g. tax is correctly calculated from cell references for a profit number and a tax rate).

Commercial review: checking that appropriate business, accounting, and tax rules are being applied correctly.

Analytical review: checking if the numbers look reasonable for the scenario being modelled.

so…it’s just giving things flashy names without actually telling you how to review a spreadsheet?

How do I identify each of the errors without recreating the whole process?

When a library gets updated, a testing suite is executed and if any of the tests encounter a bug, you need to address it.

A testing suite is written cumulatively over the course of many years by many people and it consists of individual use cases.

I’m not sure how that would be implemented for spreadsheets though. This is one reason why I don’t use them.

That’s just from the beginning of the document.

They do get into some very specific techniques to build in checks, etc.