Score your controls in ten minutes

Twenty questions across the five COSO components, scored in your browser, with advice for your weakest areas.

Engraving of an eye chart of ledger lines growing finer, with a magenta trial-lens frame in front.
Read the lines you can.
Control environment

The standards, oversight and accountability that decide whether people take controls seriously.

Has everyone in finance acknowledged your code of conduct in the last 12 months?

A code that people have never acknowledged is hard to enforce when someone breaks it.

Does someone outside the finance team, such as an owner or board member, review results at least quarterly and record their questions?

Review by someone who did not prepare the numbers is the first check on management override.

Is there a written approval matrix stating who can approve payments, contracts and journal entries, and up to what amount?

Without written limits, nobody can tell afterwards whether an approval was valid.

Can staff report a concern to someone other than their own manager, anonymously if they choose?

Tips are the most common way occupational fraud is detected, and they only arrive when people have a safe route.

Risk assessment

How you identify what could go wrong in finance, including fraud and change, before it does.

In the last 12 months, has management written down how the finance process could fail or be defrauded, ranked by consequence?

Controls chosen without a written view of risk tend to protect what is easy rather than what matters.

When you add a system, bank account, entity or AI tool, does someone decide which controls it needs before it goes live?

Control gaps often open the day something new is added and nobody is assigned its controls.

Can you name the balances and estimates most likely to be wrong at year end, and who reviews each one?

Knowing where a misstatement is likely tells you where review time is best spent.

Have you considered how a senior person could override controls, for example by posting and approving their own entries?

Management override goes around ordinary controls, so it needs its own detective checks.

Control activities

The approvals, reconciliations, reviews and system settings that prevent or catch errors and fraud.

Is every change to a vendor's bank details confirmed by calling the vendor on a number already on file?

Calling a known number defeats an attacker who controls only the email thread or the voice on the line.

Do payments above a set amount need a second person to release them in the bank portal?

With dual release, one compromised login or one mistaken click cannot send a large payment.

Does someone other than the preparer review manual journal entries above a set amount?

Manual entries are the simplest way to misstate results, and a self-approved entry gets no second look.

Are bank accounts reconciled within ten business days of month end and signed off by a reviewer?

A late or unreviewed reconciliation lets missing cash and unrecorded payments age unnoticed.

Information and communication

Whether the right people get reliable information in time, and know what to do with it.

Do management accounts reach leadership within 15 business days of month end, every month?

Information that arrives late cannot catch a problem while it is still small.

Are your finance policies written, dated within the last two years and kept where staff can find them?

People cannot follow a procedure they cannot find, or one that describes a process you no longer run.

When someone finds a control exception, do they know whom to tell, and is the exception recorded?

Exceptions that are fixed quietly hide the pattern that would point to a broken control.

Could you answer a lender's or funder's controls questionnaire from documents you already keep?

Evidence created on request makes outsiders doubt that the control runs the rest of the year.

Monitoring activities

How you check that controls still work, and how you fix what you find.

In the last 12 months, has anyone checked the evidence that your key controls actually ran?

A control that is well designed but skipped in practice gives no protection.

Are audit findings and control exceptions logged with an owner and a due date, and tracked until closed?

Findings without an owner tend to reappear in next year's management letter.

Do you run automated checks over all payments or journal entries, such as duplicate invoices or weekend postings?

Testing every transaction finds what a sample misses, and it costs little once the data is exported.

Does the board or owner receive a report on open control issues at least twice a year?

An oversight body can only push for fixes it knows about.

0 of 20 answered

Answer for what actually happens

Ten minutes, one entity, honest answers. The check runs entirely in your browser and stores nothing.

  1. Pick one entity and one person

    Answer for a single legal entity or business unit, ideally as the controller or finance lead who sees the work every week. If two entities run different processes, take the check once for each.

  2. Answer yes, partly or no

    Yes means it happens every time and leaves evidence someone else could check. Partly means it happens sometimes, depends on one person or leaves no record. No means it does not happen. When you are unsure, choose partly.

  3. Read the score by component

    You get an overall percentage and a separate score for each of the five components. The lowest component, not the total, usually shows where to start.

  4. Keep your own record

    Nothing is saved on our side or in your browser, and closing the page clears every answer. Print the result or take a screenshot if you want to compare next quarter.

How the score is built

The 20 questions are grouped under the five components of the COSO Internal Control: Integrated Framework (2013): control environment, risk assessment, control activities, information and communication, and monitoring activities. Each component has four questions, so each carries the same weight in the total.

A yes earns full credit, a partly earns half and a no earns nothing. Your overall score is the share of available credit, shown as a percentage. Each component is scored the same way, and the advice beside it changes depending on whether that component scored low or high.

Reading the bands

  • 0 to 39 percent. Basic controls are missing or leave no evidence. Start with the controls that stop cash leaving the business.
  • 40 to 69 percent. The core exists but depends on particular people or leaves gaps in the record. Fix the weakest component first.
  • 70 to 100 percent. The answers describe a working system. The open question is whether the evidence would back them up.

What a self-score cannot see

The check records what you believe happens. It cannot see that a reconciliation was signed three weeks late, that an approver has never rejected anything, or that the person who keys vendors can also release payments. Those answers sit in the ledger and in system access lists.

For the access question, the Segregation-of-Duties Conflict Finder is the quickest next step. To work through gaps on your own, use our internal controls checklist for growing companies. For a tested view of the same areas, a Controls Health Check compares your answers with your actual transactions, user permissions and evidence.

What people ask before starting

Is this an assessment of our internal control?

No. It is a structured self-review built from your own answers. It is not an audit, a review or any other attest service, and Finance Ethics does not provide those. Treat the result as a list of questions to follow up, not a conclusion about your controls.

Is anything I enter saved or sent to you?

No. The questions and the scoring run entirely in your browser. Nothing is stored, nothing is sent to our servers, and your answers are gone when you close or reload the page. Our privacy policy explains the aggregate page statistics we do collect.

Why are the questions built on COSO?

The 2013 COSO framework is the structure US auditors and many lenders already use to describe internal control. A gap you find here translates directly into the language of a management letter or a lender's controls questionnaire.

What should we do with a partly answer?

Treat it as a no with a head start. Ask what would make the control happen every time and leave a record, then name the person who will do it. Many partly answers become a yes through a system setting, a checklist or a signature rather than a new hire.

Can we share the result with our board or auditors?

Yes, as a starting point for discussion. Note who answered and on what date, and say clearly that the answers are self-reported and untested.

How often should we retake it?

Quarterly while you are closing gaps, then once a year. Retake it after any change in finance staff, accounting system, bank or ownership, because those are the moments controls tend to slip.

Test the answers against your ledger

A self-check measures what you believe. A Controls Health Check tests it against your transactions, system access and evidence in three to four weeks.

Writehello@financeethics.com
Call(307) 365-5252
Office7124 Commons Dr, Cheyenne, WY 82009
HoursMonday to Friday, 9:00 to 17:00 Mountain Time