Customer Access Control: Who Can See and Do What

Two customers can use the same application while accessing different data. The key is to define permissions and check them against each requested resource.

By Daniel Sardá · Published on

6 min read1,127 words

In this article · 6 sections

Why can two accounts using the same application see different invoices? Using a service does not mean having permission to access everything it contains. The application needs to decide what each requester can view or change.

Customer access control can describe this separation of permissions. The context matters: it is not a universal technical term, and “client” can refer either to a customer account or to a piece of software. Here, the focus is access to digital resources in an application connected to a remote service.

There is another important distinction: assigning permissions by customer and performing checks on the client side are different things. One concerns who is allowed access; the other concerns where a check takes place.

One request, four questions

Imagine a portal where two companies, A and B, view their invoices. Someone signs in with an account associated with A and asks to download a document. To decide whether to provide it, the service needs to answer four questions:

  • Who is asking: which account or identity is acting and, where relevant, whom it represents.
  • Which resource is requested: the particular invoice.
  • What action is requested: viewing, downloading, modifying, or deleting it.
  • Which rule permits it: for example, the account may view A's invoices but not modify them.

Permission depends on that relationship. Being able to open the “Invoices” section does not mean being able to read every invoice in the portal. Nor does permission to view a document imply permission to change it.

OWASP explains the need to check the specific object and action in its guide to broken object-level authorization in APIs. An API is a way for an application to request data or actions from another system; the same problem arises when the portal requests an invoice from the service that stores it.

Key idea: A permission should establish who can do what to a specific resource. Access to a section does not grant access to everything in it.

An application can also be a client

In the example, “customer A” identifies a company using the service. In computing, however, the application sending the request can also be called a client: a browser, a mobile app, or an automated program.

That application may act on someone else's behalf. The OAuth 2.0 specification's definition of roles distinguishes between the client application and the resource owner. This distinction helps avoid a mistake: knowing which program sends a request does not, by itself, establish whose data is involved or which actions are permitted.

Claiming an identity, verifying it, and granting permissions

In the invoice portal, entering a username identifies an account. Proving that someone can act as that identity is the job of authentication, using whatever mechanism the service provides.

A separate question remains: what can that identity do? That is a matter of authorization. An authenticated account associated with A may be authorized to download A's invoices while having no permission to view B's.

Separating authentication from authorization prevents “signed in successfully” from being confused with “allowed to access everything.” It does not mean every resource requires a login: a service may allow people to view public information without authentication. The access rule should reflect that public status.

The OWASP authorization guidance distinguishes these functions and recommends checking permissions on every request. Passing an earlier check does not automatically settle the next one.

Where the permission is checked matters

The portal can hide the “Delete invoice” button from someone who has read-only access. This helps guide the user: it shows the actions available and avoids offering options that would be rejected.

But the interface alone cannot protect the document. In a client-server application, someone can try to send a request without using the intended buttons. Checks run in a browser or an application under the user's control should therefore not be the decisive authorization barrier.

According to the same OWASP guidance, the decision should be enforced in a trusted layer, such as a server, gateway, or functions that enforce the access policy. This does not require a single physical server: it requires that the requester cannot bypass the check by modifying their own interface.

Key idea: Hiding a button helps people use an application. Protecting the resource also depends on the system rejecting unauthorized requests sent through other means.

This clarifies the two senses of the initial expression. Permissions by customer can distinguish what A and B are allowed to access. Client-side controls can adapt what each person sees on screen. To protect remote data, authorization must remain effective beyond that screen.

The two-invoice test

Suppose A views one of its own invoices and then changes the identifier in the request to that of an invoice belonging to B. Even if the portal shows no links to another customer's documents, the service must check whether A has permission to view that specific document before providing it.

If the service responds simply because the account is signed in or can access the invoice section, an essential check is missing. OWASP describes this kind of failure in its object-level authorization documentation. Hard-to-guess identifiers are no substitute for permission either: the decision must hold even if the requester knows the identifier.

The correct outcome depends on the service's policy. If A and B do not share documents, the request should be rejected. If B has granted A valid access to that invoice, viewing it may be allowed within the scope of that grant. Account isolation means preventing unauthorized access, while allowing resources to be shared when a rule authorizes it.

The action matters too. Shared read permission does not by itself grant permission to change the amount or delete the invoice.

Permissions with clear limits

Limiting access to data helps protect the privacy and autonomy of those who use it. In the portal, it means that subscribing to a shared service need not expose a customer's invoices to every other account.

This protection has a specific scope. Permission to view an invoice does not determine how a downloaded copy should be used afterward. Nor do portal permissions alone guarantee the security of the entire system or resolve every question about data use.

Key idea: A useful policy states both what it permits and its limits: the resource, the action, and the conditions of access.

To understand how a service works, return to the specific request: which account is asking for the invoice, whom it belongs to, what the requester wants to do with it, and which component checks the rule before responding. When those answers are clear, “customer access control” becomes less ambiguous and describes a decision that can be checked.

About the author

Daniel Sardá is an SEO Specialist, a university-level technician in Foreign Trade from Universidad Simón Bolívar, and editor of Libertatis Venezuela. He writes on liberalism, political economy, institutions, propaganda and individual liberty from an independent, non-partisan perspective.

Read full profile