Eight Checks Every Growing Business Needs for IdentityOps

Eight Checks Every Growing Business Needs for IdentityOps

Small and mid-sized companies tend to guard the front door well and forget the guest book. Someone joins, gets a laptop and a dozen logins, and years later nobody can say with confidence how many of those logins still exist. IdentityOps is a label some practitioners use for the routine work of keeping that guest book honest. Here are eight practical points, ordered from the cheapest to start to the most demanding, for business owners and technology managers who want a clear picture before investing.

1. Start with a definition you can defend

The Ops Atlas notes that IdentityOps is still an emerging term with no single agreed definition, and that different vendors stretch it to cover different ground. That is not a reason to wait. Pick a working scope for your company and write it down. The Atlas offers a tidy editorial split: IdentityOps concerns who exists and how they prove it, while neighbouring practices handle what they may do and how access is requested or reviewed. If you want to read how the Atlas frames the boundary, see its published entry on IdentityOps.

2. Treat the HR record as the starting gun

Every person account should trace back to a hire, a transfer or a departure recorded by whoever manages people data. The Atlas calls this the joiner, mover and leaver cycle and puts it at the centre of the discipline. In practice that means the HR system triggers account creation, changes group memberships when someone moves teams, and starts the shutdown when someone leaves. Where there is no automation, a shared checklist owned by a named person is still a large improvement on memory.

3. Make the departure an end-to-end event

A frequent mistake is to call a departure finished when the main login is disabled. The Atlas distinguishes disabled from offboarded: an account can be switched off at the central directory while a copy of it survives in an application nobody connected. Its worked example is instructive. Of forty leavers, thirty-six were disabled on time, though a trio of that group retained a working login in a less critical tool, so the honest figure was eighty-two and a half percent rather than the flattering ninety.

If you track one number from this list, track that one, measured from the application side rather than from the directory side.

4. Remember that sessions linger

Turning off an account does not instantly end the sessions and tokens already handed out. Depending on the platform, those can remain valid for an hour or more, and some collaboration tools keep their own sessions that the central directory cannot clear. A departure routine worth trusting therefore includes a revocation step and a spot check afterwards. It is a small addition to the runbook with a disproportionate effect on risk.

For a broader view of how technology choices fit a growing business, our earlier piece on choosing and combining business technology solutions sets useful context for where identity tooling sits among everything else.

5. Count strong sign-ins, not enrolled users

Many dashboards report that nearly everyone has a second factor. The Atlas warns against claiming full coverage when text-message codes and approve-a-prompt notifications are included, because attackers have learned to defeat both. It separates two measures. In its example, 1,190 of 1,400 people had a phishing-resistant method registered, roughly eighty-five percent, while barely two out of three real logins were made with one. The gap means the policy still allows weaker routes, and the weaker routes are the ones that get exploited.

A sensible rollout order, according to the Atlas, is the administrators of the identity platform first, then people with email, remote-access or finance reach, then the rest of the company.

6. Give every automated account an owner

Business systems talk to each other through service accounts, API keys and tokens. These non-human identities rarely appear in HR records, so nobody notices when they linger. The Atlas lays out a lifecycle: inventory them, record who owns each, what it is for, when it expires and how often it is reviewed; prefer credentials with short lives; rotate them; and decommission with evidence by disabling first, waiting, then deleting. A simple spreadsheet with those five columns is a respectable first version.

It also warns about a practice common in small teams: several humans sharing one service login. Shared credentials make it impossible to say who did what, and impossible to remove one person’s access without breaking everyone’s.

7. Design the recovery path before you need it

When someone loses a phone, the temptation is to let the help desk fix it quickly. The Atlas treats the help desk as a poor default, because convincing impersonation calls have been used to take over accounts at large organisations. Its recommendation is graduated. Self-service recovery relies on a second authenticator that the person already holds. Assisted recovery hands over a single-use code only once the person’s manager vouches for them face to face or on a live video call. Recovery for privileged accounts calls for supervised proofing plus a second administrator. Every recovery event should also send a notification to the account holder, which echoes guidance from the US standards body.

8. Put the checks on a calendar

Practices decay unless they are scheduled. The Atlas suggests a rhythm that a small team can actually keep:

  • A daily reconciliation between the directory and the applications.
  • A weekly sample of recovery events to confirm the script was followed.
  • A monthly report listing machine credentials close to expiry.
  • A quarterly drill in which a mock departure is followed through every system, plus a review of which outside identity providers the company still trusts.

A dormant-account sweep fits into the same calendar. The Atlas proposes a window of ninety to one hundred eighty days, with exclusions for people on HR-approved leave, emergency accounts and machine credentials with a registered longer review cycle.

Spending wisely on identity tooling

Business buyers often ask whether a governance suite is the first purchase. The Atlas advises caution: if HR data is untidy, buying an expensive product only automates the confusion. It also reminds readers that many software-as-a-service products reserve provisioning features for their higher-priced plans, and that some identity capabilities depend on particular licence tiers from the platform vendor. Check those details before committing to a design, not after.

If you are planning the wider digital roadmap of your company, the article on sequencing digital growth plans offers a way to slot identity work into a sequence alongside other priorities.

One more misconception is worth clearing up for business readers. People sometimes avoid single sign-on because it seems to put every login behind one door. The Atlas answers that it concentrates control, not just risk, and that the central directory should be treated as the most sensitive system in the company, protected accordingly. Giving staff one well-guarded entrance and one place to cut access is generally easier to manage than scattered passwords.

A realistic ninety-day outline

For teams that like a staged plan, the Atlas sketches three checkpoints at thirty, sixty and ninety days, and a modest crew could adopt them nearly verbatim.

  1. Days one to thirty: assign owners, list the applications in use, put hardware security keys in the hands of administrators and write the recovery steps down.
  2. Days thirty-one to sixty: connect HR data for one group of staff, automate provisioning for the ten applications that matter most and build the machine-account inventory.
  3. Days sixty-one to ninety: add session revocation to departures, run the first complete drill and set up alerts for password resets and changes to federation settings.

What the public incidents have in common

The Atlas collects several widely reported breaches, among them a test environment with no second factor, a service credential saved in a personal cloud folder and a data platform where credentials went unrotated for years. Its observation is that each case defeated a business process, not an encryption scheme. That is good news for a company of any size, because processes are exactly the thing you can fix with a checklist, an owner and a calendar.

The takeaway is modest and actionable. Know who exists, remove them properly when they leave, strengthen the sign-ins that matter most, give machines owners and rehearse recovery. Those five habits make up most of what the label describes.