Publicité
ERP IMPLEMENTATION
🇫🇷 Lire en français →

Orphaned ERP: Take Back Control in 30 Days After Your ERP Lead or CIO Leaves

Your CIO or ERP administrator just left: a 4-week recovery plan, a risk matrix, a system-mapping template and the mistakes to avoid.

Orphaned ERP: Take Back Control in 30 Days After Your ERP Lead or CIO Leaves

On a Monday morning, IT leadership learns that the ERP administrator has resigned, or that the CIO who led the project is leaving the company. The notice period is three months, sometimes one week. Nobody else really knows how the ERP is configured, which interfaces run overnight, or where the passwords for the technical accounts are kept.

This is key-person risk applied to the ERP. It hits mid-market companies hardest, where one or two people hold years of decisions that were never written down. This guide is for the CEO, the CFO and the incoming CIO who inherits the situation: a four-week recovery plan, with a concrete deliverable at the end of each week. It covers the departure of an internal person. If your dependency is on the implementation partner instead, read our guide on knowledge transfer with your integrator.

The Scenario: What Breaks When the ERP Administrator Leaves

An ERP does not stop on the day someone leaves. It keeps running, and that is precisely the trap: the risk is silent. Incidents show up weeks or months later, when an event that only the departed person knew how to handle comes along. Failures fall into four families.

Access

The administrator often holds the privileged accounts: application super-user, database account, service accounts used by interfaces, access to the vendor’s support portal. If those credentials are personal or stored in a local file, the departure creates two opposite problems: nobody can do anything without them, and they can still do everything until their access is revoked. The second point is a security issue, covered in our guide to ERP access governance.

Interfaces and Batch Jobs

Flows to the bank, payroll, e-commerce, logistics or the e-invoicing platform run on scheduled tasks, scripts and certificates. A nightly job that fails without an alert, a certificate that expires, a service-account password that changes: nobody knows where to look, because the logic is documented nowhere.

Undocumented Configuration

Why is this general ledger account excluded from automatic reconciliation? Why does this pricing rule apply to this one customer only? These choices had a reason, but it lived in the administrator’s head. Without it, every change becomes a gamble.

The Business Calendar

Month-end close, VAT returns, dunning runs, stock counts: the administrator knows the sequences and the usual sticking points. A departure rarely lands at a convenient time, and the first month-end close without them is the real test.

Week 1: Secure

The goal of the first week is not to understand, it is to avoid the irreparable. Map nothing until access and backups are under control.

Privileged Accounts

  • List every administrator account: application, system, database, vendor portal, service accounts.
  • Tie each account to a named individual or a designated owner. A generic account with no owner is a risk.
  • Create at least two named administrators on the company side (the new CIO and a deputy), so you no longer depend on a single person.
  • Schedule the deactivation of the departing person’s access for their exit date, not after.

Passwords and Secrets

Collect technical-account passwords, API keys, certificates and interface tokens, and place them in a company-managed secrets vault. Once the secrets have been transferred, rotate them: a secret known to someone who is leaving should be treated as compromised on principle, without implying any bad intent.

Vendor and Integrator Contracts

Check who the named contact is at the vendor and the integrator, which support and maintenance contracts are active, and which address receives renewal notices. The reference contact is sometimes the departing person’s personal mailbox. Have the contacts and ticket-raising rights updated. For how these contracts are structured, see our guide to the application management (AMS) contract.

Tested Backups

A backup you have never restored is a hypothesis. Ask for a test restore on a separate environment and record the actual duration. Also verify the backup of everything outside the database: custom developments, scripts, configuration files, custom reports.

Week 1 deliverable: privileged-account register, populated secrets vault, up-to-date vendor and integrator contacts, proof of a successful restore.

Week 2: Map

Once the system is secured, take inventory. The aim is a shared overview, not exhaustive documentation. If the departing person is still around, this is the week when every hour of their availability counts: schedule short sessions, recorded with their consent.

Mapping Template

Copy this outline and fill it in, module by module:

  1. Active modules: which ones are actually used, by which users, at what licence level.
  2. Custom developments: name, function, author, date, last change, whether the source code sits in a version-controlled repository.
  3. Interfaces: source and target system, protocol, frequency, format, business owner, monitoring method, recovery procedure on failure.
  4. Scheduled jobs: name, schedule, dependencies, usual runtime, failure alert, impact if the job does not run.
  5. Sensitive configuration: non-standard business rules, with the business reason when known.
  6. Close calendar: monthly, quarterly and annual milestones, with the associated ERP actions and who performs them.
  7. Environments: production, test, development, and their synchronisation status.
  8. Contacts: vendor, integrator, hosting provider, key users by domain.

For each line, add a “source” column showing where the information came from: existing documentation, interview, or observation of the system. This separates what is certain from what is assumed, which matters in week 3.

If you are also preparing a migration, the same mapping feeds the audit of your existing ERP.

If the Person Is Still Here: The Handover Interview

During the notice period, the scarcest resource is the departing person’s time. Avoid vague “documentation sessions”. Prepare five questions instead, asked module by module:

  • What has broken before, and how did you fix it?
  • Which tasks do you do by hand, regularly, that no document describes?
  • Which configurations or developments should never be changed without care, and why?
  • Who at the vendor, the integrator or in the business actually knows the answers?
  • What worries you about the next quarter?

The last question is often the most productive: it surfaces risks the person knows about but never had time to address. Record the answers (with consent) and attach them to the mapping lines. If the departure is acrimonious or immediate, put the same questionnaire to their colleagues, key users and the integrator, who can reconstruct part of the missing memory.

Week 2 deliverable: the completed mapping, with its “unknown” areas openly acknowledged.

Week 3: Assess the Risk

You now have a list of objects. The next step is deciding which deserve immediate action. Cross two axes: criticality to the business (what happens if this stops?) and the level of knowledge available today (who can fix it, and is there documentation?).

Criticality x Knowledge Matrix

Score each line from 1 to 3 on each axis. Risk is high when criticality is high and knowledge is low. Here is an example table with 10 lines, to adapt to your context:

ItemCriticalityKnowledgePriority
Monthly financial close31Immediate
Bank interface (statements, payments)31Immediate
E-invoicing flow32High
Payroll interface32High
Nightly stock valuation job21High
Custom pricing development21High
E-commerce interface22Medium
Custom management reports11Medium
User permissions and profiles23Low
Test environment12Low

This table is an illustration, not a benchmark. The scores come from your own teams. Have them validated by business owners: a CFO will judge the criticality of the close better than IT.

The “3 / 1” lines are your emergencies. For each one, ask a single question: can it be documented, tested or taken over by a third party before the end of the notice period?

Week 3 deliverable: validated matrix, top 5 risks with an owner and a due date.

Week 4: Recovery Plan

You know what is critical and fragile. The last week is for choosing how to rebuild the capacity to run the ERP.

Hire, Outsource or Rely on the Integrator

OptionWhen to prefer itWatch-out
Hiring an administratorCentral ERP, frequent changes, enough volume to justify a roleLong hiring lead time, ramp-up period
Managed services or contractors from a providerNeed for responsiveness, scarce profile on the marketDependency is shifted, not removed
Original integratorRecent project, heavy custom developmentCommercial leverage, see the knowledge-transfer guide

In practice, a mixed model is common: an internal person who holds the business knowledge and makes the calls, backed by a provider for technical depth. In that case, make sure the provider documents what it does.

Minimum Documentation

Do not launch a full documentation programme. Aim for the useful minimum: one runbook per critical process in the matrix (what to check, how to restart, who to call) and the week 2 mapping, kept up to date.

Buddy System

Name a deputy for each critical item. A buddy arrangement has to be tested: ask the deputy to carry out a procedure alone while the owner watches. If the procedure fails, the documentation is incomplete.

Week 4 deliverable: 90-day recovery plan, chosen support model, budget, first runbooks, named deputies.

Prevention: Making Sure It Does Not Happen Again

The best recovery is the one you never have to run. Three practices durably reduce key-person risk.

Living minimum documentation. Runbooks and the mapping are updated with every change, as a condition for closing a change request. A document that is not reviewed every quarter goes stale.

Rotation. Rotate sensitive operations (the close, restarting an interface) between at least two people. Rotation exposes gaps while the owner is still there to fill them.

Reversibility clauses. In contracts with the vendor, integrator and hosting provider, require handover of documentation, source code for custom developments and configuration, plus transition assistance. The topic ties into vendor dependency and reversibility.

Mistakes to Avoid

MistakeCountermeasure
Waiting until the last day to actStart week 1 as soon as the departure is announced, while the person is available
Asking the leaver for exhaustive documentationTarget the 5 risks in the matrix and record short sessions
Forgetting to revoke or change access and secretsBuild revocation into the exit procedure and rotate shared secrets
Hiring a replacement without an assessmentDo the mapping and the matrix before writing the job description
Handing everything to the provider with no frameworkRequire runbooks, deliverables and reversibility in the contract

Taking Action

An orphaned ERP is fixed faster than you might fear if you proceed in order: secure, map, assess, decide. If you would like an outside view of your situation, get in touch for an ERP audit: access review, mapping and risk matrix in about two weeks.

To go further, read our guide on knowledge transfer with your integrator and the one on auditing your existing ERP.