Go-live has happened, the project team is dispersing, the integrator moves into support mode and leadership asks: “So, who looks after the ERP now?” In many mid-sized companies and SMEs, the honest answer is: nobody in particular. A bit of the IT director, a bit of the finance director, a bit of the ticket open with the vendor.
This article is not about the maintenance contract: that is the subject of our ERP maintenance contract guide. It is about the organisational decision that comes first: who owns the ERP day to day, and in what form. Three models compete, and a fourth (the hybrid) is in practice the most common. We compare them, then recommend one model per company profile.
A note on method: the cost figures below are working assumptions drawn from our own project experience (Dicentur / DigitsLane), not market statistics. They are meant to compare the models with each other, not to build a budget without quotes.
The Organisational Vacuum After Go-Live
During the project, governance was clear: a steering committee, a project manager, an integrator, business referents. The day the project closes, all of that disappears within a few weeks, just as the ERP starts its real life.
This vacuum has three origins.
The integrator walks away. They remain reachable, but their project team is reassigned. The consultant who knew your configuration has moved to another client. The support contract guarantees a response time, not knowledge of your company. We cover this in our article on knowledge transfer from the integrator.
The project manager role ends. Often internal, sometimes on a fixed-term contract, they return to their original job or leave the company. If nobody inherited the role, you end up with an orphan ERP.
No dedicated budget or position. The project had an envelope. The run phase often has neither a budget line nor a job description. We broke down these costs in the article on the 7 line items of ERP run cost.
The symptoms are familiar: tickets with no owner, change requests piling up, Excel workarounds coming back, key users losing heart. We listed them in the article on post-deployment governance and the centre of excellence, and in the one on the 8 adoption KPIs. Here the question is more practical: who gets the wheel?
Before choosing a model, let us pin down the four responsibilities that must be covered, whichever model you pick:
- Arbitrate: decide what goes into the backlog and in what order (a business responsibility).
- Operate: handle incidents, manage access rights, monitor interfaces (a technical responsibility).
- Evolve: configure, develop, test and deploy changes (a technical and functional responsibility).
- Report: measure, present and defend the budget in front of leadership (a governance responsibility).
Each model covers these four functions differently. None covers all of them without trade-offs.
Model 1: In-House ERP Product Owner
The role
The ERP product owner (PO) is an employee who treats the ERP as a product: they own the vision, maintain the backlog, set priorities with the business and sign off deliveries. They do not necessarily do the configuration themselves. They are the single point of contact on the company side, facing both users and suppliers.
The profile
The right profile is rarely a pure IT person. It combines three skills:
- knowledge of the company’s processes (finance, purchasing, production, sales) and of the configuration done during the project;
- the ability to arbitrate and to talk with both leadership and operational teams;
- enough technical literacy to challenge a supplier without depending on them.
The natural talent pool is the most engaged project manager or key user from the project. Spotting them before go-live, not after, is one of the most profitable decisions you can make. Our article on the super-user programme describes how to bring these profiles out.
The workload
Working assumption: in a company of 100 to 300 employees with an ERP covering finance, purchasing, sales and inventory, the role takes 0.5 to 1 full-time equivalent (FTE) in the first year. It drops in steady state, but rises again with every version upgrade or new regulatory obligation. Below 100 employees it is more like 0.2 to 0.4 FTE, which immediately raises the question of combining it with another job.
The risks
- Key-person risk: knowledge of the configuration sits with one person. It is the first factor that tips a company into the orphan ERP. Remedy: a pairing arrangement, living documentation, a trained deputy.
- Overload: the PO combines their original job with the ERP role. At closing time the run work takes second place and tickets age.
- Technical isolation: alone, they cover neither operations, nor development, nor security. They need a supplier or a team to operate.
- Hierarchical positioning: a PO without credibility with leadership cannot arbitrate between two departments.
When it fits: as soon as the ERP is the operational heart of the company and business knowledge is the limiting factor. It is the only model that builds lasting in-house expertise.
Model 2: Fractional CIO or Part-Time ERP Manager
The use case
The fractional CIO (also called outsourced CIO) is an experienced professional who works a few days a month, or a few days a week, for several companies. They bring governance, not production: scoping, arbitration, supplier management, budget, security, roadmap.
This model suits companies with no CIO, or with a small IT team focused on user support and infrastructure and lacking the skills to steer an ERP.
The typical scope
- set up and run a quarterly change committee;
- manage the integrator or the managed-services provider: review of tickets, SLAs and invoices;
- own the run budget and the roadmap (versions, regulatory obligations, integrations);
- carry security and compliance topics to leadership;
- coach the key user or future PO as they build up their skills.
The limits
- Availability: a few days a month is not support. A blocking incident on a Thursday during closing cannot wait for the next visit.
- Detailed ERP knowledge: the fractional CIO sees your ERP from a distance. They do not know your configuration quirks as well as an employee or a dedicated consultant.
- Dependence on one person: key-person risk exists on the supplier side too. It is covered by contract (replacement, documentation, access to deliverables).
- Perceived cost: the day rate looks high, but it should be compared with the full cost of a full-time executive, including payroll costs, management and recruitment.
Working assumption: a package of 2 to 4 days a month covers the steering of an SME or mid-market ERP, excluding production work. Ask for a detailed quote and compare it with the full cost of an internal position.
When it fits: for companies of 50 to 150 employees without a CIO, or when the existing CIO is a generalist who needs reinforcement on ERP governance. Also during a transition: for the 6 to 12 months while the future PO is recruited or trained.
Model 3: Managed Services and Application Maintenance
What the model covers
Application maintenance outsourcing (known in France as TMA) hands corrective, sometimes evolutive, maintenance of the ERP to a provider. Managed services go further: the provider takes over operations, monitoring, version upgrades and sometimes hosting, under contractual service levels.
For the detail of the clauses (SLAs, scope, indexation, penalties), see our ERP maintenance contract guide. Here we stick to the organisational decision.
When it fits
- the company has neither the size nor the wish to build an ERP team;
- the ERP runs on a specialised technical platform (SAP, Dynamics 365, Oracle) where skills are scarce and expensive to hire;
- the company requires extended coverage (on-call, multiple time zones) that an in-house team cannot provide;
- the ERP is standard, lightly customised, with a predictable ticket volume.
The limits
- Managed services do not decide for you. They process the tickets you submit. Without someone to prioritise on the company side, the backlog fills with contradictory requests and the invoice follows.
- Business knowledge is low: the provider knows the tool, not how you work. They execute, they do not push back.
- Dependence is at its highest: if the provider alone holds the knowledge of the configuration and of custom developments, changing provider becomes a project in itself.
- Conflict of interest: a provider paid for time spent has no interest in simplifying your ERP.
Exit clauses
Whichever provider you choose, three clauses determine your future freedom:
- Up-to-date documentation delivered continuously (configuration, custom developments, interfaces), not just at the end of the contract.
- Direct access to the environments, the ticket logs and the source code of custom developments.
- A costed reversibility plan: transition duration, scope transferred, cost, overlap period with the new provider.
A simple rule: if the contract does not say how you get out, you have not negotiated the exit, you have only negotiated the entry.
Hybrid Models, Key Users and the Change Committee
In practice, no company of 100 to 500 employees applies a pure model. The combinations that work rest on one idea: keep in-house what requires business knowledge and decision-making power, and outsource what requires technical capacity and availability.
The three most common combinations
- In-house PO + managed services: the PO arbitrates and validates, the provider executes. This is the healthiest combination for a mid-sized company with a standard ERP.
- Fractional CIO + managed services: the CIO steers the provider and the roadmap, the provider operates. Suited to SMEs with no in-house ERP resource.
- Fractional CIO, then in-house PO: the CIO secures the first 6 to 12 months and trains the PO, then scales back.
The minimal RACI
| Activity | ERP PO | Key users | CIO / fractional CIO | Provider / integrator | Leadership |
|---|---|---|---|---|---|
| Prioritise the backlog | R/A | C | C | I | I |
| Handle an incident | A | R (level 1) | I | R (level 2) | I |
| Deliver a change | A | C (acceptance) | C | R | I |
| Steer SLAs and invoices | R | I | A | C | I |
| Run budget and roadmap | R | I | R | C | A |
| Security and access rights | C | C | A | R | I |
R: responsible. A: accountable. C: consulted. I: informed. Adapt it: what matters is that every row has exactly one A.
The change committee and the backlog
Three rituals are enough to hold governance together in a company of this size:
- Weekly ticket review (30 minutes): the PO, a representative of the provider, tickets open for more than ten days.
- Monthly or quarterly change committee (1 hour): PO, key users, CIO, finance leadership. The backlog is arbitrated on a single criterion, for example business impact divided by effort.
- Annual review with leadership: budget, roadmap, indicator review, decision to keep or change the model.
The backlog must be single and visible: one tool, one ordered list. Requests that arrive by email, by phone or at the coffee machine are copied into that list, or they do not exist.
Steering indicators
Three families of indicators let you judge the chosen model, whichever it is:
- Tickets: volume per module, age of the stock, share of reopened tickets.
- Resolution time: median and high-percentile time by priority level, compared with contractual commitments.
- Adoption: actual module usage, rate of workarounds outside the ERP. See the 8 adoption KPIs.
Set targets from your own data over the first three months rather than borrowing generic benchmarks.
Comparison Table
The ratings below are qualitative, drawn from our project experience. Costs are assumptions to be validated by quotes.
| Criterion | In-house PO | Fractional CIO | Managed services |
|---|---|---|---|
| Cost (order of magnitude, assumption) | Full cost of a 0.2 to 1 FTE position | Package of a few days a month | Fixed fee or time-and-materials, varies with scope and SLA |
| Nature of cost | Fixed, payroll | Semi-fixed, adjustable | Variable, indexed |
| Incident responsiveness | Good in office hours, fragile during leave | Low (not designed for support) | High if SLAs are contractual |
| Business knowledge | Strong | Medium | Low to medium |
| Arbitration capacity | Strong if credible | Strong (outside view) | None (executes) |
| Dependence | Low on third parties | Medium (one person) | High |
| Key-person risk | High (knowledge concentrated) | Medium (coverable by contract) | Medium (provider turnover) |
| Expertise built | In-house, lasting | Partial, transferable | At the provider |
Reading: no model wins on every row. The PO is strong on knowledge and arbitration, weak on technical capacity. Managed services are strong on capacity and responsiveness, weak on knowledge. The fractional CIO fills the governance gap between the two.
Recommendation by Company Profile
The thresholds below are guideposts, not rules. The complexity of the ERP (number of modules, customisations, interfaces) weighs as much as headcount.
50 to 100 employees
- Recommended: fractional CIO (2 to 3 days a month) + integrator support or light managed services. A key user appointed as “ERP referent” at 10 to 20% of their time.
- Why: the volume does not justify a dedicated PO, but the lack of steering is costly in drift.
- Watch point: do not leave the ERP referent without an official reduction of their original workload.
100 to 300 employees
- Single site: in-house ERP PO at 0.5 FTE or more + managed services for level 2 and changes. A fractional CIO only if the internal IT department does not cover governance.
- Multi-site: PO full-time or nearly so + key users per site + managed services with extended coverage. The change committee must include a representative from each site, otherwise headquarters arbitrates alone and the sites work around the system.
- Why: business knowledge and arbitration become the limiting factor; outsourcing them costs more than internalising them.
300 to 500 employees
- Recommended: a small in-house ERP team (PO + a functional administrator) + managed services for capacity and on-call. A light centre of excellence, as described in the article on post-deployment governance.
- Why: the volume of requests and number of processes justify an internal pair, which also reduces key-person risk.
- Watch point: avoid letting the provider become the sole holder of technical knowledge.
Situations that call for a change of model
- A major version upgrade or a cloud migration: temporarily strengthen governance (fractional CIO or project reinforcement).
- External growth (acquisition, new site): move from a part-time PO to a dedicated PO.
- A backlog that durably exceeds processing capacity: review the provider or the size of the internal team.
90-Day Setup Plan
This plan assumes go-live has happened or is imminent. It complements the 90-day stabilisation checklist, which focuses on technology and adoption.
Days 1 to 30: scope
- Map who does what today, in reality: tickets, arbitration, budget, access rights.
- Appoint the owner (PO or fractional CIO) with a mission letter and an explicit time allowance.
- Collect from the integrator the documentation, the access rights and the list of custom developments. See knowledge transfer.
- Open the single backlog and load the pending requests into it.
Days 31 to 60: contract and equip
- Choose the model and any provider; negotiate the exit clauses (see above) and the SLAs following the maintenance contract guide.
- Draft the RACI and have leadership validate it.
- Appoint key users by domain and allocate them time.
- Define the indicators and their data source (ticketing tool, usage logs).
Days 61 to 90: make it run
- Hold a first full cycle: weekly ticket review, first change committee, first dashboard.
- Measure the gap between the planned model and reality: tickets with no owner, delays, the PO’s actual workload.
- Adjust: allocated time, provider scope, committee frequency.
- Present leadership with a review and the roadmap for the next twelve months.
At the end of the 90 days, you should be able to answer one question without hesitation: if the person who runs the ERP leaves tomorrow, what happens? If the answer is “we don’t know”, governance is not finished.
Going Further
Is your run organisation right for the size and complexity of your ERP? We carry out ERP organisation audits: role mapping, analysis of actual workload, model recommendation and transition plan. Contact us for a 30-minute conversation, or subscribe to the newsletter to receive our upcoming guides on ERP run.