OCI Managed Services Pricing: Retainer Models Explained

Managed services pricing is where OCI buyers most often sign something they do not fully understand. A project quote is at least bounded: a scope, a price, an end date. A retainer is open ended by design, and the pricing model you accept on day one determines how the relationship behaves for years, how the provider earns, what they are motivated to do, and what happens to your bill when your estate grows, shrinks, or has a bad month. The model matters more than the number, because the number changes once and the model shapes every invoice after it.

This article is part of our complete guide to hiring an OCI partner, and it takes one slice of that decision, how ongoing management of an OCI estate is priced, and goes deep. If you are still weighing whether to outsource operations at all, start with in house vs outsourced OCI and come back here once the direction is set.

The five pricing models in use in 2026

Almost every OCI managed services contract on the market is built on one of five models, or a hybrid of two of them. Each one allocates risk differently, and each one creates a different incentive for the provider.

Percentage of cloud spend

The provider charges a monthly fee calculated as a percentage of your OCI consumption, typically somewhere between 8 and 20 percent depending on estate complexity and service depth. It is simple to administer and scales automatically with the estate, which is why large providers like it. The structural problem is the incentive: the provider earns more when your cloud bill grows, which is precisely the opposite of what you want from the team advising you on architecture and capacity. A percentage of spend provider who finds a way to cut your bill by 40 percent has just cut its own revenue by 40 percent. Some providers manage that conflict honestly. The model itself does not.

Tiered flat retainer

A fixed monthly fee for a defined service tier covering a defined estate. This is the most common model among specialist firms and the one we use for our own managed monthly engagements. It is predictable for both sides, easy to budget, and neutral on the size of your cloud bill, the provider has no stake in your consumption going up. The discipline it requires is a clear estate definition and a change control process, because the fee is built on assumptions about how much there is to manage. When the estate changes materially, the tier should change with it, in either direction.

Per resource and per OCPU pricing

A unit price per managed item: per database, per compute instance, per OCPU, per environment. This model is transparent and granular, and it suits estates that change shape frequently, because the bill tracks the inventory automatically. The weaknesses are administrative overhead, since someone has to keep the inventory honest, and the temptation it creates to leave marginal resources unmanaged to save units, which is how shadow infrastructure is born. It also says nothing about service depth, a cheaply managed database and a properly managed one can carry the same unit price on paper.

Time bank and prepaid hours

You buy a block of hours each month or quarter at a discounted rate and draw against it for whatever comes up. This is flexible and feels low risk, but it is not really managed services, it is discounted reactive consulting. Nobody owns outcomes, nobody is watching the estate at 3 a.m. unless an hour is being burned, and unused hours either expire, which means waste, or roll over indefinitely, which means deferred work piling up. A time bank is a reasonable supplement to a retainer for project overflow. As the primary model for production operations it is a warning sign.

Outcome based pricing

Fees tied to measured results: uptime achieved, incidents resolved within target, or, most powerfully, cost savings delivered. The purest version is the optimization model, where the fee is a percentage of verified savings and nothing is owed if nothing is saved. Across our optimization engagements the average outcome is a 40 percent reduction in OCI spend, and the model works because verification is contractual, the savings are measured against an agreed baseline before any fee exists. Outcome pricing rarely covers a whole operations contract, but it is an excellent component, and the mechanics are covered in full in OCI cost optimization fees.

What a retainer should actually include

The model sets how you pay. The inclusions set what you get, and this is where cheap quotes earn their price. A managed services retainer for a production OCI estate should include, as standard and not as paid extras, the following.

• 24/7/365 monitoring and alerting with a human response path, not just a dashboard you can log into. If the provider only watches during business hours, you do not have managed services, you have a help desk.
• Patching and lifecycle management across the OS, database, and platform layers, on a published cadence, with maintenance windows agreed in advance.
• Incident response with defined severity levels, response and restoration targets, and an escalation chain you have tested, not just read about in the proposal.
• Monthly cost review of the estate against budget, with rightsizing and commitment recommendations. A provider that manages your OCI estate but never mentions your bill is leaving the easiest value on the table.
• Security posture management, including Cloud Guard findings triage, identity and access reviews, and vulnerability tracking against an agreed baseline.
• Quarterly architecture review, a structured session where the provider tells you what should change, what risk is accumulating, and what the next quarter of improvements looks like.
• Documentation and runbooks kept current, owned by you contractually, so the knowledge does not walk out the door if the provider does.

Anything on that list that appears in the quote as a chargeable extra should be treated as a price increase in disguise. The most common stripped items are the cost reviews and the quarterly architecture sessions, which is telling, because those are the two inclusions that most often shrink your spend rather than grow the provider's.

How tiers are typically structured

Most providers package the inclusions above into three tiers, whatever they choose to call them. The pattern is consistent enough across the market that you can map any provider's brochure onto it in a minute.

The entry tier is monitoring and incident response only: the provider watches, alerts, and reacts, but improvement work, patching beyond critical fixes, and advisory time are excluded or metered. It suits nonproduction estates and organisations with a capable internal team that just needs overnight coverage. The standard tier adds proactive operations: full patching cadence, cost reviews, security posture work, and a defined allowance of change and improvement effort each month. This is where most production estates belong. The comprehensive tier adds ownership: the provider runs the estate end to end, drives the improvement roadmap, attends your planning, handles audits and compliance evidence, and carries tighter response targets with financial teeth. Pricing steps between tiers are commonly 1.5x to 2x per step, and the right tier depends less on the size of the estate than on how much of the operational burden you genuinely want to transfer. We break the inclusion patterns down further in OCI support tiers explained.

The commercial fine print that decides whether the price is real

The monthly fee is the visible price. Four other terms decide what you actually pay over the life of the contract.

Onboarding fees

Most providers charge a one time onboarding fee, often one to three times the monthly retainer, to cover discovery, documentation, tooling deployment, and runbook creation. The fee itself is legitimate, onboarding is real work, but ask what you own at the end of it. If the runbooks, monitoring configuration, and documentation belong to the provider, the onboarding fee has quietly purchased your own switching cost.

SLA credits

Service level targets without consequences are marketing. Real contracts attach credits, a percentage of the monthly fee returned when response or restoration targets are missed, escalating with repeated failures and with a termination right if performance stays below target for consecutive quarters. Credits will never compensate you for a real outage, their purpose is to make failure expensive enough that the provider staffs properly. How to negotiate targets and credits that actually bind is the subject of OCI SLA negotiation.

Scope creep and change control

Estates grow. New databases appear, a project leaves behind three new compartments, and eighteen months in, the provider is managing 40 percent more than the contract priced. Good contracts handle this with a defined estate baseline, a simple mechanism for adding and removing items at published unit prices, and a scheduled true up. Bad contracts handle it with a renegotiation in which the incumbent holds all the leverage. Insist on the mechanism up front, and insist it works in both directions, removal of workloads should reduce the fee as automatically as additions increase it.

Contract length and exit terms

Twelve month terms with 90 day notice are the market norm. Three year terms earn a discount, typically 10 to 15 percent, but only accept one alongside strong exit provisions: a defined offboarding process, knowledge transfer obligations, your ownership of all documentation and configuration, and no fees for handing over what is already yours. The quality of a provider is most visible in how they behave when you leave, and the time to secure that behaviour is before you arrive.

Hybrids are common and often sensible. A flat retainer for the defined estate plus a small time bank for project overflow covers most realities without distorting incentives. A per OCPU schedule bolted onto a flat base handles seasonal scaling cleanly. The combination to avoid is percentage of spend pricing paired with architecture authority, because that hands the provider both the pen that draws your estate and a commission on how large it becomes. If a provider insists on percentage of spend, fence it: cap the percentage band, exclude committed spend you negotiated yourself, and put cost reduction targets in the contract so the incentive at least has a counterweight.

A framework for evaluating a managed services quote

What good looks like

The healthiest managed services arrangement we know is a flat monthly retainer at a clearly defined tier, an estate baseline with working change control, SLA credits with real teeth, and an optimization layer priced on verified savings sitting alongside it. That combination gives the provider a stable, predictable revenue line and gives you a partner whose only path to earning more is delivering more, either by taking on a genuinely bigger estate or by finding savings that are measured before they are billed. It is how our own OCI managed services practice is structured, backed by 24/7/365 monitoring, 500+ OCI engagements of accumulated runbook knowledge, and 20+ years of combined Oracle experience, and with no Oracle resale margin anywhere in the chain, because we are independent, not a reseller, and we earn nothing when your consumption grows.

Pricing models are not paperwork. They are the operating system of the relationship, and they quietly decide whether your provider profits from your efficiency or from your sprawl. Take the week it costs to understand the model, the inclusions, and the exit terms before you sign, and the retainer becomes what it should be: the cheapest insurance your OCI estate will ever carry.