The Services Provider License Agreement lets a service provider license Microsoft software to host applications and services for third parties on a pay as you go basis. There is no purchase and no perpetual ownership. The provider reports usage monthly and pays for what was deployed. The model rewards accurate measurement and punishes the provider who reports defensively, reports on outdated SKUs, or never reconciles deployed against reported. SPLA is metered licensing, and the meter favors whoever reads it most carefully.
SPLA is a monthly usage agreement signed between a service provider and Microsoft, almost always through a SPLA reseller. The provider deploys software to serve external customers, measures usage each month, and reports and pays through the reseller. The mechanics turn on what gets counted, which SKU it gets counted as, and how the provider proves the count if Microsoft asks.
There is no upfront purchase and no perpetual license. Each month the provider reports the peak usage of each licensed product and pays for that usage at the agreed rate. The model converts software cost into a monthly variable expense that tracks the service business. Accurate monthly measurement is the whole discipline, because the provider pays exactly what it reports and Microsoft expects the report to match the deployment.
Most SPLA products license either per subscriber, through a Subscriber Access License, or per physical core on the host. The choice of metric for a given workload can swing the monthly cost substantially. A workload with few users on dense hardware may cost far less on a subscriber metric, while a workload with many users on light hardware may favor core based licensing. Choosing the metric deliberately for each workload is a primary cost lever.
SPLA carries strict rules on shared versus dedicated hardware and on which products may run in multitenant environments. The rules determine how the provider may architect its hosting platform and how usage must be counted on shared infrastructure. Misreading the dedication rules produces either an architecture that overpays or a deployment that is technically out of compliance. The rules reward a provider who designs the platform with the licensing model in mind.
Providers report through a SPLA reseller that aggregates reports, sets margin, and is the provider primary point of contact with the program. The reseller relationship shapes the monthly rate, the support quality, and the audit posture. Providers who never test the reseller against alternatives accept whatever margin and rate the incumbent applies, often for years.
SPLA leverage lives in measurement accuracy, metric selection, and architecture. Because the provider pays exactly what it reports, every improvement in how usage is counted flows straight to the monthly bill. The leverage is recurring rather than one time, which makes it compound.
Providers who report defensively round up to protect against audit risk and pay the rounding every month. Precise measurement built from real telemetry rather than worst case assumption removes the defensive premium while keeping the report fully defensible. Accuracy is the largest recurring lever in the program.
Choosing subscriber or core licensing per workload rather than applying one metric across the platform matches the cost to the shape of each deployment. The workload audit that surfaces which metric wins for each product often reduces the monthly bill without changing the service at all.
Designing the hosting platform around the dedication and multitenancy rules can move workloads onto the most favorable licensing footprint. Consolidating onto fewer, more efficiently licensed hosts, or separating workloads that are penalized by shared counting, reshapes the monthly cost structurally.
SPLA price lists and SKUs change, and providers reporting on outdated SKUs may be paying for editions superseded by better priced equivalents. Refreshing the reported SKU set against the current price list captures pricing improvements the provider has not noticed.
The SPLA reseller sets margin on the provider monthly spend. Tendering the SPLA across competing resellers surfaces the margin and the rate, and the recurring nature of SPLA spend makes even a small margin improvement materially valuable across a year.
The single most valuable SPLA discipline is the monthly reconciliation of what was deployed against what was reported. Providers carry retired workloads on the report, double count subscribers across environments, and report development and test usage that may qualify for different treatment. The reconciliation strips usage the provider is paying for but no longer running. Because the saving recurs every month, the reconciliation pays back many times over within a single year, yet most providers have never run it rigorously.
SPLA traps are recurring rather than one time, which makes them more expensive than they appear. A small error in how usage is counted does not produce a single overcharge. It produces an overcharge every month for as long as the error persists.
Providers afraid of an audit report generously to ensure they are never under licensed. The defensive padding becomes a permanent monthly premium that nobody revisits. Precise, telemetry based measurement is fully defensible and removes the padding. The goal is a report that is accurate, not one that is inflated for comfort.
Applying a single licensing metric across every workload ignores that subscriber and core licensing each win for different deployment shapes. The uniform metric overpays on the workloads that suited the other model. The fix is a per workload metric assessment that assigns the cheaper compliant metric to each product.
Retired workloads, decommissioned hosts, and departed subscribers stay on the monthly report because the reporting process is disconnected from the operational reality. The provider pays for usage it no longer runs. A reconciliation that ties the report to live deployment removes the phantom usage and the recurring cost it carries.
The SPLA reseller margin gets set at signing and runs unchallenged for years. The recurring spend compounds the unexamined margin into a large cumulative cost. The discipline is to tender the SPLA periodically and to treat the reseller relationship as a competitive one rather than a fixed cost of doing business.
The most damaging SPLA trap is treating the monthly report as administrative paperwork rather than the legal record of what the provider licensed. SPLA carries audit rights, and Microsoft can review the provider reporting against the actual deployment for prior periods. A provider that has under reported faces back charges across the review period, and a provider that has over reported has been paying a premium that no refund will recover. The report must be both accurate and defensible, supported by measurement evidence the provider can produce on demand. The provider who builds a disciplined, evidenced monthly reporting process protects against the back charge exposure and recovers the over reporting premium at the same time. The provider who treats the report as a form to file carries both risks silently until an audit surfaces them.
We treat SPLA as a measurement and architecture problem first and a commercial problem second. The recurring nature of the spend means every improvement compounds, so the work concentrates on getting the monthly report accurate, defensible, and minimal.
We start by reconciling reported usage against actual deployment. Using the provider own telemetry and infrastructure records we rebuild what is genuinely running and compare it to what has been reported and paid. The reconciliation surfaces phantom usage from retired workloads, double counted subscribers, and defensive padding. Because each unit of over reporting recurs monthly, the reconciliation produces savings that compound across the year.
We then assess the licensing metric per workload. Subscriber and core licensing each win for different deployment shapes, and applying one metric uniformly overpays on the workloads that suited the other. We model both metrics for each product against the real deployment and assign the cheaper compliant metric to each. The reassignment reduces the monthly bill without any change to the service delivered.
We review the hosting architecture against the dedication and multitenancy rules. Where the platform design forces unfavorable counting, we identify the consolidation or separation that moves workloads onto a more efficient licensing footprint. Architecture changes are the structural lever, and on a recurring cost model they pay back quickly.
We refresh the reported SKU set against the current price list and tender the reseller relationship. Outdated SKUs and stale reseller margin are both quiet recurring costs that the provider rarely revisits. Bringing the SKUs current and testing the reseller against the market captures pricing the provider has been leaving on the table.
Throughout, we build the evidence trail that makes the monthly report defensible. The accurate report is only valuable if it can be proven, so the measurement process produces the documentation an audit would require. Our buyer side independence keeps the work aligned with the provider economics rather than any reseller margin. The same independence underwrites our EA renewal work for providers that also run a corporate Microsoft estate.
Anonymized but verifiable on reference call. Drawn from active engagements in the trailing twelve months.
The provider had reported defensively for years, applied a single metric across mixed workloads, and carried retired tenants on the monthly report. We reconciled reported against deployed, reassigned metrics per workload, refreshed the SKU set, and built the evidence trail to make the leaner report fully defensible. The recurring saving compounded across the contract year.
We had padded the report so heavily out of audit fear that we were the ones overpaying. The evidenced report is both cheaper and safer.Chief Operating Officer · Managed hosting provider
Two analyst calls. No pitch. We tell you what we would do, what the leverage actually is, and whether we are the right firm for this engagement.