© 2026 Procr
An enterprise licence agreement binds your organisation to a software vendor for years at a time. The pitch is predictable pricing and a meaningful discount. The risk, when the deal is structured poorly, is three to five years of paying for software your team stopped using at month six.
An enterprise licence agreement (ELA) is a multi-year contract between an organisation and a software vendor that grants the right to use one or more products across the entire organisation in exchange for a fixed annual or upfront fee. The defining characteristic is the all-in commitment: unlike per-seat or usage-based contracts that can be resized at renewal, an ELA fixes your spend and often your product scope for the full term.
ELAs are most common with large software vendors, including IBM, Oracle and Microsoft, and with SaaS platforms that have grown large enough to push organisation-wide agreements: Salesforce, ServiceNow, and Workday are common examples. Agreements typically run three to five years and cover an agreed set of products or modules. Some include unlimited deployment rights within the covered entity; others cap usage at a defined number of seats but fix the price for that cap.
The core value proposition is straightforward: you commit to a larger, longer relationship and the vendor gives you a lower unit cost. Discounts for Microsoft Enterprise Agreements typically reach 30 to 50 percent off list prices for covered products. Salesforce ELAs tend to run 20 to 40 percent off list, depending on deal size and how competitive the situation looks to the vendor's sales team. IBM discounts are similarly in the 30 to 50 percent range depending on product scope.
Understanding the vendor's motivation helps you evaluate whether an ELA is right for your organisation or right for theirs.
ELAs are revenue certainty for vendors. A three-year agreement with annual invoicing locks in revenue regardless of whether your usage grows, contracts, or stagnates. For publicly traded vendors managing quarterly targets, this is extremely valuable, which is why procurement teams routinely receive ELA proposals when a vendor's sales rep needs to close before quarter-end.
Vendors also use ELAs to expand product scope. Bundles frequently include products you do not currently use: Power BI added to a Microsoft EA, Revenue Intelligence layered into a Salesforce deal. The incremental cost looks modest at signing. Over three years, those line items compound even when adoption never materialises. The vendor's financial incentives and yours are not perfectly aligned in an ELA negotiation.
The headline discount is real. The total cost of the agreement is often higher than that number suggests.
Maintenance and support fees. ELAs for on-premises and hybrid software frequently include annual maintenance at 16 to 25 percent of the licence fee. On a $500,000 ELA, that adds $80,000 to $125,000 per year in ongoing costs.
Bundled products you do not need. Vendors condition their best discounts on buying a broader product set. This is negotiable. Procurement teams that push back on scope and insist on holding the discount for a reduced bundle frequently succeed, particularly when negotiating at quarter-end. But the default vendor position is to bundle.
The commitment you cannot downsize. Per-seat SaaS contracts can be reduced at renewal. ELAs typically cannot be reduced mid-term. If headcount shrinks, a business unit is divested, or an adoption project stalls, you continue paying for capacity you no longer need until the term expires. Early termination fees in multi-year software agreements typically run 50 to 100 percent of remaining contract value.
Shelfware is software you have paid for but are not using. It is the most common hidden cost in enterprise licence agreements, and it is consistently underestimated at signing.
Research across enterprise software deployments finds that 25 to 40 percent of licensed software goes unused in any given year. For Microsoft E5 bundles specifically, shelfware rates in enterprise deployments run 30 to 50 percent. An additional 45 percent of applications are underutilised, meaning less than half of their licensed capacity sees regular use.
For mid-market organisations, the dynamic is the same: you sign a three-year ELA based on a projected headcount or an ambitious rollout plan, and the deployment runs 12 months behind schedule. By year two, you are paying for 400 seats on a platform where 180 users are active. The unlimited-use model some ELAs offer cuts both ways. Unlimited deployment rights sound attractive when you project rapid growth. If growth slows, you have paid for a ceiling you never approached.
An ELA is a reasonable structure when three conditions are true simultaneously.
Usage is stable and documented. You have run the product for at least 12 months, actual seat count is consistent, and you can model realistic usage for the full contract term based on real data rather than projections.
The switching cost is high. The product is deeply embedded in your workflows. This does not mean you are stuck, but it does affect the calculus when evaluating a long commitment.
The discount offsets the flexibility you are giving up. A 15 percent discount over three years does not compensate for the inability to reduce seats if your team restructures. A 40 percent discount might. Compare the three-year ELA total against three annual per-seat renewals at current volume plus realistic growth, and run the numbers with a downside scenario included.
For most mid-market teams evaluating a first ELA with a vendor, the better default is to negotiate one more annual renewal at improved per-seat terms before committing to a multi-year structure. You preserve downside flexibility and accumulate more data about actual usage patterns.
If you are evaluating an ELA, these terms require close attention before the agreement is finalised.
Scope definition. What products and modules are covered, and which are excluded? What you believe is included should be written into the agreement explicitly.
True-up mechanism. If the ELA is usage-based with annual true-ups, confirm what triggers additional charges, how they are calculated, and whether there are caps on year-over-year increases.
Price escalation on renewal. Multi-year ELAs frequently include built-in annual escalation clauses of 3 to 7 percent. Negotiate a cap before signing.
Termination rights. Most ELAs do not include a termination-for-convenience right during the initial term. Where exit rights exist, they typically become exercisable only after 12 to 24 months, and early termination fees of 50 to 100 percent of remaining contract value are standard. Understand exactly what you are locked into.
Renewal notice window. ELAs auto-renew on terms set in the original agreement. Notice windows to opt out or renegotiate typically run 30 to 120 days before the renewal date. Missing the window means another term on the existing terms.
A standard SaaS subscription is typically annual and per-seat: you pay for active users and can adjust the count at renewal. An ELA locks in both product scope and price for a multi-year term, usually three to five years, in exchange for a material upfront discount. The trade-off is cost certainty against reduced flexibility to right-size your commitment as your needs change.
Discounts vary by vendor and deal size. Microsoft Enterprise Agreements typically offer 30 to 50 percent off list price for covered products. Salesforce ELAs generally run 20 to 40 percent off depending on total contract value and sales quarter timing. IBM discounts run 30 to 50 percent. If a vendor offers less than 20 percent in exchange for a three-year commitment, the economics do not favour the buyer.
Rarely, and not cheaply. Approximately 70 percent of enterprise software contracts include some form of early exit right, but these typically only become exercisable after an initial 12-to-24-month commitment window, and early termination fees generally run 50 to 100 percent of the remaining contract value. Before signing, negotiate either a termination-for-convenience clause with a capped fee or milestone-based reduction rights tied to specific business events such as a headcount reduction above a defined threshold.
Shelfware is licensed software that is not actively used. It is particularly costly inside an ELA because you cannot reduce the contracted scope mid-term. Industry data puts unused software rates at 25 to 40 percent across enterprise deployments; for bundled agreements like Microsoft E5, shelfware rates run 30 to 50 percent. Before committing to a multi-year ELA, audit actual usage data on your current subscriptions to establish a realistic baseline and identify which committed products have genuine adoption.
Avoid an ELA when your usage is growing fast and has not yet reached a stable plateau, when the vendor or product is still unproven in your environment, or when the discount does not meaningfully exceed what you would get by negotiating a better annual rate without the multi-year commitment. Organisational restructuring is also a red flag: headcount reductions or business unit changes that shrink your software needs mid-term will leave you holding a contract sized for a larger organisation.
Procr
See what Procr does with your real vendor portfolio.