The perpetual licence is not being withdrawn so much as quietly starved. It still appears on price lists, it can still be bought from most vendors if you insist, and the sales team will still process it. What has changed is everything around it: the new functionality lands in the subscription product first, the implementation partners are trained on the subscription product, the roadmap slides show the subscription product, and the discount available on a perpetual deal has become noticeably less generous than the one available on a multi-year subscription. ERP subscription licensing has won by making the alternative unattractive rather than unavailable. Most buyers have accepted this as an accounting change, capital expenditure becoming operating expenditure, and negotiated accordingly. That framing misses the part that matters. Subscription does not primarily change how you pay. It changes who holds leverage, and when.
Under perpetual, the negotiation happened once
The old model had an ugly shape and one redeeming feature. You paid a large sum at the start, you were over-sold, you deployed perhaps sixty per cent of what you bought, and you then paid maintenance on the full amount forever. It was wasteful and everybody knew it. The redeeming feature was that after signing, the vendor had limited power over you. You owned a right to use the software indefinitely. If the relationship deteriorated, you could stop paying maintenance and keep running, unsupported and unpatched but operational, for years. That option was rarely exercised, but its existence disciplined every subsequent conversation. Subscription removes it. The software stops when payment stops. This is not a criticism of the model, it is simply its defining characteristic, and it means the negotiation is no longer a single event at purchase. It recurs at every renewal, for as long as the system runs your business, and your position at each of those renewals is weaker than it was at the first one, because by then the system holds your data, your processes and your integrations.
There are three prices, and buyers negotiate one
The entry price gets all the attention, and it is the one the vendor is most willing to concede, because a discount on year one costs them very little against a contract they expect to renew for a decade. The renewal price is the one that determines what you actually pay. Without contractual protection, renewal uplift is set by the vendor's circumstances rather than yours, and it arrives at the precise moment when switching is least practical. Price protection beyond the initial term, expressed as a cap on annual increase, is the single most valuable clause in an ERP subscription agreement and the one most often left out. The exit price is the one nobody prices at all. What does it cost to leave, and can you? Specifically: can you extract your data in a usable structured form, including historical transactions and attachments, not just a report; how long do you have to do it after termination; and what does the vendor charge for the extract. A contract that is silent on this is a contract where the answer will be decided when you are least able to argue.
| Price | Question before signing |
|---|---|
| Entry | Which users and modules are actually required? |
| Renewal | What uplift, growth and reallocation rights are agreed? |
| Exit | What data, format, window and extraction cost are provided? |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
What to put in the contract
A cap on renewal increases, for the life of the relationship, not just the first term. Anything else is an option written in the vendor's favour. Growth ramps that reflect your plan. If you expect to add entities or users, price them now. Buying incremental seats later at list is how subscription costs run away. Definitions of user types that survive contact with reality. Full users, limited users, self-service, read-only, external. The classification argument is where audits are won and lost, and the definitions in the standard agreement are usually written loosely enough to be reinterpreted. An explicit position on indirect and digital access. Machines, portals and integrated applications that touch the system have been the subject of enough litigation over the past two years that the vendor's current licensing document should be attached to the contract and its interpretation agreed in writing. Swap rights between modules. You will buy something you do not use. The ability to reallocate that spend at renewal, rather than only add to it, is worth real money. Data extraction terms. Format, scope, timeframe, cost, and a commitment that the export is machine-readable.
The shelfware problem did not go away
A popular argument for subscription is that it ends shelfware. It does not. It converts shelfware from a sunk cost you stopped noticing into an annual charge you keep paying, which is better only if somebody actually reviews it. That review is the highest-return licensing activity available to most organisations and it takes a day. Count active users in the last ninety days, by licence type, against what you are paying for. In an organisation with five hundred subscribed users, the gap is routinely ten to twenty per cent, and under subscription that gap is recoverable at renewal in a way it never was under perpetual.
Practical Guidance for an ERP Licensing Strategy Consultation
- Negotiate renewal caps before you negotiate the entry discount. Trade the headline discount for the cap if you must; the cap is worth more over five years.
- Run a usage audit ninety days before every renewal. Active users by type, module usage, dormant accounts. Go into the conversation with your own numbers.
- Model the ten-year cost of both models honestly. Include the upgrade projects that perpetual requires and the price escalation that subscription permits. The answer is not always subscription.
- Get indirect and digital access in writing. Attach the vendor's current definition and record your agreed interpretation for your specific integrations.
- Secure data extraction rights explicitly. Structured format, full history, defined window, stated cost. Test the export once during the term rather than discovering its limitations at exit.
- Watch the mandatory upgrade clause. Subscription usually removes your right to stay on a version. That transfers regression testing effort to you, on the vendor's schedule, and it belongs in your resourcing plan.
- Keep the contract in the name of an entity that can actually enforce it. Regional buyers frequently sign through a reseller, which changes who you can hold to the service commitments.
- Diarise the notice date. Auto-renewal clauses with ninety-day notice periods have cost more organisations more money than any licensing dispute.
The Regional Angle
Subscription pricing interacts awkwardly with how regional businesses actually approve spending. Capital purchases here are often decided once by an owner or chairman and then treated as done, whereas recurring operating costs face annual scrutiny in a way capital does not. The practical consequence is that finance leaders in Gulf family groups find themselves re-justifying the ERP every single year to someone who believed they had bought it in 2017. That is an internal communication problem rather than a commercial one, but it consumes real management time and it is worth pre-empting by presenting the subscription as a five-year committed cost rather than an annual discretionary one. The purchasing route matters more here than in most markets. Regional buyers typically contract through a local reseller or implementation partner who bundles licences, implementation and support into a single agreement. That arrangement has genuine advantages, local support and a single invoice, but it obscures the licence economics, makes benchmarking difficult, and means the renewal conversation happens with the party who also holds your configuration knowledge. Splitting the licence agreement from the services agreement, even when both are with the same partner, restores a meaningful amount of leverage. Tax treatment is now a live consideration in its own right. Since value added tax arrived across the Emirates and Saudi Arabia, cross-border software subscriptions are within scope, generally handled by the recipient through the reverse charge mechanism. For a fully registered and recoverable business that is an administrative step. For entities with restricted recovery, or group structures where the contracting entity is not the consuming entity, it is a real cost and a real compliance exposure. Getting the contracting entity right at signature is far cheaper than restructuring it later. Currency is a mixed picture. For UAE, Saudi, Qatari and Omani entities the peg to the dollar removes most of the exchange risk that makes dollar-denominated subscriptions uncomfortable elsewhere, which is a genuine regional advantage. Groups with Egyptian, Turkish, Pakistani or sub-Saharan operations do not have that comfort, and a subscription priced in dollars for an entity earning in a floating currency is an unhedged escalating liability that belongs in the treasury conversation, not just the IT one. One final point specific to the moment. Subscription generally means hosted, and hosted currently means outside the Gulf, because the hyperscale regions announced for the UAE and Bahrain are not yet serving production workloads. Organisations in banking, insurance, healthcare and government should settle the hosting question before the licensing one, because the sector rules on where regulated data may sit can make the subscription product unavailable to them regardless of price.
The objection worth taking seriously
The objection is blunt: subscription is a price rise wearing a business model. Over any reasonable life, a subscription costs more than a perpetual licence plus maintenance, which is precisely why vendors have pushed it so hard, and the benefits offered in exchange, continuous updates, lower entry cost, predictable spend, are benefits the vendor also wanted. Analyst estimates of the crossover point vary, but on most sensible assumptions the subscription is more expensive somewhere between year five and year eight, and enterprise systems routinely run for fifteen. The harder version concerns dependency. Under perpetual you owned a deteriorating asset; under subscription you own nothing at all, and a business whose finance, supply chain and payroll run on rented software has handed a third party the ability to switch off its operations for non-payment. Combine that with mandatory upgrades on the vendor's schedule and the loss of version control, and the customer's autonomy has narrowed considerably in exchange for a smoother invoice. This is all true, and the useful response is not to argue with it but to price it. If subscription costs more over the life, then the contract has to buy back some of the difference: renewal caps, swap rights, extraction terms, upgrade notice periods. If the dependency is real, then it should be managed the way other critical dependencies are, with an exit assessment reviewed annually and a documented, if unattractive, alternative. What does not work is signing a standard agreement, treating the annual charge as a utility bill, and discovering the terms of the relationship during the renewal where you have no options. The model is not going to reverse. The negotiating posture is the only variable still under the customer's control.
Common Questions
Can we still buy perpetual licences?
From most vendors, yes, though the discount structure, the roadmap and the partner ecosystem all now favour subscription. The relevant question is whether new capability will reach the on-premise product, and increasingly the answer is later or not at all.
Is subscription cheaper?
Cheaper to start, generally more expensive over a long life. It also shifts upgrade cost from periodic projects to the vendor, which is a genuine saving that perpetual comparisons often ignore. Model both over ten years with realistic escalation rather than accepting either vendor's arithmetic.
What is the most commonly missed clause?
Renewal price protection beyond the initial term, followed closely by data extraction rights. Both are cheap to obtain at signature and effectively unobtainable afterwards.
What should we expect over the next twelve months?
Expect vendors to keep concentrating new functionality in cloud subscription editions and to price on-premise renewals less attractively. Expect the 2025 maintenance horizon on the previous large-suite generation to be used as a commercial lever in every enterprise negotiation from here. Expect continued clarification and continued disputes around indirect and digital access as integration counts rise. And expect mid-market vendors to keep pushing multi-year committed subscriptions with escalators, which makes the renewal cap the clause to fight for now rather than later.
ERP Licensing Strategy Consultation — we audit what you actually use, model both licensing models over a realistic life, and negotiate the two clauses that decide what you pay in year six.