Retrospective context. The original 6 January 2007 date is retained. The 2026 section is later analysis, not contemporaneous 2007 evidence.
If you sat on an ERP selection committee in January 2007, your shortlist had already been written for you. SAP or Oracle. Everything else on the long list was either too small to survive a board conversation, or had recently been bought by Oracle. That was the defining feature of SAP vs Oracle in 2007: it was not a comparison between two products so much as a choice between two theories of how enterprise software should be assembled. One company built a suite and extended it. The other bought the market and promised to integrate it later. Nineteen years on, both companies are still here, most of those implementations are still running, and the selection criteria used at the time turn out to have missed the thing that actually mattered.
Oracle Had Bought the Alternatives
By 2007, Oracle's acquisition programme had removed most of the independent competition from the enterprise applications market. PeopleSoft had fallen in an eighteen-month hostile campaign concluded in January 2005 for more than $10 billion, taking JD Edwards with it. Siebel followed in 2006. Retek, i-flex and a long tail of vertical specialists went the same way. 2007 continued the pattern: Hyperion, the dominant financial consolidation and planning vendor, was acquired in the spring for roughly $3.3 billion, and Agile Software's product lifecycle management business followed in the summer. Before the year was out Oracle had also moved on BEA Systems. For buyers this created a specific and underappreciated problem. Products that had been competing independently were now stablemates, each with its own roadmap, its own installed base, and an Oracle commitment — branded Applications Unlimited — to keep supporting them. Meanwhile Fusion Applications, the next-generation suite that would supposedly unify everything, existed mainly as slideware. Selecting Oracle in 2007 meant selecting a product line and trusting the convergence story.
SAP's Answer: Depth, Ecosystem and a Bet on Analytics
SAP's counter-positioning was continuity. One codebase lineage, deep industry configuration, the largest implementation partner ecosystem in the market, and a customer base that had already absorbed the pain of R/3. But 2007 was also the year SAP made two moves that told you where it thought the market was going. In September it launched Business ByDesign, a hosted, subscription-priced suite aimed squarely at mid-market companies that would never buy a traditional SAP licence. The product underdelivered commercially for years, but the strategic read was correct: the mid-market was going to be won by software delivered as a service. Then in October, SAP announced the acquisition of Business Objects for approximately $6.8 billion — just over $59 a share, a roughly 20% premium, and by a wide margin the largest deal in SAP's history at the time. The target had more than 40,000 customers and had spent 2007 doing its own consolidating. The message to buyers was blunt: reporting and planning would no longer be the reason to bring in a second vendor.
The Lawsuit Running Underneath Every Sales Cycle
The commercial rivalry turned genuinely hostile in March 2007, when Oracle sued SAP over TomorrowNow, a third-party support subsidiary SAP had acquired to service Oracle's newly acquired customers at a discount. Oracle alleged systematic unauthorised downloading of its support materials. SAP eventually shut TomorrowNow down, and the litigation ran for years — a jury award north of a billion dollars, subsequently reduced, appealed, and finally settled at a fraction of the original figure. For buyers at the time, the relevant detail was not the damages. It was the confirmation that third-party support of acquired ERP products was going to be a legal battlefield, which removed a genuine option for holding maintenance costs down.
What the Licence Actually Cost
Enterprise licence cost in 2007 followed a model that is still with us in modified form: a perpetual licence priced on users and modules, negotiated hard at quarter end, plus annual maintenance in the high teens to low twenties as a percentage of list price. Two things about that structure caused more damage than the headline number:
- Maintenance was calculated on list price, not the discounted price paid. A 60% discount on licences did not produce a 60% discount on the following decade of support.
- Maintenance rates were not fixed forever. SAP's later attempt to move its base to a higher-priced enterprise support tier triggered enough customer-group resistance that the company had to moderate it — a useful reminder that the recurring line was always the negotiable one, in both directions. Over a ten-year horizon, maintenance typically exceeded the original licence spend. Committees that modelled three years got the ranking wrong.
The Risk Nobody Priced Properly
The selection debate consumed months. The implementation decided the outcome. The following years produced a well-documented set of failures that had little to do with which vendor was chosen: Waste Management's lawsuit against SAP over a revenue management implementation, filed in 2008 and later settled; a US jewellery retailer citing its ERP rollout as a contributing factor in a 2009 bankruptcy filing; and a long list of quieter overruns that never reached a courtroom. The common factors in those failures are consistent and boring. Scope defined by the software rather than the business. Data migration discovered late. Process redesign delegated to consultants who would leave. Executive sponsorship that evaporated after the contract was signed. Testing compressed to protect a go-live date chosen for a board meeting. Vendor choice explains a small fraction of ERP outcomes. Governance explains most of it.
Running the Same Decision in 2026
The two-horse race persists in modified form: Oracle sells Fusion Cloud ERP for large enterprises and NetSuite below that, while SAP sells S/4HANA under the RISE commercial model with a hard clock on legacy ECC maintenance that has focused a great many minds. What has genuinely changed is that the mid-market now has credible alternatives — the category Business ByDesign was aimed at in 2007 is now well served — and that subscription pricing has replaced the perpetual-licence-plus-maintenance structure with something more transparent and, over a long enough horizon, not obviously cheaper. If you are running the decision now, the useful questions are the ones the 2007 committees mostly skipped:
- What is the ten-year total cost, including subscription uplifts, integration, and the internal headcount to run it?
- How many of our differentiating processes will the standard product support without modification, honestly assessed?
- Who on our side owns the outcome, and will they still be here at go-live?
- What is the exit cost if this vendor relationship deteriorates?
- Are we buying software to fix a process problem that software will not fix?
| Selection concern | Question to carry into the review |
|---|---|
| Long-term cost | What is the total cost including uplifts, integration and internal operations? |
| Business fit | Which differentiating processes fit the standard product without modification? |
| Accountability | Who owns the outcome through go-live? |
| Exit | What would leaving the vendor relationship require? |
| Problem definition | Is the underlying issue a process problem rather than a software gap? |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
Common Questions
Was SAP or Oracle the better choice in 2007?
For most large enterprises, neither choice was decisive. Both platforms could support the required processes. Outcomes varied far more by implementation governance, data quality and change management than by vendor.
Why did Oracle acquire so many application vendors?
To buy installed base and maintenance revenue, and to deny that base to competitors. Maintenance streams from acquired products were highly profitable and predictable, which funded further acquisitions.
Is the mid-market still forced to choose between tier-one vendors?
No. Cloud suites aimed specifically at mid-sized companies now offer functional depth that was unavailable in 2007, at implementation timelines measured in months rather than years.
What is the most common ERP selection mistake today?
Modelling licence or subscription cost over three years instead of total cost of ownership over ten, and treating implementation capability as a procurement detail rather than the primary risk.
ERP Vendor Selection Review — Outpace offers selection reviews for mid-market and multi-entity groups: ten-year cost modelling, fit-gap against your actual processes, and an honest read on whether your organization can absorb the implementation. Vendor relationships, referral or reseller terms and compensation must be confirmed for the engagement; no unverified independence or no-commission guarantee is made here.