The decision to leave Oracle Premier is the first half of the journey. The decision of which third party support vendor to engage is the harder half. The market has consolidated around ten serious providers, each with different coverage strengths, pricing posture, account team depth and contract flexibility. The wrong selection produces hidden costs that exceed the year one saving. The right selection produces a five year arrangement that holds quality, controls cost and protects the underlying license estate.
This is the complete pillar guide to choosing a TPS vendor for an Oracle estate. It is drawn from 200 plus buyer side advisory engagements and is the field summary of the TPS Vendor Selection Guide white paper. The framework is straightforward. Define the requirement against the estate. Invite three to five qualified vendors. Score against weighted criteria. Test claims with references. Negotiate twelve specific contract clauses. Close in eight to twelve weeks.
The market and the ten vendors that matter
The third party support market has consolidated around ten providers. Two are at scale, with thousands of customers and global delivery reach. Several mid market providers serve specific Oracle product sets with strong coverage. Several boutiques specialise in narrower niches such as Hyperion, Agile PLM or Oracle Industry Global Business Unit products. Each has a different commercial posture. Each has different bench strength by Oracle product. Each prices differently for the same nominal scope.
A serious selection considers all ten. A common buyer side error is to invite only the two market leaders and conclude after a price exchange. That process produces a price but not the best price and rarely the best fit. Rimini Street and Spinnaker Support are the obvious starting points, but the broader pool typically produces stronger competitive tension and better terms. The selection framework starts with mapping the estate to the vendor coverage matrix rather than starting with brand recognition.
Defining the requirement against the estate
The first task is an inventory of what actually needs to be supported. This is not the Oracle CSI master list. It is the operating reality. Which products are in active production. Which are in steady state with light change. Which are scheduled for retirement inside the next 36 months. Which are running customisations that the buyer needs the vendor to support. Which countries are represented in the user base for tax and regulatory content. Which database options and middleware components are in use.
The inventory drives scoping. A vendor cannot price intelligently without knowing whether the buyer needs payroll tax updates for 14 countries or two, whether the GoldenGate replication topology counts as one environment or five, whether the EBS instance includes the GL Localisation pack, whether the Hyperion estate runs Essbase ASO or BSO. The buyer who delivers a complete inventory at RFP time gets sharper pricing. The buyer who delivers a vague inventory gets padded pricing or finds change orders in year two. The pre RFP inventory is the single highest ROI activity in the selection process.
The twelve weighted selection criteria
Across 200 plus advised selections the criteria distribute consistently. Coverage depth by Oracle product carries 18 to 22 percent. References from comparable estates carry 12 to 15 percent. Account team experience and continuity carry 10 to 12 percent. Contract flexibility (term, exit, scope change) carries 10 percent. Tax and regulatory content (country coverage, update cadence, lead time) carries 8 to 10 percent. Security posture (SOC 2, ISO 27001, data residency) carries 6 to 8 percent. Financial stability of the vendor carries 5 to 7 percent.
Price carries 25 to 35 percent of the total scoring weight, not 50 percent. This single shift in weighting is the largest determinant of selection outcome. Buyers who weight price at 50 percent default to the cheapest qualified bid. Buyers who weight price at 30 percent and total cost of ownership including coverage risk at the remaining 70 percent select on durability. Across the advised portfolio the second approach produces materially better five year experience and lower total cost despite a higher headline price.
Across 200 plus advised selections, lowest price wins only 14 percent of decisions. The other 86 percent of buyers choose a higher priced vendor on the strength of coverage, references or contract terms.
Reference calls that produce real signal
Vendor reference lists are curated. The references on the list will say positive things about the vendor. The value of the call is not the headline endorsement but the texture beneath it. The questions that produce signal are about specific incidents. Tell me about the most difficult ticket your team has opened with this vendor in the last twelve months. How was it escalated. Who from the vendor was on the call. How long until first engineering response. Was the root cause documented in writing. Were the recommendations actionable.
The references that matter most are not on the list. Two to three back channel calls into the buyer side network typically produce the most honest assessment. Industry peers who have moved away from a vendor are particularly valuable. They explain why the relationship ended and whether the issue was structural or specific. The combination of three on list and two off list calls produces a fair picture. The vendor selection advisory service runs the structured reference process on behalf of the buyer.
The RFP document that gets sharp responses
A weak RFP document gets weak responses. A strong RFP document gets sharp, comparable, scoreable responses. The structure that works is eight sections. Background and estate inventory. Scope of services requested. Coverage and SLA expectations. Tax and regulatory content requirements. Security and compliance requirements. Commercial structure and pricing instructions. Contract clause expectations. Vendor question template with no narrative latitude.
The narrative latitude question is the most important. Vendors who can respond in flowing prose will pad their response with marketing language. Vendors who must respond in a structured question and answer template will give comparable answers that score cleanly. The template should ask for specific bench numbers, specific named engineers, specific past incidents, specific service levels. The buyer side advisor builds these templates from a market benchmark. The quote comparison service uses the same template across the shortlist for direct comparison.
The twelve contract clauses that determine the five year position
Twelve clauses determine whether the contracted scope holds across the five year horizon. Price protection capped at one to two percent. Scope change definition including the addition of products and users. Exit rights including data return and knowledge transfer. SLA definitions with measurable thresholds. Penalty clauses with real economic teeth. Tax and regulatory content commitments with deadlines. Indemnification scope including intellectual property risk. Insurance and financial responsibility levels. Data security and residency terms. Subcontracting limits. Audit rights for the buyer over the vendor. Term and renewal mechanics.
None of these clauses appears in the vendor's standard contract in the form a buyer wants. Each is negotiated. The negotiation produces meaningfully different five year economic positions for materially the same nominal price. Across the advised portfolio, the gap between a well negotiated and a poorly negotiated contract on the same vendor at the same headline price is 8 to 14 percent of total five year value. The clauses are covered in chapter four of the TPS negotiation framework white paper.
Pricing posture by vendor
The two market leaders price at the higher end of the TPS market. They invest more in account team coverage and global delivery, and the pricing reflects this. The mid market providers (Support Revolution, Origina) typically price 10 to 20 percent below the leaders for nominally similar scope. The boutiques specialise narrowly and price competitively in their niche but lack breadth for multi product estates. The pricing posture is consistent across the market but the gap between opening offer and best and final offer varies by vendor.
Vendors who hold tight on opening price are signalling that they are at the high end of their economic envelope. Vendors who move significantly on best and final are signalling that the opening had margin. Neither pattern is intrinsically better. A vendor with a narrow gap and superior coverage may still be the better five year choice. A vendor with a wide gap and weaker coverage may still be the better one year value play. The trade off is explicit and modellable. The framework is in chapter five of the vendor selection guide.
The selection timeline
A disciplined selection runs eight to twelve weeks. Week one and two run the inventory and the requirement document. Weeks three and four issue the RFP and run the question and answer cycle with each invited vendor. Weeks five and six receive responses and complete initial scoring. Week seven shortlists to two finalists and runs the reference programme. Weeks eight and nine run the oral presentations and the best and final pricing. Weeks ten through twelve run contract negotiation and legal close.
$2.9M Oracle Premier baseline. Five vendor RFP. Shortlist of two after week six. Best and final at $1.54M with one percent escalator and twelve clauses negotiated. Year one net saving 47 percent. Five year cumulative saving $7.3M.
The buyers who run faster than eight weeks compress the negotiation phase and leave value on the table. The buyers who run longer than twelve weeks lose vendor attention and momentum. The pacing matters. The buyer side advisor enforces the cadence on behalf of the customer team.
The selection that holds for five years
The selection is not over at signing. The structures that hold quality across the five year term are quarterly business reviews with measurable SLA tracking, named primary and secondary engineering points of contact with succession plans, an annual mid term review against the contract terms, and a transparent change order process with capped administration cost. Buyers who design these governance structures into the contract get better five year outcomes than buyers who treat governance as a year two afterthought.
The other determining factor is the buyer side relationship owner. A named buyer side owner of the TPS relationship, with quarterly reporting to procurement and IT leadership, keeps the vendor honest. Without a named owner, the relationship drifts and vendors deprioritise the account. Most buyers do not have such an owner internally. The Oracle Database estate is typically the largest line in the support spend and the most operationally critical, so the relationship owner is usually drawn from the database team or the procurement function with database literacy.
Frequently asked questions
How many TPS vendors should be invited to the RFP?
Three to five is the sweet spot. Two is not enough for genuine competitive pressure. Six or more dilutes attention and produces lower quality proposals. Three to five forces each vendor to invest in the response and creates real downward pressure on price.
What weight should price carry in vendor selection?
Across 200 plus advised selections, price typically carries 25 to 35 percent of the scoring weight, not 50 percent. The other 65 to 75 percent is coverage, references, account team, contract terms, financial stability and security posture. Lowest price wins only 14 percent of advised engagements.
Should the incumbent Oracle reseller be invited?
No. Oracle resellers are conflicted by definition. Their commercial relationship with Oracle compromises their ability to advise against Oracle Premier. The TPS RFP should include only independent third party support vendors with no commission relationship with Oracle Corporation.
How long does a TPS vendor selection take?
A disciplined process runs eight to twelve weeks from kickoff to signed contract. RFP issue takes two weeks, vendor response two to three weeks, scoring and shortlist one week, reference calls and orals two weeks, negotiation two to three weeks, legal close one to two weeks.
What is the most common selection mistake?
Choosing on headline price without testing coverage depth. The largest cost in a TPS programme is not the contract price but the cost of inadequate coverage on a critical incident. Vendors with thin engineering benches in a buyer's specific Oracle product set produce hidden costs that dwarf the price differential.