CCEP’s transport tender: broader bid coverage is not autonomous buying

External-source research synthesis, not an OfferSignal client result. Based on the retained October 2026 research and source-specific dates.

A large transport tender becomes difficult when one lane may need more than one equipment type, many lines must be represented, and suppliers need to return comparable bids. The CCEP case is useful because it addresses that operational problem without establishing the stronger claim that software negotiated, awarded or paid for transport.

What CCEP used the optimizer for

A Keelvar-hosted customer story and matching 93-second interview, uploaded on 16 January 2023, describe a CCEP sourcing analyst using Sourcing Optimizer for a “large, detailed” transport tender. The analyst says CCEP was trying to handle more than one equipment type on the same lane. The tool enabled CCEP to include “all of our lines,” create and send the tender more easily, and keep supplier activity in one place. “Lines” is not defined, and the public record does not disclose the tender country, business unit, freight mode, number of lanes, equipment types or carriers. The video upload date is not the tender start, close or award date. 1 2

The reported outcome is qualitative rather than financial. The analyst says suppliers had no access problems, gave positive feedback, and that some previously problematic suppliers were submitting bids. There are no invitation, eligibility, completion or bid-return counts, so the account cannot establish a response-rate lift or attribute the change solely to software. No tender value, rate baseline, award, savings, service, emissions, cycle-time, return-on-investment or software-price result is published. 2

Optimization is not award authority

Current Keelvar documentation describes a product class in which suppliers can bid across items, lanes or services using conditional prices, volume discounts and regional constraints. The system evaluates combinations and produces ranked recommendations against buyer-defined objectives; Keelvar says the team makes the call and controls when multi-round requests for quotation close. That is consistent with a human-controlled optimizer or recommender, not evidence of an autonomous CCEP award. The documentation establishes possible capability, not the configuration enabled in this tender. 3

Keelvar’s transport page lists possible dimensions such as cost, transit time, carbon dioxide, carrier reliability, capacity and lane alternatives, plus benchmark imports, spot workflows and conditional awarding. None is proven for CCEP, and newer references such as Kai must not be backdated to January 2023. CCEP’s supplier page says electronic transacting through the Ariba Network is standard after a purchase-order notification; that does not show that this tender used Ariba, that Keelvar awarded it, or that a carrier was paid automatically. 4 6

The analyst’s reference to Autonomous Sourcing was future-oriented: CCEP hoped it could help with tail spend and larger, more challenging events. It is not a reported deployment, pilot, negotiation or saving. A separate IBM procurement-data programme reported more than 98% of direct and indirect spend classified and analysed, and more than US$40 million in benefits including US$5 million in annual cost and avoidance savings. IBM does not mention Keelvar, transport or this tender, so those figures must not be used as C17 savings or return on investment. 5

An illustrative next-step test

For an illustrative manufacturer or distributor running a multi-carrier tender, preserve a versioned record of each lane, equipment or service requirement, supplier identity, bid formula, capacity and service constraint, and approval boundary. Separate eligible, invited and responding suppliers. Record missing or invalid fields, re-bids, scenarios reviewed, overrides and the final human approver. Compare awarded served cost with service, capacity and carbon constraints rather than using a vendor-wide savings percentage.

A useful test would ask whether the system expands valid tender coverage and reduces supplier-access failures while keeping the buyer in control. It should not infer an award, contract, purchase order or payment instruction from a ranked recommendation. The evidence supports optimization-assisted transport procurement, not an autonomous purchasing agent.

References

Sources and next reading

Keelvar Sourcing Optimizer — Public product reading page; current features do not establish CCEP’s 2023 configuration.

Keelvar application login — Official authenticated entry point; no login or trial was used.

CCEP transport-tender customer interview — Public vendor-hosted case account, not an application to bid, negotiate or validate CCEP results.

CCEP Supplier Information — Public supplier-process context; later access was CAPTCHA-blocked.

Explore OfferSignal · Focused Business & AI Workshop · All research