Strategy
Part of A retail media plan for 2027 with evidence gates, named owners and practical stop rules
Ninety days to investigate retail media with a rollback and a hold option at every gate
Sequence a reversible retail-media investigation across ninety planning days, with source gates, synthetic tests, named owners, rollback and a hold option.
This retail media ninety day plan is a sequencing convention, not a normal launch duration or a revenue forecast. It takes one England retailer from a bounded question to an evidence-based proceed, revise, pause or stop decision. Production delivery remains disabled throughout unless a later, separately authorised launch process passes every gate.
Days 1 to 15: define the exchange
Name the retailer or marketplace legal entity, advertiser, agency authority, controlled inventory and affected customer task. Select one on-site, off-site or in-store environment and one approved product scope. IAB UK's retail media definition provides those industry categories, but it does not establish legal authority or likely results.
The commercial owner records the decision, current route and evidence gap. Product and advertising owners create versioned claim and creative registers. Finance leaves media, labour, review, technology, incident and exit costs blank until exact evidence exists. Acceptance at day 15 means the parties, property, exchange and exclusions are unambiguous. Otherwise pause.
Days 16 to 30: map customer and data effects
Trace the advert from brief through approval, delivery, landing or store task, measurement, correction and withdrawal. Record product information, total price, mandatory charge and optional-choice evidence. The CMA's final price transparency guidance is the current official starting point for those consumer presentation questions.
Map every identifier, storage or access operation, audience instruction, recipient, exposure record and outcome record. The ICO's final storage and access technologies guidance must be applied by a privacy specialist to the actual operations. Define a refusal path and keep necessary shopping functions separate from advertising measurement.
Days 31 to 50: specify measures and gates
For each measure, write the event, eligible population, numerator, denominator, period, Europe/London timezone, source version, duplicate and late-event handling, exclusions, correction status and owner. Do not set a target from an unrelated benchmark.
Create independent gates for advertising recognition, consumer presentation, privacy and PECR, accessibility, security, IP and finance. CAP's ad-recognition rules require assessment of the rendered communication and its context. Gate approval cannot be inferred from a template field or supplier badge.
Acceptance at day 50 requires testable definitions, assigned reviewers and complete expected evidence. An unresolved gate means hold.
Days 51 to 70: run a controlled fixture
Use synthetic or specifically authorised product, audience and event records. Test approved and withdrawn creative, tracking acceptance and refusal, duplicate and late events, accessible customer and staff tasks, incorrect price correction, access revocation, incident escalation, export and rollback. No live targeting, purchase or customer message is needed.
The NCSC's secure online-service guidance supports evidence questions about identity, privileged access, logging, monitoring, incidents and supply-chain responsibility. Record what the test actually demonstrates. A successful fixture does not prove the service secure or the campaign effective.
Days 71 to 85: reconcile and challenge
Match each observed output to its source and query version. Keep delivery events, attributed outcomes and causal claims separate. Use HM Treasury's Magenta Book as methodological guidance for counterfactual thinking, while noting its public-policy remit.
Ask a reviewer who did not design the test to inspect failures, missing events, inaccessible paths and contrary evidence. Re-run only corrected cases, preserving the original record. Exercise export, isolated restore and rollback before calling the route reversible.
Days 86 to 90: decide, do not drift
The accountable owner receives the proposition, evidence index, cost ledger, measure dictionary, test results, gate verdicts, open risks and recovery proof. Proceed means only that the bounded planning test is ready for the next authorised stage. Revise names a correction and retest. Pause preserves the evidence while a gap is resolved. Stop retires the proposition and executes access removal, export, deletion or archive obligations.
Set source, contract and control recheck dates. A change to creative, product scope, data recipient, channel, measurement definition or customer journey reopens the affected gates. This article remains on publication hold until named UK specialists review the current records and the precise proposed exchange.