geopolitical intelligence decision support

A geopolitical intelligence subscription earns renewal only when it improves a decision inside the available clock.

7 min read

The seam

The geopolitical-intelligence purchasing system touches decision support when a supplier deliverable must reach a named owner early enough to alter a live option or approval.

What you leave with

Start procurement with one decision, named users, evidence need and current performance baseline before naming providers.
Give shortlisted suppliers the same bounded demonstration and preserve scores, evidence and dissent.
Calculate usable decision time by subtracting implementation and communication lead time from the last responsible moment.
Count approval hops and retain only those contributing a consequence judgement or reserved power worth their latency.
Scale option comparison and independent challenge with novelty, irreversibility and remaining decision value.
Renew against recorded use and changed decisions, with explicit data, licence and internal-knowledge exit terms.

The strategy committee receives a respected country brief every month. Its analysts discuss policy pathways across several operating markets. Users rate the writing highly. When a tariff announcement threatens a supplier commitment, the brief reaches procurement after the commitment can still be changed only at significant cost.

The subscription has delivered good analysis and weak decision support. Reputation, readership and accuracy cannot repair a missed decision window. The supplier may never have been told which choice its work had to alter, who held that choice or when the useful clock ended.

That failure begins in procurement. A broad request for geopolitical coverage rewards familiar brands, polished platforms and expansive country counts. One current product combines daily analysis with monthly outlooks. Another covers 100 countries. Neither cadence nor breadth proves that a named owner can act on the relevant evidence in time.

Write one decision before you name a provider. A product specification starts the comparison too late because it embeds an answer before the evidence need is clear. Define the recurring choice first. In this case, the supply committee must decide whether to keep, reduce or redirect a cross-market commitment after a stated policy trigger.

Name the decision owner, users, evidence needed, geographic scope, horizon, confidentiality and last useful receipt time. Then record current performance. Capture staff-hours, source cost, missing evidence, elapsed time and recent decisions that better input could have changed. FAR Part 10 binds covered US acquisitions, but its order is useful elsewhere: describe the need, then research whether capable commercial sources exist.

The baseline prevents a common renewal fiction. Twelve months later, a buyer can count logins and briefings while remaining unable to show that any decision improved. Record the existing route from question to evidence, evidence to judgement and judgement to authorised action. That is the process the service must beat.

Give every shortlisted supplier the same bounded test. Free-form demonstrations fail because each provider chooses its strongest historic work. The committee then compares one firm’s dashboard with another firm’s senior analyst and calls the result judgement.

Issue the same scenario, public-source pack, output limit, deadline and challenge session. Ask for the decision-facing answer you intend to buy. The Belfer vendor checklist includes trials or demonstrations, customisation, source qualification and subcontractor oversight. Turn those prompts into scored evidence requests.

Preserve the prompt, output, questions, scores and dissent. FAR 15.304 requires covered US source selections to tailor factors, address quality through non-price factors and disclose their importance. Your corporate process may differ, but comparability still depends on weights set before evaluators see the favourite provider.

Test separation of evidence, assumption and judgement. Ask what would change the conclusion and require delivery through the proposed operating team. A fine sample still leaves security, data handling and continuity for separate due diligence. Analytical quality answers only one part of the award.

Subtract the work after the decision. A deadline stated as “before the committee meets” conceals the hours needed to turn a judgement into action. The last responsible moment occurs when the preferred option can still be implemented without losing its intended value.

Calculate backwards. Usable decision time equals that moment minus implementation, communication and coordination lead time. Use the slower credible lead time when lateness destroys the option. The whole credible downstream requirement never belonged to analysis or approval.

A US Army rapid-planning paper uses a one-third/two-thirds guide to protect subordinate preparation. Treat approximately 33.3% of usable time as a ceiling for central work, leaving two-thirds downstream. A novel irreversible choice may consume more, but the chair must name the execution time being surrendered.

Recalculate after every delay. Time already spent cannot remain in the analysis budget. A provider response window also needs a receipt condition: uploaded, sent, acknowledged or discussed with the decision owner. These states carry different decision value when the clock is short.

Remove authority that contributes only delay. A respected deliverable can still expire while travelling through regional security, group risk, legal and the committee secretariat. Each handoff adds queuing and translation. Counting approval hops makes the hidden route visible.

Start with zero hops for a delegated local action. Add a level only when it contributes a material consequence judgement or exercises a reserved power over money, legal entities or irreversible commitments. MCDP 6 links a lower decision threshold with a faster cycle, while recognising that higher levels may see wider consequences.

Measure your authority round-trip in a demonstration or exercise. One unavailable executive can make a single hop slower than several prepared hops. Assign a deputy, pre-authorised branch or direct escalation route where the measured round-trip exceeds usable time. Every retained level should be able to state what judgement or power justifies its latency.

Scale analysis to what can still change. A fixed option count wastes time on familiar reversible choices and understates novel commitments. Depth should rise with novelty, irreversibility and consequence, then stop at remaining decision value.

For a familiar reversible move, test one plausible course against its failure conditions. A rapid comparison can examine a leading option, a materially different alternative and delay. Use deeper comparison and independent challenge when supplier termination, market exit or a public commitment cannot be readily undone.

A review of 32 high-risk-event studies found analytical strategies more prominent with less time pressure. That supports a conditional design, not a quota. Another review cites 5–10 people for group decisions and notes subgroup formation above 4. Add challengers only when each brings distinct information without consuming the clock in briefing.

Stage an irreversible choice where possible. Authorise reversible preparation now, preserve the point of no return and reserve deeper comparison for the later commitment. Stop analysis when another source or challenger cannot plausibly alter action before implementation must begin.

Record use at the decision, rather than at the platform. User surveys fail because satisfaction is easy to remember and decision influence is not. A widely read monthly note may change nothing. One rarely used escalation call may preserve an option worth more than the annual subscription.

Create a contemporaneous log. Record the requirement, promised and actual receipt, correction, analyst access, named user and decision reference. State whether the work changed an option, condition, timing, approval or confidence. FAR Subpart 42.15 requires covered US agencies to evaluate performance at least every 12 months and at completion. A monthly private log can be lighter while preserving the evidence renewal needs.

Compare that record with the original baseline before the 12-month term ends. Separate consumption from value. Review analyst continuity, product changes, corrections, security events, unused hours, price movement and whether the original decision still exists. An accurate forecast followed too late belongs in the timeliness failure column.

Contract for an exit the decision process can survive. Exit language fails when it covers account closure but leaves evidence, licences and internal memory inside the platform. The live decision route may still cite expiring content after procurement declares the service closed.

Export permitted material, reassign open requirements and identify decisions relying on the supplier. Resolve reuse rights for extracts embedded in board papers, models and country files. Capture internal notes on sources, assumptions and past challenge without copying content the licence prohibits.

Where the supplier processes personal data for the buyer under UK GDPR, Article 28(3) requires specified contract terms across 8 minimum areas. At service end, the processor terms provide for return or deletion at the controller’s choice, subject to legal retention. Apply that rule only where the actual processing roles bring it into scope.

The renewal decision is now narrow enough to defend. The service reached the owner before the last responsible moment, survived the required challenge and changed a recorded choice. If it did none of those things, excellent analysis may still deserve respect. It has not earned this subscription’s renewal.

Questions answered

What should a geopolitical intelligence procurement requirement contain?

A geopolitical intelligence procurement requirement should name one business decision, its owner and users, the evidence gap, geographic and time scope, confidentiality, delivery deadline, consequence of delay and current performance baseline before any product form or provider is selected.

Why is provider reputation insufficient evidence of decision value?

Provider reputation does not establish fit with a particular decision, evidence need, delivery clock or approval route. Decision value requires evidence that the service reached the authorised owner in time and changed an option, condition, timing, approval or recorded confidence.

How should shortlisted geopolitical intelligence providers be demonstrated?

Shortlisted providers should receive the same scenario, public-source pack, output limit, deadline and challenge session. The record should preserve the prompt, response, questions, scores, supporting evidence and evaluator dissent.

How is usable decision time calculated?

Usable decision time equals the last responsible moment minus credible implementation, communication and coordination lead time. The remainder is available for framing, options, challenge and approval, and should be recalculated whenever delay consumes part of it.

How much usable time should central decision work consume?

A US Army rapid-planning paper uses a one-third/two-thirds guide to protect subordinate preparation. In corporate use, approximately one-third of usable time is a protective ceiling for central work, while approximately two-thirds remains for downstream preparation and execution.

How should approval hops be assessed?

Approval hops should be counted from the evidence holder to the authorised decider. A hop should remain only when it adds a material consequence judgement or exercises a reserved power; its value must justify the measured queuing, translation and availability delay.

When should a geopolitical intelligence decision receive deeper option analysis?

Option depth should rise when the decision is novel, hard to reverse or consequential and enough usable time remains. A rapid comparison can test a leading option, a materially different alternative and delay; deeper analysis should stop when another increment cannot alter action before the deadline.

How should independent challenge be sized?

Challenge should be sized by distinct information needs, independence and remaining time. A group-decision review cites five to ten people while noting that groups larger than four can form subgroups, so those figures are design references rather than a universal crisis-team size.

What should a geopolitical intelligence performance log record?

A performance log should record the requirement, delivery and receipt times, accepted outputs, corrections, analyst access, use by named decision owners, options or approvals changed, confidence change, hours consumed, product changes and unresolved issues.

What evidence should support renewal of a geopolitical intelligence subscription?

Renewal should compare contemporaneous performance records with the original decision baseline. Usage, changed decisions, timeliness, corrections, analyst continuity, security events, unused capacity, price movement and continued need should be assessed separately.

What should happen when a geopolitical intelligence service ends?

Exit should export material the licence permits, capture internal decision knowledge, reassign open requirements, close accounts, resolve citations and embedded content, confirm buyer-data return or deletion, and identify live decisions still relying on expiring access.

What does UK GDPR Article 28 require at the end of processor services?

Where UK GDPR Article 28 applies, processor terms must provide for return or deletion of personal data at the controller's choice and deletion of existing copies, unless UK law requires retention. Actual processing roles determine whether the requirement applies.