The Decision System Behind Strategy Execution

Strategy execution is a decision system

Strategy does not become real when it is approved. It becomes real when leaders make a sequence of choices about where to invest, what to stop, which risks to accept, who may decide and how the organisation will know whether its choices are working. Execution therefore depends on the quality and pace of decisions across the enterprise, not simply on the quality of the strategy document.

Atlassian’s new portfolio-management research opens with the line, “Most strategies don’t fail. They stall.” Its survey reports a striking distance between perceived clarity and operational connection: 80% of leaders say strategy is clear, but only 11% report that the work of execution is connected to it in a system. It also reports that only 8% use dynamic, outcome-based funding and that fewer than one in three decide within a week of a performance signal. Nearly two-thirds take a month or longer. These figures come from a vendor-sponsored survey of 308 leaders and should be read as directional evidence, not as a representative measure of every market or sector.[1]

The pattern is useful despite the caveat. A strategy can be well communicated yet fail to shape weekly trade-offs about capacity, investment and attention. When portfolios lack visible links to outcomes, leaders see activity without strategic contribution, while teams may face unclear funding, dependencies or authority.

The executive test is whether material evidence can change the allocation of people, money, technology and attention quickly enough to matter. Strategy must become a small set of outcomes, initiatives linked to those outcomes, and authority to continue, change or stop work as evidence shifts.

Strategy is also a coordination problem. A choice made in one function can alter another’s assumptions, costs or risk. Leaders need a shared view of outcomes and dependencies, with room for local decisions inside enterprise guardrails. This keeps strategy coherent without routing every operational choice through the centre.

Connect strategic intent to resource choices

The portfolio is where intent meets constraint. When capital, skills, energy, infrastructure or supplier capacity are tight, strategy execution is an ongoing allocation problem: which outcomes deserve capacity now, what can wait, and what should be reduced to make room?

The Atlassian findings on outcome-based funding point to a familiar tension. Annual budgets provide control and predictability, but may make it hard to redirect investment as assumptions change. Dynamic funding does not mean abandoning controls or moving money on impulse. It means agreeing decision thresholds, evidence requirements and delegated authority in advance, so that a team can adjust a portfolio without waiting for an annual planning cycle to reopen every choice. The board remains responsible for risk appetite and strategic direction; leaders closer to delivery need defined room to move within those boundaries.

Outcome-based funding also requires a credible link between money and results. If the intended outcome is vague, a flexible budget can simply make activity move faster. Leaders should state the assumptions underpinning investment, the evidence to review and who may change a commitment. This protects accountability while allowing resources to follow stronger evidence.

A strategy-linked portfolio should show the outcome each initiative is meant to change, the evidence of progress, the people and capabilities required, and the decision that follows when an agreed threshold is crossed. This makes the review about value, uncertainty and choice rather than status alone.

Review the initiatives competing for the same outcome, capacity or constraint, not only each project’s status. A delayed data platform, for example, can affect customer, risk and reporting work. Leaders should decide whether to resequence, resource, simplify or stop, reducing the delay between a changed assumption and changed work.

Riviera Partners’ survey reinforces the organisational dimension. It reports that 19% of respondents’ organisations were classed as Advanced in AI execution maturity; 57% were Developing and 25% Emerging, with the totals affected by rounding. Advanced organisations were more likely to report significant business impact, but the study is cross-sectional and its categories are defined by the publisher. It does not prove that one structure causes better outcomes. It does, however, give executives a prompt to examine whether ownership, governance and operating structures help the organisation move from intention to delivery.[2]

Build decision intelligence around context

Decision intelligence is the discipline of making important decisions explicit: what decision is being made, what outcome it serves, who is accountable, what evidence is relevant, what options are available, and how the organisation will learn from the result. It is broader than analytics and more practical than a catalogue of dashboards. Analytics can surface a signal; decision intelligence connects that signal to context, authority, action and feedback.

A CX Today analysis of TheyDo-commissioned research describes a “decision-confidence gap”. Among 1,000 decision-makers, 77% said dashboards, CRM data and customer feedback sometimes tell different stories; 43% delay or avoid decisions at least weekly for lack of context. The sample covers large organisations in Belgium, Germany, the Netherlands, the United Kingdom and the United States, not South Africa. It illustrates how information can be plentiful while context remains fragmented.[3][4]

The survey also reports that 38% experienced an AI-driven decision causing an unintended customer-experience problem. This is not proof that AI caused it: the finding is self-reported and the survey sponsor sells customer-journey technology. It does raise a sound question: can leaders see the cross-functional effects of a changed price, service, eligibility rule or product journey before complaints or churn reveal harm?

A short decision record names the trigger, owner, contributors, intended outcome, guardrails, evidence and assumptions; captures the action and uncertainty; and sets a review date. Reserve this proportionate control for choices with material financial, customer, operational, regulatory or sustainability consequences.

The record also makes latency measurable: elapsed time from signal to decision, then from decision to implementation. This separates analysis from avoidable waiting in committees, reconciliation or unclear sign-offs. Speed alone is not the goal; the aim is timely, explainable action with feedback.

For a high-value decision, map a simple chain: signal, interpretation, decision, action, outcome. It exposes where meaning is lost. A sales forecast may exclude returns; a service metric may hide a segment; an AI suggestion may omit a constraint. The remedy depends on where the chain breaks, not on adding another dashboard.

Treat master data as decision infrastructure

Master data management becomes strategically material when the same real-world entity is described differently across systems. If one unit treats a customer as a billing account, another as a household and a third as a digital identity, a group-level customer decision may combine unlike things. If a product hierarchy differs between finance, sales and operations, contribution margin, inventory exposure and campaign results may not reconcile. If a supplier is represented under several names, procurement concentration and sustainability risk can be understated.

The master-data question is not “Have we cleaned the database?” It is “Which decisions depend on this definition, and what happens when it is inconsistent?” Customer, product, supplier, employee, asset, location and entity records enable joins, roll-ups, segmentation and accountability. Trusted identifiers and hierarchies can distinguish a decision-grade group view from one that hides duplication or exclusions.

Joe Ouko’s CIO Africa analysis argues that reference and master data allow AI to generalise across an enterprise. He notes that a model can misfire when “customer”, “product” or “account” does not have a trusted, shared definition. He also connects metadata and lineage to explainability: a leader, regulator or auditor needs to know what data meant, where it originated and how it was transformed. Ouko’s piece is practitioner commentary rather than original empirical research, but its operational point is important: if a recommendation cannot be traced to its source and definitions, accountability for acting on it becomes difficult.[5]

MDM alone does not create decision intelligence. A clean customer master cannot explain churn, and an asset register cannot choose the best maintenance investment. MDM supplies dependable identities, relationships and definitions; decision intelligence adds the business question, operating context, decision rights and outcome measures. Both matter when a choice spans systems or functions.

Nor does a reliable data foundation require every record to sit in one central platform. A federated design can preserve systems of record while governing shared identifiers, hierarchies, definitions, stewardship, quality rules and lineage across them. The design choice should follow the use case, risk and operating model. The business outcome is consistency where it matters, with clear authority to resolve conflicts—not centralisation as an end in itself.

Judge an MDM service by the decisions it improves: supplier concentration and performance, product profitability, entity and asset reporting, or consistent customer choices across service, sales and risk. This puts MDM alongside performance management and risk control, rather than periodic technical clean-up.

That calls for business ownership of definitions. Domain owners set the meaning and permissible use of critical records; data stewards maintain standards and resolve exceptions; technology teams implement controls and integration. Decision owners state the quality threshold needed for each use. Clear roles stop data issues being passed indefinitely between business and IT.

Design governance into AI and leadership

AI increases the reach and speed of both useful decisions and flawed assumptions. That changes governance from a late-stage approval step into part of decision and system design. Riviera Partners’ 2026 material reports that 87% of its Advanced-maturity respondents involve cybersecurity, legal and governance review at initial design, while none reported handling governance immediately before deployment or on an ad hoc basis. Among Emerging organisations, 56% used those later approaches. The study finds an association; it does not isolate early governance as the cause of greater maturity.[2]

The execution logic is compelling. If privacy, security, lineage or customer safeguards are considered after a workflow is built, teams discover constraints late and redesign. Involving risk, legal, data owners and operations early clarifies purpose, permissible data, human review and escalation. The aim is to settle material questions early, then delegate within clear boundaries.

KPMG Middle East’s leadership paper describes three shifts for the Intelligence Age: “from directing to orchestrating”, from risk avoidance to intelligent risk tolerance, and from delegation to accountability proximity. This is a leadership framework, not a measured universal law. Its useful implication is that a senior leader cannot outsource judgement to a model or a vendor. Leaders must shape how people, automated systems and controls work together, stay close to decisions with material consequences, and remain answerable for outcomes.[6]

Accountability needs an operating form. For each automated recommendation, leaders should know who owns the outcome, who monitors performance, how exceptions reach a person, and what evidence can suspend or reverse action. Monitoring should include drift, data changes, overrides and unintended effects. This makes responsible use a working practice rather than a policy statement.

For African organisations, legacy platforms, uneven connectivity, limited specialist capacity and varied market contexts make sequencing valuable. Start with a consequential decision and the minimum trustworthy evidence it needs. Set a human owner and escalation route; test real workflows and exceptions before extending authority. Scale when quality, control and value are visible together.

Connect ESG reporting to operating decisions

Sustainability information matters strategically when it changes a choice about investment, assets, suppliers, products, markets or risk. Reporting should therefore be connected to the same management system that governs financial and operational performance. If emissions, energy use, workforce indicators or supplier information are prepared only for an annual disclosure cycle, leaders lose the chance to use those signals while they can still alter the operating plan.

IFRS S1’s objective is to require disclosures about sustainability-related risks and opportunities that are useful to users making decisions about providing resources to an entity. It covers governance, strategy, risk management, and metrics and targets.[7] This is a useful design reference because it links reporting information to strategy and decision usefulness. It is not, by itself, a statement about which rules apply to every South African organisation; each entity should determine the requirements relevant to its jurisdiction, listing status and reporting framework. The JSE’s sustainability disclosure guidance is a local reference point, and should be treated according to its stated status as guidance rather than assumed to be a blanket mandate.[8]

Master data affects sustainability confidence through boundaries and relationships. Which legal entities sit in the reporting group? Which sites, meters, vehicles or assets belong to each entity? Which supplier relationships are in scope? How are products, locations and business units mapped between operational records and reporting categories? If those relationships are ambiguous, a figure may be numerically precise but incomplete or inconsistent. Data ownership, definitions, lineage and evidence retention let a finance or sustainability leader explain how the reported measure was assembled and what decision it informs.

Review ESG signals alongside capital allocation, procurement, operations and risk. Energy intensity may change maintenance or investment timing; supplier exposure may change sourcing; workforce indicators may reveal capacity or safety risks. Reporting confidence needs accountable owners, repeatable calculations and traceable evidence. Decision quality improves when leaders use these signals before the reporting deadline.

For South African groups spanning subsidiaries, sites and supply chains, the reporting perimeter is itself a management question. Entity structures, shared facilities, outsourced operations and supplier tiers affect what a measure represents. Mapping source records to reporting boundaries helps explain omissions and year-on-year changes, while making remediation ownership visible when evidence is incomplete.

Operate one connected execution rhythm

The system is a recurring sequence: set a few strategic outcomes; identify the decisions that affect them; connect initiatives and resources; define needed evidence, shared data and guardrails; give an accountable owner authority to act; then review results and update the portfolio. Keep it practical at executive cadence and rigorous enough to withstand scrutiny.

Balance speed, quality and results. Track decision latency and follow-through; the share of material initiatives with outcomes and accountable owners; exceptions in decision-critical data; lineage for critical measures; and customer, financial, operational, risk and sustainability outcomes.

Set baselines before targets, reflecting risk and context. A regulated credit decision may require tighter evidence than a reversible marketing experiment. Prioritise data defects that change valuable choices, create exposure or impair accountability.

A compact review can ask five questions: What changed in the evidence? Which strategic outcome is affected? What choice is now available? Who has authority and by when? What result and risk will be reviewed afterwards? This focuses leadership attention on action while retaining a clear record of why the choice was made. Where the signal is uncertain, the best response may be a controlled experiment, a smaller commitment or a decision to gather specific evidence—not indefinite analysis.

Cadence should match the decision. Operational teams may review leading indicators daily or weekly; portfolio and executive forums can resolve cross-functional constraints monthly; boards focus on strategic exposure and outcome trends, with escalation for material events. Forums should exchange decisions and dependencies rather than repeat a status deck. A shared decision log preserves continuity.

A practical first 90 days

In the first 30 days, select one or two outcomes where delay or fragmented information has visible cost, such as retention, reliability, working capital, production availability, supplier resilience or sustainability exposure. Map each key decision: trigger, evidence, owner, approval path, systems, hand-offs and elapsed time. Interview decision-makers and implementers, not only dashboard owners. Record conflicting definitions and reconciliation delays.

Produce a decision map and baseline, not a new architecture programme. Estimate signal-to-decision and decision-to-action time; name the outcome and guardrails; identify critical data dependencies; and agree what evidence would change course.

Share the baseline with affected teams before the pilot. Explain the decision being tested, workflow changes, who can raise concerns and how feedback will be used. This matters when people work alongside automated recommendations or when a new data definition changes established performance reporting. Adoption depends on understanding both the purpose and limits.

From days 31 to 60, assign decision owners, authority, escalation thresholds and participating functions. Align the definitions and master-data relationships needed to interpret evidence. Name a steward and response time for contested sources; document critical lineage; specify AI use, human review, exceptions, traceability and stop conditions.

From days 61 to 90, run a pilot through the real cadence. Test whether the owner can act on a timely signal, then review outcomes and unintended effects. Compare with baseline: Was the decision faster? Did action reach operations? Did the measure improve within customer, risk and sustainability guardrails? Were data exceptions fixed at source or patched in a spreadsheet?

At day 90, expand, redesign or stop based on decision and outcome evidence, not technical completion. If performance did not improve, test whether the decision was remote from the outcome, the signal late, authority unclear, definitions wrong or capacity missing. Diagnose the failure before scaling.

The Emergent Africa point of view

Emergent Africa sees strategy execution, decision intelligence, MDM and sustainability reporting as connected executive disciplines. A portfolio view is limited if measures cannot be reconciled; data creates little value without decision rights; AI cannot accelerate a choice if people cannot interpret evidence or own the consequences; and ESG confidence needs explainable boundaries and sources.

The opportunity is a shorter, more accountable path from intent to result. Start with consequential choices, define their evidence, connect trusted data and give authority to people who can act. Review quickly enough to adapt and carefully enough to protect trust. That is how strategy becomes executable and decision intelligence becomes an operating capability.

Point of view word count: approximately 3,057 words, excluding references and LinkedIn pack.

References

[1] Atlassian, Leading the Human + AI Enterprise: A New Playbook for Strategic Portfolio Management, 2026 research page. Atlassian describes primary research with 308 leaders. Accessed 30 September 2026. Open source

[2] Riviera Partners, 2026 Future of Tech Leadership Report and related research. Landing page dated 22 September 2026; report release statement dated 23 September 2026. Survey of 958 senior technology executives in North America and Europe. Open source

[3] Francesca Roche, CX Today, Your Enterprise May Have a Decision-Making Problem, September 2026. Commentary on the decision-confidence findings. Accessed 30 September 2026. Open source

[4] TheyDo, The Decision Confidence Gap Report: What 1,000 Enterprise Leaders Told Us They Need to Trust AI-Powered Decisions, September 2026. Company-sponsored survey; fieldwork reported as June 2026 across Belgium, Germany, the Netherlands, the United Kingdom and the United States. Open source

[5] Joe Ouko, CIO Africa, Data Governance Is the Precondition for Trustworthy AI — Not a Compliance Afterthought, 25 September 2026. Practitioner commentary. Open source

[6] Mazhar Hussain, KPMG Middle East, Lead first: The human playbook for the Intelligence Age, 2026. Leadership paper; accessed 30 September 2026. Open source

[7] IFRS Foundation, IFRS S1 General Requirements for Disclosure of Sustainability-related Financial Information, issued June 2023; effective for annual reporting periods beginning on or after 1 January 2024, subject to applicable adoption and reporting frameworks. Open source

[8] Johannesburg Stock Exchange, Sustainability Disclosure Guidance, with accompanying climate disclosure guidance. Use as guidance and confirm applicability against current requirements. Open source