An independent study reference written by Dr Phuc V. Nguyen. It is not official subject material — for assessment requirements always follow your subject outline and vUWS.

How Analytics Gets Builtintermediate

Analytics capability

Q: An organisation has a well-documented warehouse, six capable analysts and a modern platform, but no manager may change a price without quarterly committee approval. How is its analytics capability best described?

Limited by its weakest component, which here is the authority to act on a result. Capability is a chain. The output of the whole system is bounded by the slowest link, so a quarterly approval cycle caps the value of everything upstream of it regardless of how strong that upstream is. Averaging hides exactly this, which is why the minimum is the more honest summary and why the useful intervention here is a delegation rule rather than more analytics investment.

Q: Which organisation is more mature in the sense that maturity models intend?

One that reliably produces simple, monitored, documented analytics on a published schedule. Maturity is about repeatability, not peak sophistication. A process that produces a modest result dependably can be relied on, improved and audited. A process that occasionally produces something excellent cannot be planned around, and its output decays quietly once the person who built it moves on.

Analytics capability is an organisation's durable ability to turn data into decisions repeatedly, as distinct from having produced one good analysis. It has several components that must all hold: data that is accessible and documented, people with the relevant skills, a process for choosing and delivering work, technology that runs reliably, governance covering ownership and access, and, most easily forgotten, the authority to act on a result. Capability behaves like a chain rather than a sum. Strengthening the component that is already strongest changes nothing, which is why platform purchases so often produce no measurable improvement.

Why it matters

Think of a relay. Data has to reach an analyst, the analyst has to reach a decision-maker, and the decision-maker has to be allowed to change something. If any runner drops the baton the race is lost, no matter how fast the others are. Most organisations know which runner drops it. The uncomfortable part is that the fix is usually organisational, and buying more of what already works is far easier than fixing the weak link.

Before you read on — recall

An organisation has a well-documented warehouse, six capable analysts and a modern platform, but no manager may change a price without quarterly committee approval. How is its analytics capability best described?

Formulas

Capability as a bottleneck, not an average

C = \min\left(d_{\text{data}},\, d_{\text{people}},\, d_{\text{process}},\, d_{\text{tech}},\, d_{\text{govern}},\, d_{\text{act}}\right)

Each

d

is a rough score for one component on a common scale. Writing capability as a minimum rather than a mean is deliberately crude, but it encodes the argument that matters. Raising the highest score leaves

C

unchanged, and staged maturity models work the same way, since an organisation reaches a level only when every component at that level holds.

Worked examples

Scenario

A general insurer has a documented data warehouse, six capable analysts and a modern platform. Prices can only be changed by a committee that meets quarterly, and any change needs a paper submitted twenty-one days in advance.

Solution

Score the components. Data, people and technology are strong. The ability to act is weak, and it sets the ceiling. A pricing model that could be refreshed every seven days delivers value four times a year at best, and each delivery is delayed by the submission cycle. No amount of further investment in the strong components moves the outcome. The intervention with actual leverage is a delegation rule, for example allowing movements inside an agreed band without committee approval. That is a governance change, not an analytics project, which is exactly why it tends to go unaddressed.

Scenario

A logistics firm hires eight data scientists into a central team. A year later they report spending most of their time requesting and reconciling extracts from four operational systems.

Solution

The people component was raised while the data component was left alone, so total capability did not move. The visible symptom is expensive staff doing low-value integration work, and the usual response is to hire more of them, which makes the imbalance worse. The correcting investment is unglamorous: documented, refreshed, permissioned access to the four systems. Once that exists, the same eight people produce far more, and this ordering, data access before headcount, is one of the more reliable patterns in capability building.

Common mistakes

✗Buying a platform creates capability. Technology is one component. Without documented data, someone who understands the business question and a decision-maker permitted to act, a capable platform runs a small number of reports at high cost.
✗Capability is measured by how many analysts an organisation employs. Headcount is one component and is easily over-weighted. Analysts whose data access is poor spend their time on extraction and reconciliation, so the marginal analyst adds very little.
✗Capability must be centralised or it does not exist. Centralised, hub-and-spoke and fully embedded structures all work in practice. What distinguishes the ones that work is that ownership of data, of priorities and of decisions is written down, not which shape the organisation chart takes.
✗A maturity model measures how advanced your techniques are. Maturity models mostly measure repeatability. An organisation that reliably delivers simple, monitored, documented analytics scores higher than one that occasionally produces a brilliant model nobody maintains, and that ranking is the right way round.

Revision bullets

•Capability = repeatable ability to turn data into decisions, not one good project
•Components: data, people, process, technology, governance, authority to act
•Behaves like a chain, so the weakest component sets the ceiling
•Investing in the strongest component buys nothing measurable
•Maturity models reward repeatability, not sophistication
•Structure can be central, hub-and-spoke or embedded; explicit ownership is what matters

Quick check

Which organisation is more mature in the sense that maturity models intend?

Connected topics

What Analytics Is Why Analytics Stalls Analytics Leverage Into Production Agile Analytics Data Governance Analytics as a Service Analytics Roles

More in How Analytics Gets Built

Why Methodology CRISP-DM Structuring a Problem Business Model Mapping Design Thinking

Sources

Davenport & Harris (2007)
Davenport, T. H., & Harris, J. G. Competing on Analytics: The New Science of Winning. Harvard Business School Press, 2007.
Argues that analytics is an organisational capability spanning data, people, leadership and process rather than a technology purchase.
Davenport, Harris & Morison (2010)
Davenport, T. H., Harris, J. G., & Morison, R. Analytics at Work: Smarter Decisions, Better Results. Harvard Business Press, 2010.
Source of the DELTA components (data, enterprise, leadership, targets, analysts) used widely as a capability checklist.
Chrissis, Konrad & Shrum (2011)
Chrissis, M. B., Konrad, M., & Shrum, S. CMMI for Development: Guidelines for Process Integration and Product Improvement. 3rd ed. Addison-Wesley, 2011.
The staged maturity convention in which a level is reached only when every component at that level is satisfied, which is the origin of the minimum rule used above.

How to cite this page

Dr. Phil's Quant Lab. (2026). Analytics capability. Derivatives Atlas. https://phucnguyenvan.com/concept/ba-analytics-capability