Every test team runs out of time before it runs out of things to test. That is not a planning failure; it is arithmetic. The number of possible tests for any real system is effectively infinite, and the calendar is not. So the useful question is never “did we test everything?” but “did we spend our time on the things that would hurt most if they broke?” Risk-based testing is the name for answering that question deliberately rather than by habit.

A worked example before the theory

Picture a team releasing a new version of a motor insurance app. The release touches six areas. The test lead gathers a developer, the product owner and a support team member for an hour, and they score each area on two scales from 1 to 5.

  • Likelihood: how probable is it that this area contains a defect? Developers usually know: new code, complex logic, a junior author, a rushed change, a history of bugs.
  • Impact: if it fails in production, how bad is it? The product owner and support usually know: money lost, customers affected, regulatory exposure, reputation.
AreaLikelihoodImpactRisk scoreBand
Premium calculation (new add-on covers)4520High
Payment and policy issuance3515High
KYC document upload (new OCR vendor)4312Medium
Claim intimation248Medium
Profile editing224Low
FAQ and help pages111Low

The score is likelihood multiplied by impact. Here the team set bands of 15 to 25 for high, 8 to 14 for medium and 1 to 7 for low. The result guides three decisions at once.

Depth. Premium calculation gets the full toolkit: decision tables for the add-on rules, boundary values for vehicle age and IDV, comparison against a reference calculation, and a review of the rate tables themselves. FAQ pages get a quick check that they load and links work.

Order. High-risk areas are tested first, so if the schedule collapses, what remains untested is the low-risk tail rather than the core.

Who and how. The most experienced testers take the high-risk areas. Automation investment goes where risk is high and the area changes often.

What “risk” means here

In testing, a risk is a possible event that would have a negative consequence. ISTQB distinguishes two kinds. Product risks are about the software itself failing: wrong premium, lost payment, data leak, slow page. Project risks are about the effort to build and test it: the environment arriving late, a key tester leaving, a vendor API sandbox being unreliable. Risk-based testing in the usual sense is driven by product risk, but project risks shape the plan too, because they decide how much test time actually exists.

Impact is not only financial. A defect that exposes Aadhaar numbers might affect few customers but carries legal consequences under India’s Digital Personal Data Protection Act, 2023, plus serious reputational damage. Accessibility failures can lock out entire groups of users. A good impact score considers customers, money, compliance and trust together.

Where the scores come from

The scoring session is where most of the value lies, more than in the numbers. Sources of likelihood evidence include:

  • How much code changed, and how complex it is.
  • Defect history: modules that produced bugs before tend to produce them again.
  • New technology, new vendors or new integrations.
  • Unclear or late-changing requirements.

Sources of impact evidence include usage analytics (which flows carry the most traffic or revenue), support ticket categories, contractual SLAs and regulatory obligations. When developers, product owners and support staff score together, they often disagree, and the disagreement is informative. A developer who rates KYC upload as likelihood 2 while support rates it as a top complaint has revealed something worth testing.

Pitfalls of the multiplication

The 5 × 5 grid is popular because it is simple, but it has known weaknesses.

Rare but catastrophic risks get under-weighted. Likelihood 1 and impact 5 scores 5, the same as likelihood 5 and impact 1. Most teams would not treat a one-in-a-thousand chance of wrongly cancelling every active policy the same as a near-certain typo on the help page. A common fix is a rule that any impact of 5 is at least medium, regardless of likelihood.

The numbers look more precise than they are. These are ordinal judgements, not measurements. A score of 12 is not “50 percent more risky” than 8. Use scores to sort and band, not for arithmetic beyond that.

Low risk is mistaken for no testing. Low-risk areas still get some coverage, usually through automated regression and light exploratory checks. The point is proportion, not abandonment.

The register is never revisited. Risk changes during a project. If testing finds twelve defects in an area scored likelihood 2, the score was wrong and should go up, which means more testing there, not less. A risk register that is not updated after the first week of testing is a historical document, not a plan.

Reporting in the language of risk

The quieter benefit of risk-based testing appears at release time. Instead of “we executed 840 of 900 test cases”, the test lead can say: “All high-risk areas are fully tested with no open serious defects. Of the medium-risk areas, KYC upload has two open defects with workarounds. The 60 unexecuted cases are all in low-risk areas.” Business owners can weigh that statement. They cannot weigh a raw count.

It also makes trade-offs negotiable. When the release date moves forward by a week, the team does not quietly test less of everything. It shows the risk table and asks which of the medium and low rows the business is willing to leave at reduced depth. That conversation is uncomfortable, but it puts the decision with the people who own the consequences.

Scaling it down and up

For a small hotfix, risk-based thinking might take ten minutes: what did this change touch, what is downstream of it, and what would be expensive if wrong? For an enterprise programme, it might be a formal register reviewed at every steering meeting, linked to requirements and test cases. Standards such as ISO/IEC/IEEE 29119 describe risk-based approaches in detail for organisations that want a formal process. The underlying habit is the same at every size: name the risks out loud, rank them honestly, and let that ranking decide where the hours go.

Written by the RP Strativa Team. The insurance app, areas and scores are illustrative.

Leave a Reply

Trending