Test Prioritization — Risk First
It is Friday afternoon. Basketly's release goes out in two hours and you have ideas for two days of tests. You will not run them all. The question is not "can I test everything?" but "if I only test some things, which ones?"
The rule
Risk is likelihood times impact: test the biggest risks first.
You can never test everything; even a single quantity box has thousands of possible inputs. So testing is always a choice, and risk is the fairest way to make it.
- Likelihood: how probable is a failure here? New code, changed code, complex logic and past bug hotspots all raise it.
- Impact: how bad is it if this fails? Money, safety, data loss, legal trouble and the main user journey all raise it.
Score each from 1 (low) to 3 (high) and multiply. The result, from 1 to 9, sorts your list.
| Area | Likelihood | Impact | Risk |
|---|---|---|---|
| Checkout payment (code changed) | 3 | 3 | 9 |
| SAVE10 coupon (untouched) | 1 | 3 | 3 |
| Search filters (new) | 3 | 2 | 6 |
| Footer links | 1 | 1 | 1 |
Prioritisation does not mean skipping low-risk areas forever. It means the order you test in, and being honest about what was not tested when time ran out.
The shortcut
For each area, ask two quick questions: "Did it change?" for likelihood and "Does it touch money or the main journey?" for impact. Two yeses put it at the top.
List the areas
Payment, coupon, search filters, footer links. Keep the list short enough to see at once.
The trap is testing in the order things appear on screen, or testing what is easiest. The footer is quick to check, but a broken footer link loses nobody any money.
Drills
Answer fast — trust the rule.
Why do testers prioritise at all?
Likelihood 2, impact 3. What is the risk score?
Which area deserves the highest likelihood score?
Time runs out before you test the footer links. What should you do?
Coupon: likelihood 1, impact 3. Newsletter signup: likelihood 3, impact 1. Which is riskier?
Remember
- Risk is likelihood times impact: test the biggest risks first.
- Ask "Did it change?" and "Does it touch money or the main journey?"
- When time runs out, report what was not tested; never pretend it passed.