
In short
- A decision table lists conditions as rows, combinations as columns, and the expected action for each combination.
- It forces every combination to be thought about, which is where business rules usually have gaps or contradictions.
- Each column becomes one test case, and columns with the same outcome can often be merged to keep the set small.
On this page
The product owner writes the shipping rules on a sticky note: free shipping for members, free shipping on orders over $50, express delivery costs $15 extra, and nothing ships free outside the US. Four rules. They seem clear until you ask what happens to a member ordering $30 with express delivery to Canada.
Nobody knows. That’s not because the product owner is careless. It’s because four yes-or-no conditions make sixteen combinations, and nobody thinks about sixteen combinations on a sticky note. A decision table does.
01What a decision table is
A decision table has two parts. The top lists the conditions, each true or false. The bottom lists the actions, the results you expect. Each column is a rule: one combination of conditions and the actions it should produce.
| Conditions | R1 | R2 | R3 | R4 |
|---|---|---|---|---|
| Customer is a member | Y | Y | N | N |
| Order over $50 | Y | N | Y | N |
| Express delivery | N | N | N | N |
| Ships within the US | Y | Y | Y | Y |
| Action: standard shipping free? | Yes | Yes | Yes | No, $6.99 |
| Action: express charge | n/a | n/a | n/a | n/a |
With four conditions there are 2 × 2 × 2 × 2 = 16 columns. Writing them all out feels slow. It’s also the fastest way to find the questions nobody has answered.
02Building the full table, step by step
- List the conditions. Only things that change the outcome. Here: member, over $50, express, inside the US.
- List the actions. Shipping fee, express fee, and anything else that varies, like an error message.
- Write every combination. For n yes/no conditions, 2ⁿ columns. A simple trick: alternate Y/N every column in the last row, every 2 in the row above, every 4 above that, and so on.
- Fill in the expected action for each column from the requirements. Every time the requirement doesn’t say, write a question mark and ask.
- Merge columns that don’t matter. If outside the US always means "no free shipping" whatever else is true, those columns can collapse into one with "don’t care" marks.
03From table to test cases
Each remaining column is one test case. Pick concrete values that match the column, and use boundary values for any numeric condition.
| Test | Setup | Expected |
|---|---|---|
| R1 | Member, $60 order, standard, to Ohio | Shipping $0 |
| R4 | Guest, $30 order, standard, to Ohio | Shipping $6.99 |
| R5 | Member, $30, express, to Ohio | Shipping $0 + express $15 (as agreed) |
| R9 | Member, $60, standard, to Canada | International rate, no free shipping |
| Boundary | Guest, exactly $50.00, standard, to Ohio | Per the agreed answer on "over $50" |
04When to reach for a decision table
- Pricing, discounts, shipping and tax rules.
- Insurance, loan or credit decisions with several eligibility criteria.
- Permissions: role × resource × action.
- Any requirement written as a list of "if this and that, then…" sentences.
If the conditions aren’t yes-or-no, you can still use the technique. A condition like "customer tier: bronze, silver, gold" just has three values instead of two, and the number of columns multiplies accordingly.
05Keeping it manageable
Ten yes/no conditions give 1,024 columns, which is too many to test by hand. Three ways to cope:
- Collapse "don’t care" columns wherever a condition makes others irrelevant.
- Split the table. Shipping rules and tax rules rarely interact; give each its own table.
- Use pairwise selection for the leftovers, which tests every pair of values without every full combination.
Decision table checklist
- Every condition actually changes an outcome
- Every combination is listed, or deliberately collapsed with a reason
- Every unclear outcome was turned into a question and answered
- Each column has one test case with concrete values
- Numeric conditions also have boundary tests
- The table is kept with the requirement, so it’s updated when the rules change
Frequently asked questions
What is decision table testing?
A black-box test design technique that lists every combination of input conditions and the expected result for each, then turns each combination into a test case. It is used when several rules interact.
How many test cases does a decision table produce?
One per column. With n yes/no conditions that is 2 to the power n before simplifying. Collapsing columns where some conditions don’t matter usually cuts that down a lot.
What is the difference between a decision table and boundary value analysis?
Boundary value analysis picks values at the edges of a single input range. Decision tables deal with how several conditions combine. They work well together: use the table for combinations and boundaries for the values inside each one.
What is a "don’t care" condition?
A condition whose value doesn’t affect the outcome in that column. Marking it lets several columns merge into one, which keeps the table and the tests smaller.
Is decision table testing in the ISTQB syllabus?
Yes. It is one of the black-box test techniques in the ISTQB Foundation Level syllabus.
Written by
Vihari Panyam
QualityAIQ