Testing is simple if you know the steps.
Step 1: Select a tool and a panel to run your tests
You don’t have to hit the streets with a clipboard. There are online tools for testing. They collect the answers and analyze the results.
High rolling readers can buy fresh eyeballs for each questionnaire.
- Filter for age, gender and other things that represent your audience.
- You can ask the same question multiple times.
- Mercenaries are brutally honest.
- Can become expensive quickly.
- Mercenaries won’t always offer quality input when asked.
Frugal readers can ask their friends or customers to answer questions.
- Free! Yay!
- Soul mates are more generous with their own input.
- Family and friends don’t necessarily represent your audience.
- Soul mates go easy on the criticism.
- The results will come in slower.
- Volunteers tire. You can’t keep asking them.
Step 2: Present the panel with a real choice
Setting up a test is an investment for you and your panel. Make it usefull.
- Provide real alternatives for every
- Make a selection of 3 logos, not
- Propose 3 headlines, not
But don’t offer too many, for people are most comfortable choosing if the choice is limited. This is called ‘choice paralysis’.
The test audience doesn’t know everything you know. Make sure what you present your audience is:
- Relevant. Does it send the message you want it to?
- Distinctive. Does it stand out amongst similar brands?
- Memorable. Does it evoke enough emotion to be remembered?
Step 3: Ask questions about the choice
You can ask the test audience why they chose the design they did. Besides preference, you can ask how cool, trustworthy or clear it is.
Good questions find out if your brand promise has been made:
- Which design looks the most trustworthy?
- Which design looks the easiest to use?
- What would you have done differently?
The feedback allows you to find out areas for improvement. Or merge the best parts of each tested designs plus the feedback into a hybrid.
Step 4: Make sure your results are statistically significant
Statistical significance means your design isn’t beating the others by chance. If you ask just 10 people and the result is close, it’s not really significant.
But your results are conclusive and significant, you can safely rebuff team members who still want to champion their preference.
While there’s no accounting for taste, numbers never lie.