A good survey question is a flashlight, not a floodlight

 
 

Never ask a young kid, “What do you want for dinner?”

The question is so broad that they’ll default to their favorite carb-heavy meal, then complain when you say “no.”

Just give them two options: "Shepherd's Pie or chicken and potatoes?”

They’ll still complain and demand mac n’ cheese. 

But trust me: you'll control the subsequent negotiation.

My parenting hack illustrates the difference between “floodlight” and “flashlight” questions.

A floodlight question, like 'What do you want for dinner?', is broad and difficult to answer.

A flashlight question, like '“Shepherd's Pie or chicken and potatoes?”, is narrow and easier to answer.

Yet the defining feature of a 'flashlight' question isn’t that it’s forced-choice.

Instead, it's the way it directs people toward specific, verifiable actions.

Case study: Squarespace's reason-for-cancellation survey 

Last week, after canceling my subscription for a Squarespace website designed for a now-defunct fantasy football league, I was asked, “Did you cancel your subscription for another website builder?”

I clicked 'no.’

It's a good "flashlight question" because it focused on what users actually did (whether they switched to a different website builder or not), which should help Squarespace to reduce cancellations.

The follow-up question, "Anything we can do to improve?" is a classic "Floodlight question."

It might gather a few insightful responses -- particularly if one suggestion is consistently mentioned -- but extracting meaningful insights from these open-ended responses can be challenging.

Plus, the vast majority of users will skip it.

 

Applying “flashlight questions” to ecommerce cancellations and unsubscribes

Here are three things you can do right now:
 

  1. Prompt shoppers onsite with a survey instead of sending one via email. Onsite surveys have a higher response rate.
     

  2. Ask about a specific behavior such as, "Are you leaving because you've subscribed to another brand?"
     

  3. Test other questions asking about specific behaviors, including:

    → "Did you stop using our product/service in the last three months?"

    → "Did a product issue or defect influence your decision to unsubscribe?"

    → "Do you need a specific feature or service that we don't provide?"

 

Next Steps

Tired of floodlight questions?

Need help designing flashlight questions?

Click on the link below and get a survey Roast. 

I’d love to help.
 
https://www.sammcnerney.com/survey-roast


Sam

PS: If you enjoyed this content, please like, comment, and/or repost on LinkedIn.

Previous
Previous

How to cut your survey in half without removing a question

Next
Next

Never ask respondents to perform more than one task