Scrappy Audience Feedback Loops
Use repeated lightweight audience checks to improve creative decisions
- Difficulty
- Easy
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 97%
Scrappy Audience Feedback Loops introduce real audience evidence into creative development through small, rapid rounds. Begin with a clearly defined audience and a concrete choice, such as ten possible newsletter titles. Contact relevant customers or prospects through whatever channels are readily available and ask a simple question that requires them to compare the options. Gather responses from as many appropriate people as practical, identify recurring preferences or misunderstandings, and revise the shortlist. A second round can test whether the revision improved the reaction. The method prioritizes speed and repeated learning over formal research infrastructure, but it does not pretend that a convenience sample is statistically representative. Its purpose is to challenge internal assumptions and improve creative decisions with genuine audience signals before committing substantial resources.
Origin
Jason Keith recommends running proposed newsletter ideas past customer-service leaders through LinkedIn DMs, an email list, or any other quick and scrappy channel, then repeating the feedback process across multiple rounds.
Core principles
- 01Creative judgment improves when exposed to target-audience evidence
- 02Quick imperfect research is often better than isolated speculation
- 03Questions should reveal audience preference without overexplaining
- 04Multiple rounds convert feedback into an iterative process
How to run it
- 1
Define the Audience Decision
Specify whose response matters and the exact creative choice you need evidence about.
Pro tip Recruit people who resemble actual customers or plausible customers rather than convenient colleagues.
Watch out Feedback from the wrong audience can confidently steer the work away from its market.
- 2
Prepare Concrete Options
Create a concise set of titles, concepts, hooks, or prototypes that respondents can compare.
Pro tip Ten options can provide useful range without requiring a long explanation.
Watch out Abstract descriptions make it difficult for respondents to react to the real choice.
- 3
Ask a Simple Question
Ask which option they would most want to read, use, click, or explore, based on the material shown.
Pro tip Collect their first interpretation before explaining what each option is intended to mean.
Watch out Leading questions can turn the exercise into a request for agreement.
- 4
Use Scrappy Channels
Reach respondents through LinkedIn DMs, an email list, customer conversations, or another available path.
Pro tip Choose the lowest-friction channel that still reaches relevant people.
Watch out Ease of access should not override audience relevance.
- 5
Find Patterns and Revise
Aggregate recurring preferences, confusion, and language, then update or narrow the options.
Pro tip Distinguish repeated signals from one person's unusually strong opinion.
Watch out Do not claim statistical certainty from a small convenience sample.
- 6
Repeat the Loop
Test the revised options with another group or return to the audience after meaningful changes.
Pro tip Multiple rounds are especially useful when early feedback reveals a hidden requirement.
Watch out Avoid repeated testing that makes no material change to the decision.
In the wild
A company creates ten possible names for a customer-service leadership newsletter. It sends the list to customers and prospects through LinkedIn DMs and email, asking which title they would be most interested in reading. Responses reveal that clever names obscure the leadership focus, so the team revises the top candidates and runs a second smaller round.
→ The final title reflects audience interpretation rather than the internal team's assumptions.
Common mistakes
Polling Only Colleagues
Internal peers may share the same assumptions and vocabulary, weakening the value of the feedback.
Explaining Before Asking
Explaining what an option is supposed to mean prevents the team from learning how people interpret it unaided.
Overclaiming the Evidence
Scrappy feedback improves judgment but does not provide the certainty of representative quantitative research.
Is it for you?
Best for
It is best for early newsletter, content, positioning, naming, and campaign decisions that can be shown as concrete alternatives.
Not ideal for
It is not ideal for statistically precise conclusions or decisions requiring representative samples and controlled research.
From the transcript
“and literally just running it by customer service leadership people out there in the wild that either are customers or could be customers and saying…”
“it could do whatever is like scrappy and quick because those iterative feedback pieces are really important and doing multiple rounds of it is even…”
“doing more kind of feedback loops and whatever way you can”
From the episode
I Tested 5 AI Prompts That Replace Weeks of Creative Work