Skip to main content
Packaging has seconds to communicate what a product is and why it’s worth buying — on a crowded shelf or a search-results grid. This playbook measures first impressions, identifies which design elements actually land, benchmarks against competitors, and iterates the design to a validated winner. The loop is designed to be run autonomously by an AI agent (Claude, ChatGPT, Cursor, or any client with PickFu MCP / CLI / API access) with minimal human intervention.

​
When to use this playbook

Use it whenSkip it when
You’re designing or refreshing product packagingThe packaging is locked and can’t change
You can produce 2+ design variationsYou only have one design and no ability to iterate
You have 2–3 competitor packages to benchmark againstYou need pricing or concept feedback, not design

​
The optimization loop

1

Test first impressions

A 5-second test reveals what the packaging communicates at a glance — what the product is and which features stand out before deliberate study.
2

See what catches the eye

A click test (heatmap) shows which areas of the design draw attention, and the written feedback explains why.
3

Benchmark against competitors

Test your current design vs. 2–3 competitor packages. Save this competitor set — you reuse it in the final validation.
4

Iterate your design

Generate or upload improved variations and run head-to-head polls until a design wins with a score of 70 or higher.
5

Re-validate against the original competitor set

Run a final poll with the winning design plus the SAME competitor packages from step 3 to confirm the improvement holds against the category.
Sample size and cost. Defaults to small, cheap polls — 50 respondents for the impression, click, benchmark, and validation tests, 15 respondents per iteration. The loop triangulates across polls, so a single noisy result self-corrects. Scale the final validation to 100–200 before committing to an expensive print run. See sample size guidance.

​
Run this playbook with an AI agent

Copy prompt for AI

Paste this prompt into Claude, ChatGPT, Cursor, or any AI agent connected to the PickFu MCP server, CLI, or REST API. The agent will run the entire loop on your behalf — creating polls, reading responses, and iterating until a winning variation emerges.
You are running the PickFu "Packaging design optimization" playbook end-to-end.
Goal: refine a product's packaging design so it communicates clearly, draws the
eye, and out-converts competitors on the shelf.

Before starting, ask the user for:
- The product and what it is (e.g. "organic dog treats, 12oz bag")
- Their current packaging design image
- 2-3 competitor packaging images (the canonical competitor set — fixed across
  steps 3 and 5)
- Target audience (default: General; refine only if the user requests it)

Run this loop:

1. FIRST IMPRESSIONS (50 respondents, five_second_test).
   Show the packaging for 5 seconds. Question: "Based on what you just saw, what
   do you think this product is? Which features or benefits stood out to you?"
   Read responses to learn what the design communicates (and what it misses).

2. WHAT CATCHES THE EYE (50 respondents, click_test).
   Question: "Which areas of this product packaging stand out to you most, and
   why?" The heatmap + comments show which elements draw attention.

3. BENCHMARK (50 respondents, ranked).
   Compare the current design vs. competitor packages. Question: "Based on the
   packaging, which product would you buy? Why?" Record the baseline ranking.

4. ITERATE (15 respondents per poll, head_to_head).
   Using insights from steps 1-2, generate or commission 2-3 design variations
   that fix what didn't land and amplify what did. Test each variation vs. the
   original with head_to_head (exactly 2 options). Question: "Which packaging
   design do you prefer and why?" Repeat until a variation wins with a 70+ score.

   Stop condition: if 5+ iterations don't yield a 70+ winner, halt and report —
   the issue may be brand recognition or product category, not the design.

5. RE-VALIDATE (50 respondents, ranked).
   Compare the winning design against the SAME competitor packages from step 3.
   Same question as step 3. A higher ranking/score than the baseline confirms
   the design improvement holds against the category.

Final report must include:
- What the design communicated in 5 seconds (step 1) vs. intended message
- The high- and low-attention areas from the heatmap (step 2)
- Baseline ranking/score vs final ranking/score
- The specific design changes that drove the improvement
- The winning design URL

Tools to use:
- save_survey + publish_survey  — create and launch each poll
- get_survey_responses          — read responses
- extract_images                — view/describe the heatmap and designs
- generate_image                — create design variations (step 4)
- upload_media                  — for designs created outside PickFu
Want to run this manually? The same flow is available as a one-click template in the PickFu app (Start the packaging playbook) with pre-filled poll URLs.

​
Step-by-step (human operator view)

​
1. Test first impressions

SettingValue
Poll type5-second test
Question”Based on what you just saw, what do you think this product is? Which features or benefits stood out to you?”
OptionsYour packaging design image
AudienceGeneral
Sample size50
What you’ll get: whether the packaging communicates the product and its key benefit in the first few seconds — the single most important test of shelf legibility. Launch the 5-second test → · See an example →

​
2. See what catches the eye

SettingValue
Poll typeClick test
Question”Which areas of this product packaging stand out to you most, and why?”
OptionsYour packaging design image
AudienceGeneral
Sample size50
What you’ll get: a heatmap of attention plus written reasons. Use it to confirm your hero element (logo, product window, key claim) is actually where eyes go. Launch the click test → · See an example →
PickFu click-test heatmap over a body wash bottle packaging design with AI insights

​
3. Benchmark against competitors

SettingValue
Poll typeRanked choice
Question”Based on the packaging, which product would you buy? Why?”
OptionsYour packaging + 2–3 competitor images
AudienceGeneral
Sample size50
What you’ll get: a baseline ranking against the competitive set. Save these competitor images — you reuse them in step 5. Launch the benchmark poll → · See an example →
PickFu ranked poll comparing three body wash packaging designs with vote share and AI insights

​
4. Iterate your design

SettingValue
Poll typeHead-to-head (exactly 2 options)
Question”Which packaging design do you prefer and why?”
OptionsVariation + original design
AudienceGeneral
Sample size15
Test one major change per iteration. If you change the color scheme and the logo placement in the same variation, a win won’t tell you which one worked.
See an example iteration →

​
5. Re-validate against the original competitor set

SettingValue
Poll typeRanked choice
Question”Based on the packaging, which product would you buy? Why?”
OptionsWinning design + the same 2–3 competitor images from step 3
AudienceGeneral
Sample size50 (bump to 100–200 before an expensive print run)
Interpret the result: a higher ranking/score than your step-3 baseline confirms the new design out-performs against the real competitive set, not just against your old design. See an example validation →

​
Troubleshooting

This is the most valuable failure to catch early. Before iterating on aesthetics, fix legibility: larger product name, a clearer product window or photo, or an explicit category descriptor. A beautiful package that doesn’t communicate the product loses on the shelf.
If eyes land on a decorative element instead of your key claim or brand, increase the visual weight of what matters (size, contrast, position) and de-emphasize the distraction. Re-run the click test to confirm the shift.
You improved on your old design but didn’t beat the category. Look at what the winning competitor does that you don’t — often a clearer benefit claim or stronger color contrast — and address that specifically.