communications Intermediate Communications Manager Program Manager ✓ Tested 0.56/10

Impact Statistic Visualization Caption

Caption for data/infographic social post

The Prompt

The Prompt

Impact Statistic Visualization Caption Generator

Purpose: Create platform-ready social captions (with accessibility and optional designer guidance) for nonprofit impact statistics that earn engagement, saves, and shares.

Best for: Program milestones, annual report highlights, campaign updates, funder accountability posts.

STEP 1: Provide Your Information (copy/paste and fill in)
Required
1) Organization: [ORG NAME]
2) Primary stat (number + what it measures): [e.g., “1,247 students served” or “89% graduation rate”]
3) Platform: [Instagram / Facebook / LinkedIn / X]
4) Target audience: [Donors / Funders / Volunteers / Community]
5) Timeframe: [e.g., “in 2024,” “Jan–Jun 2025,” “since 2019,” “as of Oct 2025”]
6) Geography/community: [e.g., “in rural Montana,” “across 12 Chicago neighborhoods”]
7) Source + year/URL: [e.g., “Internal program data, 2024” or “2024 Annual Report, URL”]

Required if using a percentage
8) Base/denominator: [e.g., “89% of 340 participants”]

Optional but recommended
9) Relatable comparison: [e.g., “equal to the population of Boise,” “enough to fill Soldier Field”]
10) Call to action (CTA): [e.g., “Donate monthly,” “Volunteer this weekend,” “Save and share”]
11) Tone preference: [FORMAL / WARM / CASUAL] (default WARM)
12) Staff quote/sign-off: [e.g., “—Maria Chen, Program Director”] (use only if adding a quote)
13) Designer notes needed?: [Yes/No] (default No)

If any required fields are missing, I will ask up to 3 concise clarification questions before generating. If a percentage lacks a base or date, I will convert to an absolute number if possible or ask for details.

STEP 2: Generate the Caption Package

Produce Two Caption Options (A & B)
- Length: 60–80 words each (see platform overrides below)
- Use the SAME primary stat in both options; vary the lead, framing, and CTA.
- A vs. B purpose: Two testable angles of the same fact.
  • Option A: Lead with program/solution impact.
  • Option B: Lead with community/individual perspective.

Required structure (per caption)
1) Opening (insight, not a number): One human takeaway or why it matters.
   - Weak: “We served 1,247 students this year.”
   - Strong: “When students get steady tutoring, confidence and grades rise.”
2) Stat sentence with full context: Number + what + who + when + where (and base if %).
   - Example: “In 2024, 1,247 students across 12 Chicago neighborhoods received free one-on-one tutoring through [ORG].”
3) Why it matters (plain language; no jargon/acronyms).
4) Relatable comparison (if provided or credibly inferrable).
5) Soft, platform-aligned CTA.
6) Source citation at the end: (Source: [SOURCE], [YEAR/URL]).
7) Hashtags: 2–4 relevant, TitleCase, at the very end.

Tone guidance
- FORMAL (funders/LinkedIn): Precise, credible, data-forward. No emojis.
- WARM (donors/community): Empathetic, hopeful, community-centered. Emoji ok (max 1).
- CASUAL (volunteers/Instagram): Friendly, plain language. Emoji ok (max 2).

Platform length and formatting overrides
- Instagram: Keep 60–80 words; front-load the hook in first 100–125 characters; 1–2 line breaks ok.
- Facebook: 60–100 words; front-load key info in first 140 characters; links in comments if needed.
- LinkedIn: 80–140 words; no emojis in FORMAL; use line breaks for scannability.
- X: Hard cap 280 characters; compress each option to ≤240 characters, leaving room for 1–2 short hashtags. Prioritize: insight + stat context + CTA + source abbrev.

Content rules for the statistic
- Always include timeframe and geography/community.
- If using a percentage, include base/denominator and timeframe (e.g., “89% of 340 participants in 2024”).
- Use round numbers unless precision is critical to credibility.
- People-first language; avoid deficit framing and guilt appeals.
- If comparison isn’t credible or may feel minimizing, omit it rather than forcing one.

Relatable comparison ideas (use one if supplied or appropriate)
- Population equivalence (e.g., “equal to the population of [Town]”)
- Venue capacity (e.g., “enough to fill [Landmark]”)
- Everyday scale (e.g., “about one full school bus per week”)

CTAs (select or adapt to platform)
- Instagram: “Save and share,” “Link in bio to donate monthly,” “Volunteer signup in bio.”
- Facebook: “Share this post,” “Donate at the link in comments,” “RSVP to volunteer.”
- LinkedIn: “Follow for data updates,” “Connect to explore funding partnerships,” “Read the full report.”
- X: “Retweet to spread the word,” “See link in thread for details.”

Alt Text (accessibility)
- Write 1–2 sentences, 120–160 characters total.
- Describe the visual style and state the exact statistic text as shown.
- No emojis, hashtags, or CTAs.
- Example: “Bold white text on navy background reads: 1,247 Chicago students received free tutoring in 2024. Small source line at bottom right.”

Designer Notes (include only if Step 1 #13 = Yes)
- Format: 1080×1080 (Instagram), 1200×628 (Facebook/X), 1200×1200 (LinkedIn)
- Visual: Single stat on simple solid background; no chart junk.
- Typography: Large, bold sans-serif; minimum 16–18pt on mobile; test color contrast ≥4.5:1.
- Hierarchy: Stat is hero; context smaller but readable; source bottom-right in 10–12pt.
- Branding: Use org brand colors; small logo in a corner.

Decision rules and conflict resolution
- Prioritize order: Platform constraints > Audience needs > Tone preference.
- If platform = LinkedIn but audience = Volunteers, keep FORMAL data clarity with a community-forward opening; avoid emojis.
- If percentage lacks a base or timeframe, convert to an absolute count or ask for details; do not post a standalone percent.
- If timeframe is ongoing, use “as of [Month Year].”
- If geography unspecified, ask for it; if unavailable, use the smallest accurate scope from the source.
- If comparison is absent or weak, omit it rather than stretching.
- If tone not specified, default to WARM.

Quality checklist (self-check before finalizing)
- Concrete and dignifying; people-first language.
- No deficit framing, guilt appeals, or clichés (e.g., “transform lives,” “change the world”).
- Verifiable: stat includes timeframe, geography, and base if %; cites a credible source.
- Mobile-first: short sentences, scannable, clear hierarchy.
- Platform-appropriate: tone and CTA match the channel.
- Numbers are rounded unless precision is needed.

Mini examples: weak vs strong openings
- Weak: “We served 2,300 meals.”
- Strong: “A dependable hot meal turns a tough day into a doable one.”

- Weak: “Our program reached 412 teens.”
- Strong: “When teens find a safe place to belong, they keep showing up—and keep growing.”

Example output (for reference)
Caption A (WARM, Instagram, donors):
A safe place after school changes the rhythm of a family’s week. In 2024, 1,247 students across 12 Chicago neighborhoods received free one-on-one tutoring through [ORG]. With steady support, kids keep up in class and build confidence that carries home. That’s enough students to fill three elementary schools. Your gifts keep the doors open. 💙 Tap the link in bio to become a monthly donor. (Source: Internal program data, 2024)
#ChicagoEducation #TutoringWorks #CommunityImpact

Caption B (WARM, Instagram, donors):
Imagine being behind in reading and finally getting the right help. In 2024, 1,247 Chicago students accessed free tutoring at [ORG] centers across 12 neighborhoods. Consistent, caring support helps students stay in school and see new possibilities. Picture three full elementary schools’ worth of learners getting a boost. Want to keep this going? Join as a monthly donor—link in bio. (Source: Internal program data, 2024)
#EducationEquity #ChicagoKids #DonorImpact

Alt Text:
Bold white text on navy background reads: 1,247 Chicago students received free tutoring in 2024. Source listed bottom right.

NOW GENERATE
Using the information from Step 1, produce:
- Two caption options (A & B) as specified (apply platform overrides if Platform = X).
- Alt text (1–2 sentences, 120–160 characters).
- Designer notes only if requested in Step 1.

How to Customize

  1. Replace all [BRACKETED] fields with your specific information
  2. Adjust tone and length as needed for your audience
  3. Review and personalize before using

Pro Tips

  1. Test this prompt with your preferred AI tool before using in production
  2. Always review AI output for accuracy and appropriateness
  3. Customize outputs to match your organization’s voice and brand

(See other prompts in the communications category)

Example Outputs

Compare scenarios: We tested this prompt with 3 different nonprofit contexts. Each scenario shows outputs from GPT-5, Claude, and Gemini. Select a model above each scenario to compare.

Small Community Org

Rural food pantry coordinating produce distributions across multiple small towns in central Montana.

View scenario details
[Organization]:Prairie Table Pantry
[Primary stat (number + what it measures)]:2,400 residents received fresh produce
[Platform]:Facebook
[Target audience]:Community
[Timeframe]:in 2024
[Geography/community]:across five towns in Fergus County, Montana
[Source + year/URL]:Internal distribution logs, 2024
[Base/denominator]:N/A
[Relatable comparison]:roughly equal to filling the Fergus High School stadium twice
[Call to action (CTA)]:Share this post
[Tone preference]:WARM
[Staff quote/sign-off]:—Tasha White, Pantry Coordinator
[Designer notes needed?]:No
[ORG]:Prairie Table Pantry
[SOURCE]:Internal distribution logs
[YEAR/URL]:2024
[Town]:Lewistown, MT
[Landmark]:Fergus High School Stadium
[Month Year]:Dec 2024

I’m ready to generate—just need a few details.

1) Please provide: Organization; Primary stat (number + what it measures); Timeframe; Geography/community; Source + year/URL.

2) Which Platform and Target audience? Any Tone preference (FORMAL/WARM/CASUAL)?

3) If the stat is a percentage, what’s the base/denominator? Optional: Relatable comparison, preferred CTA, and whether you want Designer notes.

Generated: Nov 2, 20252395 tokens$0.0122 cost
AI Evaluation Details (Score: 0.5/10)
Reasoning:

The AI completely failed to execute the task. Despite receiving all required information in a structured format, it ignored the provided scenario data and instead asked the user to provide the exact same information again. This represents a critical failure in following instructions and parsing input.

Strengths:
  • The clarifying questions are appropriately structured and cover the necessary fields
  • The tone of the request is professional and organized
Weaknesses:
  • Complete failure to recognize that all required information was already provided in the scenario values
  • Did not generate any of the requested deliverables (Caption A, Caption B, Alt Text)
  • Wasted the user's time by requesting information that was clearly present in the input
  • Shows no evidence of parsing or processing the structured data provided
  • A nonprofit professional would need to re-enter all information, making this output entirely unusable
  • Demonstrates poor AI comprehension of context and task requirements

Mid-Size Professional Org

Urban workforce development nonprofit serving immigrant and first‑generation jobseekers across Queens, NYC.

View scenario details
[Organization]:Queens Pathways Workforce Center
[Primary stat (number + what it measures)]:87% job placement within 90 days
[Platform]:LinkedIn
[Target audience]:Funders
[Timeframe]:Jan–Jun 2025
[Geography/community]:across Queens, New York City
[Source + year/URL]:2025 Midyear Outcomes Dashboard, https://qpwc.org/reports/2025-midyear
[Base/denominator]:87% of 312 graduates
[Relatable comparison]:about nine out of ten participants
[Call to action (CTA)]:Connect to explore funding partnerships
[Tone preference]:FORMAL
[Staff quote/sign-off]:—Rogelio Martinez, Director of Programs
[Designer notes needed?]:Yes
[ORG]:Queens Pathways Workforce Center
[SOURCE]:2025 Midyear Outcomes Dashboard
[YEAR/URL]:2025, https://qpwc.org/reports/2025-midyear
[Town]:Jackson Heights, NY
[Landmark]:Citi Field
[Month Year]:Jun 2025

Thanks! I’m ready to build your caption package. I just need a few details.

1) What’s the organization name, the primary stat (number + what it measures), the timeframe, and the geography/community? If it’s a percentage, what’s the base/denominator?

2) Which platform (Instagram/Facebook/LinkedIn/X) and who’s the target audience (Donors/Funders/Volunteers/Community)? Any tone preference (FORMAL/WARM/CASUAL) and preferred CTA?

3) What’s the source (name/year/URL)? Any relatable comparison to include, a staff quote to sign off, and do you want designer notes?

Generated: Nov 2, 20252330 tokens$0.0110 cost
AI Evaluation Details (Score: 0.5/10)
Reasoning:

The AI completely failed to execute the task. Despite receiving all required information in a structured JSON format in the 'Scenario Used' section, it ignored the data and asked clarifying questions instead. This represents a critical failure in following instructions and parsing provided input.

Strengths:
  • The clarifying questions are logically structured and cover the necessary fields
  • The tone of the questions is professional and organized
Weaknesses:
  • Complete failure to recognize that all required data was already provided in the scenario JSON
  • Did not generate any of the requested outputs (Caption A, Caption B, Alt Text, Designer Notes)
  • Wasted the user's time by asking for information that was clearly present
  • Shows poor data parsing capability - the AI should have extracted values from the structured scenario data
  • A nonprofit professional would receive zero usable content from this response
  • Demonstrates the AI did not read or process the full context before responding

Large Established Org

Regional environmental nonprofit restoring marine habitats throughout New York Harbor.

View scenario details
[Organization]:Harbor Renewal Alliance
[Primary stat (number + what it measures)]:1.2 million oysters restored
[Platform]:Instagram
[Target audience]:Donors
[Timeframe]:as of Oct 2025
[Geography/community]:across New York Harbor and tidal tributaries
[Source + year/URL]:2025 Annual Impact Report, https://harborrenewal.org/impact2025
[Base/denominator]:N/A
[Relatable comparison]:enough to filter about 90 Olympic-size pools of water each day
[Call to action (CTA)]:Link in bio to donate monthly
[Tone preference]:WARM
[Staff quote/sign-off]:—Aisha Khan, Restoration Director
[Designer notes needed?]:No
[ORG]:Harbor Renewal Alliance
[SOURCE]:2025 Annual Impact Report
[YEAR/URL]:2025, https://harborrenewal.org/impact2025
[Town]:New York, NY
[Landmark]:Brooklyn Bridge Park
[Month Year]:Oct 2025

I’m ready—just need a few details to make this accurate and platform-ready.

1) Core impact info - Organization name: - Primary stat (number + what it measures). If a percent, include the base/denominator: - Timeframe (e.g., in 2024, Jan–Jun 2025, since 2019, as of Oct 2025): - Geography/community (e.g., across 12 Chicago neighborhoods):

2) Channel and audience - Platform (Instagram / Facebook / LinkedIn / X): - Target audience (Donors / Funders / Volunteers / Community): - Tone preference (FORMAL / WARM / CASUAL; default WARM): - Preferred CTA (e.g., Donate monthly, Volunteer, Save and share):

3) Credibility and extras - Source + year/URL (e.g., 2024 Annual Report, link): - Relatable comparison (optional): - Staff quote/sign-off (optional): - Designer notes needed? (Yes/No; default No):

Generated: Nov 2, 20252589 tokens$0.0150 cost
AI Evaluation Details (Score: 0.5/10)
Reasoning:

The AI completely failed to execute the task. Despite receiving all required information in a structured scenario format, it ignored the data and instead asked the user to provide the exact same information again. This represents a total failure to parse input, follow instructions, or generate the requested captions.

Strengths:
  • The clarifying questions themselves are well-structured and would be appropriate if information were actually missing
  • The questions are organized logically by category (core impact, channel/audience, credibility)
Weaknesses:
  • Complete failure to recognize that all required information was already provided in the scenario values
  • Did not generate any of the requested deliverables (Caption A, Caption B, Alt Text)
  • Shows no evidence of parsing or processing the structured input data
  • Wastes user time by requesting information that was already supplied
  • Makes the tool completely unusable in this instance—a nonprofit professional would have to re-enter everything manually
  • The prompt explicitly states the AI should ask clarifying questions only if required fields are missing, but all fields were present
  • Demonstrates poor instruction-following and data extraction capabilities

Test Summary: Generated Nov 2, 20253 scenarios • 9 total outputs • Average quality score: 0.56/10 • Total validation cost: $0.0773