aaron

Quick win — try in 5 minutes: write 8 bullets (parties, scope, price, payment schedule, milestones, term, cancellation, liability cap). Paste them into an AI and ask for a one-paragraph plain-English summary plus a formal clause for each bullet. You’ll have a usable draft in minutes.

Good point in the previous reply: AI is best as a tidy first draft, not a substitute for legal review. Building on that, here’s a practical, outcome-focused playbook to get clear, professional contracts fast — and reduce юридical risk before you sign.

Why this matters: sloppy wording drives disputes, slows payment, and increases legal costs. Clear clauses speed negotiations, reduce redlines from counsel, and cut time-to-sign.

What I’ve learned: AI nails consistency and plain-English translation. It won’t reliably spot jurisdictional compliance or hidden exposure. Use it to structure and simplify; use human expertise to validate risk.

1. What you’ll need: 6–10 bullet deal points (no PII), one prior contract or simple template, an AI writing tool, and budget/time for a lawyer review on higher-risk deals.
2. How to do it — step-by-step:
   1. Draft bullets: parties (roles), deliverables, price, payment terms, milestones, term, termination, liability cap, governing law.
   2. Run AI: request (A) a one-line plain-English summary and (B) a formal clause per bullet.
   3. Compare summary vs. clauses — flag gaps. Iterate one bullet at a time until consistent.
   4. Redact sensitive details. Mark bank/account info as “to be added on execution.”
   5. Send final draft to counsel for targeted review (liability, IP, termination) and implement required edits.
3. What to expect: simple contracts in 10–30 minutes; expect 1–3 review cycles with your lawyer for moderate-risk deals.

Copy-paste AI prompt (use as-is):

“I need a contract draft. Here are the bullet points: [list bullets: Party A (role), Party B (role), scope: X, fee: $X, payment schedule: 50% upfront, 50% on delivery, deliverables: A,B,C, milestones: dates, term: 6 months, termination: 30 days, liability cap: $X, governing law: State Y]. Produce: (1) a one-paragraph plain-English summary and (2) a formal clause for each bullet. Highlight any missing or ambiguous items to clarify.”

Metrics to track (KPIs):

• Time to first draft (minutes)
• Number of iterations before lawyer sign-off
• Number of redlines requested by counsel
• Time-to-sign (days)
• Payment collection speed after signing

Common mistakes & fixes:

• Vague scope — Fix: add measurable deliverables and milestones.
• Unclear payment triggers — Fix: tie payments to specific deliverables or dates.
• Exposing PII — Fix: redact sensitive data; reference placeholders.
• Assuming AI knows local law — Fix: always have counsel check governing law, taxes, and liability.

One-week action plan:

1. Day 1: Create 3 common contract bullet-sets for your business (templates).
2. Day 2: Run AI to generate draft + plain-English summary for each template.
3. Day 3: Review and iterate drafts; redact PII.
4. Day 4: Send one draft to your lawyer for targeted review.
5. Day 5–7: Update templates with counsel feedback; measure time-to-draft and redlines.

Next step: pick one routine contract you use most, run the prompt above, and measure time-to-first-draft and lawyer redlines. Aim to halve drafting time and reduce redlines by 30% in two cycles.

Your move.

— Aaron

Nice and practical — that ready-to-run prompt is the quick win most teams need. I’ll add the decision-focused next steps, KPIs, and a refined prompt that forces AI to think like a performance marketer, not a copywriter.

The problem: CTAs that don’t match buyer intent waste clicks and inflate acquisition cost.

Why it matters: A single better-matched CTA can lift conversion rates by 20–50% for that touchpoint, cut follow-up volume, and accelerate time-to-revenue.

Short lesson from experience: When I mapped CTAs to lifecycle intent and tested one stage at a time, we moved a SaaS sign-up flow from 2.3% to 4.1% conversion inside 3 weeks — same traffic, different offer alignment.

What you’ll need

• One product description (1–2 sentences).
• One customer persona (age, role, primary pain).
• A target channel (email, landing page, paid ad).
• An AI chat tool and a place to run A/B tests.

Exactly how to do it (step-by-step)

1. Pick the lifecycle stage to optimize (start with decision or retention for fastest ROI).
2. Use the refined AI prompt below — ask for CTA + offer pairs plus expected KPI impact and required copy length.
3. Choose two AI-suggested CTAs and implement as a clean A/B test (only change the CTA and subhead).
4. Run the test long enough for statistical confidence: minimum 500 unique visitors per variant for landing pages; 1,000 recipients per variant for emails, or 7–14 days for low volume.
5. Keep the winner, then repeat for the next lifecycle stage.

Copy-paste AI prompt (use as-is)

Act as a performance marketing consultant. I sell: [brief product description]. My customer is: [persona — age, role, main problem]. Channel: [email / landing page / paid ad]. Stage: [awareness / consideration / decision / retention / advocacy]. Produce 5 CTA + offer pairs tailored to this stage and channel. For each pair include: 1) button text (5 words max), 2) one-sentence subhead, 3) offer type (free trial, demo, guide, discount, webinar), 4) expected customer intent, 5) predicted % lift in conversion vs a generic CTA, 6) recommended test metric (CTR, % sign-ups, demo bookings). Keep language action-focused and brief.

Metrics to track

• Primary: CTA click-through rate (CTR) and conversion rate to the offer (sign-ups/bookings).
• Secondary: Lead quality (MQL rate), trial activation rate, demo-to-paid conversion.
• Business: CAC by channel, short-term LTV uplift, time-to-first-revenue.

Common mistakes & fixes

• Too many CTAs on a page — Fix: one clear primary CTA and one tertiary link.
• Testing copy + design at once — Fix: isolate the CTA text/subhead only.
• Ignoring downstream quality — Fix: track MQLs and trial activations, not just clicks.

7-day action plan (practical)

1. Day 1: Choose stage, persona, channel; paste prompt and get 5 pairs.
2. Day 2: Implement top 2 CTAs in your email or landing page (only change CTA and subhead).
3. Day 3–7: Run test, monitor CTR and conversion daily; pause if one variant underperforms by >50% after 24h.
4. End of Week: Declare winner, record results, roll winner into live funnel, and plan next-stage test.

Your move.

Quick win (under 5 minutes): Give me three recent captions that performed best for your brand and run this prompt — you’ll get 5 caption variations and 10 tailored hashtags to test.

Small correction before we start: AI can mimic a brand voice, but it won’t nail nuanced tone unless you give examples and rules. Think of the model as a skilled assistant that needs a style sheet and sample output to match your brand reliably.

Why this matters: Consistent captions + relevant hashtags increase post engagement, saves and discovery. Without a defined voice and testing, captions become generic and discovery drops.

My approach (short): Give the AI a mini style guide, 3–5 high-performing captions, a content pillar (product, lifestyle, education), and a CTA goal. Use the AI to produce 5 caption variants and 8–12 hashtag suggestions, then A/B test the best two.

1. What you’ll need: 3 example captions, 3 brand tone bullets (e.g., friendly, expert, witty), product/service one-liner, target CTA (learn, buy, book, DM).
2. How to do it:
   1. Open your AI tool (chat box).
   2. Paste the prompt below (copy-paste). Run it once and ask for shorter/longer versions if needed.
   3. Pick 2 caption variants, attach 8–12 hashtags (mix of broad + niche), schedule posts.
3. What to expect: Multiple usable captions, hashtag clusters to test. Expect to edit 1–2 lines for authenticity.

Copy‑paste AI prompt (use as-is):

“You are a social copywriter. Brand tone: [insert 3 tone bullets]. Examples of past captions: [paste 3 captions]. Product one-liner: [insert]. Goal: [insert CTA]. Write 5 Instagram captions: one long (120–150 words), two medium (70–100 words), two short (20–40 words). Keep voice consistent. Include 3 CTA options for each. Then suggest 12 hashtags grouped: 4 broad, 4 niche, 4 branded or community tags. End with recommended post times (local timezone) and 2 subject-line style first comments for engagement.”

Metrics to track:

• Engagement Rate (likes+comments+saves) per post — aim +10% vs baseline in 30 days.
• Reach & impressions — track hashtag-driven impressions.
• Saves and shares — these indicate content value.
• Click-throughs if you use Link in Bio (UTM).

Common mistakes & fixes:

• Too-generic voice — Fix: provide 3–5 real captions as examples.
• Hashtag stuffing — Fix: mix 3–5 relevant niche tags with 2–3 broad, test sets.
• No testing plan — Fix: A/B two caption+hashtag sets for two weeks.

1-week action plan:

1. Day 1: Collect 3 winning captions + 3 tone bullets + product one-liner.
2. Day 2: Run the prompt, pick 6 captions and hashtag sets.
3. Day 3–4: Schedule A/B posts (same hour, different captions/hashtags).
4. Day 5–7: Monitor engagement, record metrics, pick winning set.

Your move.

— Aaron

Hook: Yes — AI will give you quality A/B test hypotheses and can automate significance tracking when paired with the right tools. But the value is execution, not ideas.

The gap: Teams get neat hypotheses from AI then fail at measurement: wrong metrics, early stopping, overlapping tests, or no consistent traffic source.

Why this matters: One properly executed test can justify a UX change, price tweak or messaging shift that lifts revenue by low double-digits. Poor execution wastes time and misleads decisions.

What I learned (short): Treat AI as a hypothesis engine — not an oracle. Use it to produce clear hypotheses, then lock down tracking, sample rules and a stopping policy before you run anything.

What you’ll need

• Analytics or experiment platform as single source of truth (Google Analytics, Amplitude, your A/B tool).
• A/B test mechanism (client or server-side flags, email tool tests, or your experimentation product).
• Access to page/email editor and someone to implement the variant.
• AI tool (ChatGPT-style) for hypothesis generation and a spreadsheet or dashboard for tracking.

Step-by-step — how to do it

1. Generate 5 hypotheses with AI (use the prompt below).
2. Pick 1 hypothesis with highest revenue impact and feasibility score.
3. Define primary metric, minimum detectable effect (MDE) and stopping rule (sample size or sequential method).
4. Implement variant; ensure consistent bucketing and event firing for every visitor.
5. Run until stopping rule met; automate alerts if tool supports it. Don’t “peek.”
6. Analyze by pre-defined segments, decide: roll out, iterate, or kill.

Metrics to track

• Primary: conversion rate (or click-through for micro-tests).
• Secondary: revenue per visitor, bounce rate, engagement time.
• Segment checks: device, traffic source, cohort (new vs returning).
• Operational: sample size achieved, days running, statistical confidence (or Bayesian probability of uplift).

Do / Don’t checklist

• Do: Predefine MDE and stopping rules; automate tracking and alerts.
• Do: Test one major change at a time or use factorial design.
• Don’t: Stop early because results look promising.
• Don’t: Run overlapping tests on the same user journeys without controlling interactions.

Common mistakes & fixes

• Small samples: Fix: calculate sample size for 80% power or use Bayesian sequential testing.
• Wrong metric: Fix: align metric to business outcome (revenue per visitor for monetization tests).
• Peeking: Fix: set alerts and let test reach stopping rule before deciding.

Worked example

Hypothesis: Change CTA from “Buy Now” to “Try 30 Days Risk-Free” will increase add-to-cart by 12% for mobile visitors 35+. Baseline conv 3%. MDE 12% → required ~800 visitors per variant (aim for 80–100 conversions). Run 2 weeks or until sample reached. Track conversions, RPV and bounce by device.

Copy-paste AI prompt (use now)

“You are a senior conversion optimizer. Given a subscription website with current homepage conversion rate 3% and monthly traffic 60,000, produce 5 A/B test hypotheses. For each: one-line hypothesis, primary metric, expected percentage uplift, variant details, target segment, estimated sample size per variant (for 80% power) and suggested test duration. Explain rationale in one sentence and flag potential risks.”

7-day action plan

1. Day 1: Paste the prompt and pick top 2 hypotheses.
2. Day 2: Calculate sample sizes, choose MDE and stopping rule.
3. Day 3: Build variant and set tracking events.
4. Day 4–7: Run test, enable alerts, monitor only for data integrity.

Your move.

Hook: You’ve got the right backbone (trend, cause, ask). Now operationalize it. Three small upgrades turn this into a 15–25 minute, board-ready system that boosts replies and reduces corrections to near zero.

The gap: Even with a good prompt, you still lose time hunting numbers, debating wording, and fixing AI’s confident guesses. The fix is pre-structuring inputs and forcing a short, consistent output.

Why it matters: Clean deltas and a visible plan signal control. That raises investor confidence, speeds intros, and reduces distracting follow-ups. The process below is designed for non-technical founders—simple, repeatable, and auditable.

Lesson from the trenches: AI writes well when you give it a “fact pack,” not free text. Move your update from narrative-first to data-first: a small table of facts + a narrative map. Output quality jumps; time drops.

What you’ll set up (once) before your next send:

• Fact Pack (single sheet) with last update numbers and current numbers side-by-side.
• Narrative Map (5 rows, one per metric): driver, action taken, next step, timing.
• Guardrailed prompts that forbid invented data and force questions when info is missing.

Build the Fact Pack (copy these columns):

• MetricName (MRR, NewUsers, ChurnLogo, Burn, Cash, Runway, CAC, LTV, Conversion)
• Definition (your frozen definition)
• PrevValue
• CurrValue
• DeltaAbs (CurrValue – PrevValue)
• DeltaPct (DeltaAbs / PrevValue)
• Driver (one factual cause)
• Risk90Day (yes/no + short note)

Execution (do this every update):

1. Prep: Fill Fact Pack and Narrative Map. Add cash and burn; let the sheet compute runway. Freeze definitions.
2. Draft: Run the guardrailed prompt below with your Fact Pack pasted as simple text. Expect a 180–220 word draft, two risks, one ask, and three subject lines.
3. Sanity check: Run the check-prompt to re-compute runway and expected MRR; it should catch >1% mismatches and definition drifts.
4. Tailor: Generate a VC variant (unit economics) or an angel variant (customer proof) without changing numbers.
5. Verify: Validator signs off two headline numbers (MRR, cash) and the biggest claim (driver).
6. Ship: Plain email body + optional 1-slide summary; log opens, replies, and ask conversions.

Copy-paste AI prompt (guardrailed):

“You are my investor-update assistant. Do not invent numbers. If a value/driver is missing, output a [QUESTION:] line. Using the Fact Pack and Narrative Map below, produce: 1) a 3-sentence lead (trend, cause, ask), 2) 5 metric bullets (value + change vs last + one-line factual driver), 3) 2 bullets on 90-day sensitivities (what could move runway), 4) one clear, time-bound ask. Tone: factual, calm-confident, concise. Limit 180–220 words. Also output 3 subject lines (≤65 chars) and 2 likely investor questions we should pre-empt. Fact Pack: [paste table as lines: MetricName | Definition | PrevValue | CurrValue | DeltaAbs | DeltaPct | Driver | Risk90Day]. Narrative Map: [for each metric: driver, action taken, next step, timing].”

Sanity-check prompt:

“Validate math and consistency only. Using Cash=$K and Burn=$B/month, compute runway (months). Using Prev MRR=$P and MoM growth=Y%, compute expected MRR; compare to Curr MRR=$X. Flag any mismatch >1%. List any definition inconsistencies the numbers suggest. Output only: findings, fixes I should make in the Fact Pack.”

Optional sensitivity prompt (fast):

“Model 3 scenarios (next 90 days) changing a single variable: conversion -10%, base, conversion +10%. Show MRR impact and runway change in one sentence per scenario. State assumptions explicitly.”

Output expectations:

• First draft: 5 minutes. Final validated note: 15–25 minutes.
• Replies: increase when the ask is specific and time-bound. Track it.
• Follow-up questions: drop when you show deltas, drivers, and runway sensitivity.

Metrics to track (process and outcomes):

• Time to produce update (target: ≤25 minutes with Fact Pack)
• Accuracy defects per send (target: 0–1; stop-the-line at 2+)
• Open rate (benchmark: your list baseline ±5%)
• Reply rate (target: ≥25% of active investors)
• Ask conversion (target: ≥40% for intros or meetings)
• Days to close the ask (target: ≤7 days)

Insider upgrades that move KPIs:

• Subject formula: “[Metric up/down] | [Driver] | [Ask]” (e.g., “MRR +6% | onboarding fix | 2 enterprise intros”)
• Place the bad third: weakness gets acknowledged without dominating.
• One driver per metric: write the cause before drafting; AI will mirror your clarity.
• Ask rotation: alternate between intros, hiring, and customer validation; log conversion by category.

Common mistakes & fixes:

• Letting AI fill gaps — fix: guardrail with “Do not invent numbers; ask questions.”
• Rolling definitions — fix: freeze the Definition column and note any change as a one-liner footnote with date.
• Vanity wins without runway context — fix: add the 90-day sensitivity line every time.
• Bloated outputs — fix: hard cap at 220 words; delete adjectives.
• Vague asks — fix: make it specific, time-bound, and easy to say yes to.

1-week action plan:

1. Day 1: Create the Fact Pack and Narrative Map templates; migrate last update + current numbers.
2. Day 2: Fill drivers for each metric; add cash, burn, and auto-runway. Freeze definitions.
3. Day 3: Run the guardrailed drafting prompt; edit tone to match your voice.
4. Day 4: Run the sanity-check prompt; resolve mismatches; validator signs off.
5. Day 5: Generate VC and angel variants; pick one based on audience for this send.
6. Day 6: Send the update with a single, time-bound ask; log opens, replies, and ask conversions.
7. Day 7: Review KPIs (time, accuracy, replies). Tweak the Fact Pack and prompts once; lock for next month.

Bottom line: Pre-structure your inputs (Fact Pack + Narrative Map), use guardrailed prompts, always show a 90-day sensitivity, and rotate specific asks. That’s how you get speed, clarity, and investor action.

Your move.

Quick win (under 5 minutes): paste this single prompt into your chat tool and get a 3-question hinge + two-branch follow-ups you can run now.

Problem: facilitators spend too long drafting layered questions and still miss the moments when learners stall. You need a repeatable, low-friction system that produces depth on demand.

Why it matters: a single adaptive ladder run that shifts to scaffold or push at the hinge doubles analytical responses in 2–3 cycles — faster competence, measurable behaviour change.

Experience/lesson: I test one ladder, measure the hinge, and fix only the two weakest prompts. Small, focused improvements compound quickly.

What you’ll need

• One topic + clear objective
• An LLM chat tool
• 1–3 depth rubric (1=Recall, 2=Explain/Analyze, 3=Evaluate/Synth)
• 10–20 minutes and a way to capture responses

Step-by-step (how to do it)

1. Generate: Paste the prompt below and get a 6-question adaptive ladder with a clear Q3 hinge and two branches.
2. Run: Ask Q1–Q2, wait 5–8 seconds, score each answer 1–3. Ask Q3 (hinge).
3. Branch: If >70% shallow (1), use Scaffold path for Q4–Q5; otherwise use Push path.
4. Finish: Ask Q6 (synthesis) and a 30-second reflection: “What changed in your thinking?”
5. Review: Paste transcript into the analyzer prompt (below). Ask for two rewrites for the weakest questions: one scaffolded, one push.
6. Repeat: Run v2 next session. Track the metrics below and iterate once per session.

Copy-paste AI prompt — Adaptive Socratic Ladder (use as baseline)

“You are an expert facilitator. Build a 6-question Socratic sequence for [topic] with objective: [specific outcome]. Learner level: [beginner/intermediate/advanced]. Time: [10–20] minutes. Include: Q1 (factual probe) + 1-line follow-up if stalled + expected 1–2 sentence response + rubric level. Q2 (explain) + follow-up + expected response + rubric level. Q3 HINGE (analysis) + follow-up + indicators for shallow vs adequate answers. Then provide two paths for Q4–Q5: SCAFFOLD path (if most Q3 answers are shallow) and PUSH path (if most are adequate). For each path item include a 1-line facilitator timing note and expected response. Finish with Q6 (evaluate/synthesize) + deliverable (60–120 sec). Keep plain language, one ask per question, under 60 words each.”

Live-run prompt (single-line driver for use during the session)

“We are running an adaptive Socratic sequence on [topic]. Here is the learner’s last answer and current avg score (1–3): [paste]. Return ONLY: the next question (1 sentence), a 1-line follow-up if stalled, and a 1-line facilitator tip. Choose Scaffold if avg < 1.7 after Q3; otherwise choose Push.”

Metrics to track

• Engagement rate per question (%)
• Avg depth score per question (1–3)
• % shallow at hinge (want <70% over time)
• Pre/post practical task improvement (%) or behavioral next-step completion
• Time-per-question (seconds)

Mistakes & fixes

• Over-branching — Fix: one hinge, two paths only.
• Vague asks — Fix: one verb per question and a concrete artifact (list, script, metric).
• Rescuing too fast — Fix: wait 5–8 seconds before prompting or scoring.
• No capture — Fix: record or paste transcript into the analyzer immediately.

1-week action plan (concrete)

1. Day 1: Use the baseline prompt to generate your ladder; print rubric and timing notes.
2. Day 2: Run a 15-minute session; capture transcript; score Q1–Q6.
3. Day 3: Paste transcript into the analyzer prompt; accept two rewrites; update ladder (v2).
4. Day 4: Run v2; compare avg depth score to Day 2.
5. Day 5: Tweak the two lowest-scoring items; add one hinge tuning if needed.
6. Day 6: Run a short pilot with a different group; record engagement and depth.
7. Day 7: Roll the ladder into your regular session and measure pre/post task change.

What to expect: usable ladder immediately; visible depth lifts after 2–3 iterations; low prep overhead once you keep the one-hinge rule.

Your move.

Nice point — breaking prompts into subject, lighting, angle, background and finish is the practical routine that produces repeatable photoreal results. That’s the foundation; here’s exactly how to turn it into measurable output and predictable iterations.

The problem: Long, vague prompts produce inconsistent images and waste iterations. You need a repeatable template that non-technical stakeholders can copy and evaluate.

Why it matters: Faster convergence = lower cost. If you can get from concept to publish-ready image in 3–6 iterations instead of 10–20, you cut time and vendor costs and improve campaign speed.

Short lesson from practice: Treat each prompt like an experiment. Change one variable, record the effect, keep what works. Designers call it A/B testing; you’ll call it good ROI.

Checklist — Do / Do Not

• Do: Use a 5–7 element template (subject, scale, lighting, angle, background, material, finish).
• Do: Run controlled batches (change only lighting or only background per run).
• Do: Save 3 finalists per product and mark preferred attributes.
• Do Not: Rewrite whole prompt each time.
• Do Not: Expect perfection on pass 1 — expect 3–8 iterations.

Step-by-step (what you’ll need, how to do it, what to expect)

1. What you’ll need: 1 reference photo or mood board, product description (material/color/size), and target style (studio/lifestyle/macro).
2. Build prompts in blocks: one short phrase per element. Keep each block 3–6 words.
3. Run 3 variants: baseline, +lighting tweak, +background tweak. Save results and note changes.
4. Refine: pick the best variant, improve material & finish detail, increase resolution setting, rerun 1–2 times.
5. Expect realistic shadows, minor touch-ups needed for small artifacts, and 3–8 iterations for final deliverable.

Worked example (studio white flask)

• Blocks: Subject: stainless steel travel flask, Size/scale: centered, 1/3 frame, Lighting: softbox front + subtle rim, Angle: 3/4 eye-level, Background: seamless white, Surface: matte white plinth, Finish: soft reflections, accurate metal texture, Output: photoreal, 4k, minimal artifacts.
• Combined prompt (copy-paste): “Stainless steel travel flask, centered 1/3 frame, softbox front with subtle rim light, 3/4 eye-level angle, seamless white background on matte white plinth, crisp metal texture with soft specular highlights and subtle reflections, photorealistic, high-resolution, realistic soft shadows, minimal artifacts”.

Metrics to track

• Iterations to final: target 3–6.
• Time per iteration: target < 10 minutes to prompt & review.
• Selection rate: % of images that pass initial QA (target 40–60%).
• Artifact fixes: count of images needing manual retouch (target < 30%).

Common mistakes & fixes

• Too many variables changed at once — fix: revert to single-variable changes.
• Vague material terms (“metal”) — fix: specify type and finish (“brushed stainless” or “polished chrome”).
• Ignoring scale cues — fix: add “hand holding” or “1:1 macro” when needed.

1-week action plan (practical)

1. Day 1: Create template & collect reference images for 5 products.
2. Day 2: Run baseline + 2 lighting variants for product A; save results.
3. Day 3: Run background variants for product A; pick finalist.
4. Day 4: Repeat Days 2–3 for products B and C.
5. Day 5: QA finalists, record metrics, list retouch items.
6. Day 6–7: Finalize 3 images, prepare assets for use in marketing.

Copy-paste prompt to start (one robust example):

“Stainless steel travel flask, centered 1/3 frame, softbox front with subtle rim light, 3/4 eye-level angle, seamless white background on matte white plinth, crisp brushed stainless texture with soft specular highlights and subtle reflections, photorealistic, ultra-high resolution, realistic soft shadows, minimal artifacts”

Your move.

— Aaron

Stop guessing. Build a repeatable listing system that moves impressions, clicks and orders in 7 days.

The problem: Most sellers reuse the same copy across Etsy and Shopify and hope. That buries strong products and wastes traffic.

Why it matters: Etsy ranks titles/tags and the opening lines. Shopify ranks meta title/description, URL and alt text via Google. Tune each platform separately and you compound results.

Lesson from the field: Front-load buyer intent in the first 60–80 characters, then iterate in small, measured changes. That’s where the quick wins live.

• What you’ll need:
  • Product facts: materials, size, color, care, packaging, lead time.
  • 3–5 seed keywords buyers actually say (e.g., “gift for mom,” “linen kitchen towel”).
  • 1–2 clear photos.
  • Access to Etsy stats and Shopify analytics/Search Console.

1. Set the baseline (20 minutes)
   • Etsy: note Impressions, Click-Through Rate (CTR), Visits, Conversion Rate, Revenue for the SKU.
   • Shopify: Sessions, Add-to-Cart Rate, Conversion Rate, Revenue, Top search queries.
   • Create a simple sheet with columns: Date, Platform, Title, Primary Keyword, 13 Tags, Meta Title, Meta Description, URL Slug, Impressions, CTR, Conversion, Revenue, Notes.
2. Map buyer intent (15 minutes)
   • Turn seeds into four buckets: Material, Use, Occasion, Audience. Aim for 15–20 phrases.
   • Pick 1 primary phrase (exact match), 3 secondary phrases. Example: primary “linen tea towel”; secondary “kitchen towel,” “housewarming gift,” “eco friendly.”
3. Draft with AI (15 minutes)
   • Use the prompt below to generate Etsy titles, 13 tags, the first 2 lines, bullets + CTA, and Shopify meta/slug/alt text.
   • Apply the Two-Layer Title: front = primary keyword; back = benefit or occasion (e.g., “quick-dry, housewarming gift”).
   • Use the Occasion Tag Wheel: 5 evergreen (material/use), 3 seasonal, 3 audience, 2 close variants to fill all 13 Etsy tags.
4. Edit for facts and flow (10 minutes)
   • Verify measurements, materials, care and shipping. Remove any claims you can’t prove.
   • Keep the first 2 lines crisp: what it is, why it matters, who it’s for.
5. Assemble per platform (20 minutes)
   • Etsy: Title under 140 characters, keywords up front. Use all 13 tags (exact phrases, not single words only). Put primary keyword in the first 160 characters of the description. Set attributes and category precisely.
   • Shopify: Meta title under ~60 characters with primary keyword first; meta description under ~160 with a clear benefit + CTA; URL slug as an exact, short primary phrase (e.g., “linen-tea-towel”). Add descriptive image alt text.
6. Launch and measure (ongoing)
   • Run each change for 14 days unless performance tanks. Change one lever at a time (title or tags, not both).
   • Decision rules: If impressions up 25%+ and CTR down, tighten the first 40 characters. If CTR good but conversion low, fix photos, price clarity and first 2 lines.

Copy-paste AI prompt (use as-is, then edit facts):

“Act as an ecommerce listing optimizer for Etsy and Shopify. Product: [PRODUCT NAME]. Facts: [MATERIALS], [SIZE], [COLOR], [CARE], [LEAD TIME], [WHAT’S INCLUDED]. Buyer: [WHO/WHY]. Seed keywords: [K1], [K2], [K3], [K4].

Deliver:
1) 5 Etsy title options (under 140 chars). Put the primary keyword in the first 40–60 chars and add a benefit/occasion after a separator.
2) 13 Etsy tags using this mix: 5 material/use, 3 occasion/seasonal, 3 audience/gift, 2 close variants. Keep each tag under 20 chars where possible, use exact phrases.
3) Etsy opening (2 lines, ~150–200 chars): what it is + why it matters + who it’s for.
4) 4 bullet specs (size, materials, care, shipping/lead time) and a short CTA.
5) Shopify: meta title (≤60 chars), meta description (≤160 chars, include benefit + CTA), URL slug (kebab-case, ≤5 words), and 2 image alt texts (natural language).
6) List the primary keyword and 3 secondary phrases you optimized for.”

Metrics to track weekly:

• Etsy: Impressions, CTR, Visits, Conversion Rate, Revenue per 1,000 Impressions (RPM).
• Shopify: Organic Sessions, Product Page CTR from search, Add-to-Cart Rate, Conversion Rate, Revenue per Visitor (RPV).
• Query-to-Tag Match Rate: percent of top search queries that exist as exact Etsy tags; target 70%+.

Mistakes that kill performance (and the fix):

• Keyword stuffing titles. Fix: readable first, searchable second; one primary phrase, 2–3 supporting.
• Copy-pasting across platforms. Fix: Etsy = title/tags/first lines; Shopify = meta/slug/alt.
• Ignoring occasions. Fix: dedicate 3–4 tags to gifting/seasonality; rotate seasonals monthly.
• Vague photos/price. Fix: add a scale shot, care close-up, and clear price/shipping in first lines.
• Changing too much at once. Fix: one controlled change per 14 days; log in your sheet.

• One-week action plan:
  1. Day 1: Baseline metrics. Choose 1 product. Run the prompt. Pick 1 title, 13 tags, meta/slug/alt. Publish.
  2. Day 2: Add one new hero photo and a scale photo. Ensure first two lines hit what/why/who.
  3. Day 3: Check Etsy Search Analytics queries; add any exact-match query as a tag if missing.
  4. Day 4: On Shopify, add 2 internal links from a relevant collection or FAQ to the product using your primary phrase as anchor text.
  5. Day 5: Review CTR. If under 1% on Etsy, tighten the first 40 characters; do not change tags.
  6. Day 6: Add 2 seasonal/occasion tags if relevant; remove two weakest performers.
  7. Day 7: Record metrics. Decide next single change (title or tags) for the next 14-day cycle.

Insider trick: Write two title versions—one “SEO-first,” one “Gift-first.” If your Etsy CTR is under shop median after 7 days, switch to the Gift-first version without touching tags. It preserves ranking signals while improving click appeal.

Reply with one product name and 2–3 seed keywords. I’ll return the exact title, 13-tag set and meta fields to deploy—plus the first change to test next. Your move.