Smartass Prompts donkeySmartassPrompts
← Advanced

Repeat-Customer Cohort Check

ROLE: Act as a customer retention analyst. Compare like-for-like repeat-purchase windows and be honest about incomplete records. INPUTS: - Business, normal purchase cycle and repeat-purchase definition: {business and definitions} - Analysis cutoff date, timezone and chosen window in calendar days: {cutoff and window} - Anonymized transactions with customer ID, transaction ID, date, qualifying net value and refund status: {transaction records} - Earliest reliable history date and exclusions such as tests or canceled orders: {history limits and exclusions} - Optional customer segment or acquisition source: {segments} PROCESS: 1. Check duplicates, date formats, refunds and missing customer IDs. Establish what counts as a qualifying purchase. If historical first purchases are unknown, call the cohorts first-observed cohorts, not new-customer cohorts. 2. Group customers by the month of their first qualifying purchase in the supplied history. For the chosen window, include in the denominator only customers whose entire window has elapsed by the cutoff date. 3. Count an eligible customer's repeat only when a later qualifying transaction occurs after the first purchase and on or before first purchase plus the window. Compute repeat rate as repeaters divided by eligible customers. Report immature customers separately rather than labeling them lost. 4. Show cohort counts, eligible customers, repeaters, repeat rate and known net sales within the window. Explain exclusions and show calculations; mark unavailable results unknown. 5. Compare only mature windows. Flag small samples, seasonality, differing segments and incomplete earlier history. Do not infer why people did not return from transaction dates alone. 6. Suggest three questions for customer research and one consent-respecting retention experiment. Treat reasons for change as hypotheses. OUTPUT FORMAT: Definitions and Data Limits; Cohort Table; Comparable Patterns; Unknown Causes; Next Research and Experiment. QUALITY CHECKS: No personal contact details. Do not claim lifetime value from a short window or mix repeat purchasing with subscription renewal. Do not treat correlation as cause. Reconcile customers and transactions separately.

Tip: fill in the {brackets}

Swap every highlighted bit for your real details before you paste. The more specific you are, the less generic the answer.

What it does

Calculates comparable repeat-purchase rates and separates mature customer windows from records that are too recent to judge.

When to use it

Use when repeat business matters and you have dated, anonymized transactions with enough history for your chosen comparison window.

FAQ

Can it identify why customers did not return?

No. Transactions show purchase patterns, not customer motives; the prompt suggests research questions instead.

What happens to recent customers?

Customers whose full observation window has not elapsed are reported separately and are not counted as failures to return.

More Advanced prompts

AdvancedTested

Offer Teardown and Rebuild

ROLE: You are a direct-response offer strategist who has rebuilt offers for hundreds of local service businesses. You are blunt, specific and allergic to fluff. INPUTS (fill in every blank you can; leave a blank empty rather than guess): - Business: {business} (e.g., Advanced Air Conditioning + Heating, an HVAC company in Cedar Park, TX) - The offer, copied word for word from my site, ad or flyer: {current offer} - Price and terms: {price and terms} - Who buys it: {ideal customer} - What customers usually object to: {objections} - Proof I actually have (reviews count, years in business, licenses, guarantees): {proof} PROCESS: 1. Teardown. Score the current offer 1 to 5 on each: clarity (can a stranger repeat it?), specificity (numbers, deliverables), risk reversal, urgency, proof, and differentiation from a generic competitor. One sentence of why for each score. 2. Diagnose. Name the single biggest reason a ready-to-buy customer still says no. 3. Rebuild three versions, each with a different lever: A. The Clarity Rebuild: same deliverables, rewritten so it is impossible to misunderstand. B. The Risk-Reversal Rebuild: adds or sharpens a guarantee or promise, using only terms I gave you. C. The Bundle or Tier Rebuild: restructures what is included or how it is packaged. For each: a headline (under 12 words), a 2 to 3 sentence offer statement, a bullet list of what's included, and the call to action. 4. Recommend which version to test first and what to measure. OUTPUT FORMAT: Headings exactly: Scorecard (a table), Biggest Problem, Rebuild A, Rebuild B, Rebuild C, Test This First. QUALITY CHECKS (run before answering): - Use only facts, prices, guarantees and numbers I supplied. If a rebuild needs a fact I did not give, write a plain note like (add your price) instead of inventing it. - No phrases like "unlock," "elevate," "seamless" or "game-changer." - Every headline must make sense to someone who has never heard of my business.

Open
AdvancedTested

Review Themes to Marketing Angles

ROLE: You are a customer research analyst who turns review piles into marketing angles. You summarize patterns; you never quote customers. INPUTS: - Business: {business} (e.g., La Dosis Coffee + Cocktails, a coffee shop and cocktail bar in Cedar Park, TX) - What we sell and who we sell to: {offer and audience} - My reviews, pasted below (Google, Yelp, Facebook; 15 or more works best): {paste your reviews} PROCESS: 1. Read every review. Group what people praise and complain about into 4 to 7 themes. Name each theme in plain words. 2. For each theme give: how many reviews mention it (approximate count), whether it's praise, complaint or mixed, and a one-line paraphrased summary. Do not copy any review word for word, and do not include reviewer names. 3. Spot the gap: which strengths do customers mention that my marketing (as I described it) never says? 4. Turn the top 3 praise themes into marketing angles. For each: an angle name, a headline, a one-sentence explanation of who it speaks to, and two places to use it (for example Google Business Profile post, menu board, homepage hero). 5. Turn the top complaint theme into an operations fix, not a marketing spin. OUTPUT FORMAT: Headings: Theme Table, What Your Marketing Is Missing, Three Marketing Angles, Fix Before You Advertise. QUALITY CHECKS: - Zero verbatim quotes, zero names. Paraphrase only. - Counts must come from the pasted reviews. If I pasted fewer than 10, say the themes are a weak signal. - Do not claim anything about the business that isn't in the reviews or my inputs.

Open
AdvancedTested

Local SEO Audit: Profile vs. Website vs. 3 Competitors

ROLE: You are a senior local SEO consultant running a fast, honest audit. You care about what moves map-pack rankings and calls, not vanity metrics. INPUTS (paste what you can see; the AI can't log in to anything): - Business: {business} (e.g., Whitestone Plumbing, a plumber in Cedar Park, TX) - Main service and city I want to rank for: {service + city} - My Google Business Profile details: primary category, secondary categories, review count and rating, hours, services listed, photos count, last post date, description: {GBP details} - My website: homepage title tag, H1, list of service pages, location or service-area pages, NAP (name, address, phone) as shown in the footer: {website details} - 3 competitors that show in the map pack for my search, with the same GBP and website details: {competitor 1}, {competitor 2}, {competitor 3} PROCESS: 1. Build a side-by-side comparison table: me vs. each competitor on primary category, review count, rating, services listed, dedicated service pages, city pages, and NAP consistency. 2. Check alignment between my GBP and my website: does the primary category match my homepage focus? Do the GBP services each have a page? Does the NAP match exactly? 3. Identify where competitors beat me and where I already win. 4. Produce a ranked fix list, ordered by likely impact on local rankings and calls divided by effort. For each fix: what to do, why it matters, effort (low, medium, high), and where (GBP or website). OUTPUT FORMAT: Headings: Comparison Table, Profile vs. Website Mismatches, Where You Win, Where You Lose, Ranked Fix List (numbered, top 10 max). QUALITY CHECKS: - Base every finding on the details I pasted. If a field is missing, write (not provided) in the table and don't assume. - Don't promise rankings or timelines. - No "build backlinks" filler without a specific, local way to do it.

Open
See all Advanced prompts

The Weekly Smart Ass

One useful prompt in your inbox, every Tuesday. No fluff.