💡 Customer Market Visit Memory & Seasonal Re-Buy Reminders

[github-actions[bot]]

Summary

After each market visit, customers can one-tap log which vendors they bought from (selected from their followed vendors list). Over time, this builds a personal purchase journal with seasonal patterns. The app sends re-buy reminders when products return to season: "You bought peaches from Green Valley Farm last August — they're back at Riverside Market this Saturday!" This creates a memory-based retention loop without the overhead of a formal loyalty program.

Market Signal

Personal food tracking apps (Yummly, Whisk, Mealime) have grown significantly, and seasonal eating is a mainstream consumer trend. No farmers market app offers purchase memory or seasonal re-buy reminders. The retention mechanism is proven at massive scale in e-commerce — Amazon's "Buy Again" drives 35%+ of repeat purchases. Applying this pattern to local food creates a powerful loyalty loop. The SARE-funded loyalty program research confirms that farmers market loyalty programs increase visit frequency, but current programs are paper-based punch cards. Digital memory-based retention requires zero vendor participation — it's purely customer-side.

User Signal

The PRD identifies "weekly returning customer rate: 10%" as a KPI. Purchase memory directly drives return visits by reminding customers of positive past experiences tied to specific vendors and seasons. This creates a personalized discovery surface that improves with each visit — a retention moat no competitor offers. Discussion #363 (Digital Market Visit Rewards) covers incentive-based loyalty with a points/rewards model; this covers memory-based retention — a complementary approach that doesn't require vendor buy-in or economic incentives. The two approaches can coexist and reinforce each other.

Technical Opportunity

The architecture already tracks customer follows (FR25-26) and vendor check-ins (FR14). Purchase logging extends this with a lightweight junction table (customer_id, vendor_id, market_day_id, products[]). The seasonal reminder system cross-references vendor product catalogs (FR12) with historical purchase dates — when a product category last purchased N months ago reappears in a followed vendor's catalog, trigger a reminder via FCM. The geofence infrastructure (if Discussion #385 ships) can trigger post-visit prompts on market departure. No new backend infrastructure required — just a new domain model and a Cloud Scheduler job for seasonal reminder checks.

Assessment

Dimension	Score	Rationale
Feasibility	high	Extends existing data models (follows, check-ins, product catalogs); no new infrastructure; lightweight junction table
Impact	med	Drives retention and repeat visits (directly supports KPI); but value accrues slowly over multiple seasons
Urgency	low	No external signal demanding this NOW; value compounds over time as purchase history grows

Adversarial Review

Strongest objection: Customers won't bother logging purchases — it adds friction after a pleasant market experience, and the re-buy value only materializes months later when the season returns.

Rebuttal: The logging is one tap from a post-visit prompt: "You were near Riverside Market today — tap vendors you bought from." The customer's followed vendor list provides the selection set — no typing required. No product-level detail needed for MVP — just vendor-level purchase memory. The delayed-gratification concern is valid for seasonal reminders, but early value comes from short-cycle reminders: "Green Valley Farm is at Riverside Market tomorrow — you bought from them 2 weeks ago." The first re-buy reminder creates an "aha moment" that validates the logging behavior and drives continued engagement.

Suggested Next Step

Design the post-visit prompt UX — evaluate geofence departure trigger (requires Discussion #385) vs. time-based prompt (evening of market day) vs. manual "Log visit" button. Define the purchase memory data model (market_visit_log junction table). Prototype the seasonal matching algorithm using vendor product catalogs and historical purchase dates.