Repeat Purchase Frequency Analysis

Goal of the analysis:

Quantify how often customers return to buy again, how quickly they do so, and which factors materially influence repeat behavior. This analysis produces core metrics (repeat purchase rate, time-to-second purchase, purchase frequency, interpurchase interval), segmented by customer, product, and channel, and links them to value outcomes (LTV, margin). For executives, it indicates the health of retention and loyalty, identifies levers to lift profitable repeat (assortment, experience, lifecycle marketing, pricing/loyalty), and informs revenue forecasts and budget allocation across acquisition vs retention.

Data required:

• Transactions and customer identity:
  • Order-level data: customer ID, order ID, order date/time, items/SKUs, quantities, net revenue, discounts, returns/cancellations, gross margin.
  • Channels and touchpoints: e-commerce, app, POS, marketplace; device, store/region.
  • Customer attributes: acquisition date/channel, geography, persona/segment, loyalty tier, cohort tags.
• Product and catalog attributes:
  • SKU category, replenishment vs durable, pack sizes, subscription availability, seasonality/expiry.
  • Assortment changes, stockouts, substitutions.
• Marketing and engagement signals:
  • Campaign exposures, emails/SMS sends and clicks, app push, retargeting impressions.
  • Loyalty program activities: points earn/burn, tier changes, offers redeemed.
• Experience and service data:
  • Fulfillment speed, delivery reliability, return experience, support tickets, CSAT/NPS.
  • Pricing and promo calendars.
• Benchmark and historical context:
  • Prior cohort repeat curves, category demand cycles, holidays/events.
  • CLV models, RFM scores, churn probabilities (if available).
• Governance and privacy:
  • Identity resolution rules, consent status, data retention, deduplication logic across channels.

Detailed step-by-step instruction on how to conduct the analysis:

1. Define scope and key metrics.
   • Observation windows: 30/60/90/180/365 days post-acquisition and rolling monthly.
   • Repeat Purchase Rate (RPR) = customers with ≥2 purchases ÷ customers with ≥1 purchase (in window).
   • Purchase Frequency (PF) = total orders ÷ active customers (in window).
   • Time to Second Purchase (T2) = median days from first to second order among converters.
   • Interpurchase Interval (IPI) = median/mean days between consecutive orders; variance to assess regularity.
   • Reorder Rate by SKU/category = orders with prior purchase of same item/category ÷ eligible orders.
2. Assemble and cleanse data. Pull orders from POS/e-comm/app; unify customer IDs (CDP/CRM), deduplicate accounts, net out returns/cancellations, normalize timestamps and currencies, tag acquisition date/channel and cohort month.
3. Create cohorts. Group customers by acquisition month/quarter and by acquisition channel/offer. Exclude incomplete cohorts for short windows (e.g., require 90 days of aging for 90-day metrics).
4. Compute core repeat metrics.
   • For each cohort and window, compute RPR, PF, T2, and IPI.
   • Build cumulative repeat curves: share reaching 2nd, 3rd, 4th purchase over time.
   • Construct frequency distribution: % of customers with 1, 2, 3–4, 5–9, 10+ orders.
5. Segment analysis. Cut metrics by persona, region, channel, device, loyalty tier, product category (replenishable vs durable), price band/AOV, and promotion exposure. Identify segments with markedly higher/lower RPR and T2.
6. Product and basket lens. Compute reorder rates and typical replenishment cycles by SKU/category using IPI distributions. Identify “gateway” SKUs (high T2 improvement) and cross-sell pathways (e.g., cat litter → litter liners within 30 days).
7. Hazard and propensity analysis (optional).
   • Hazard rate: probability of next purchase as a function of days since last order; use survival analysis to plot hazard by segment.
   • Predictive models (BG/NBD or Pareto/NBD) to estimate expected frequency and active probability; use for forecasting and LTV.
8. Experience driver diagnostics. Regress cohort-level RPR/T2 on operational variables (delivery SLA, out-of-stock rate, return friction, CSAT/NPS) and marketing variables (email cadence, loyalty status). Rank elasticities.
9. Campaign and lifecycle effectiveness. Measure lift in repeat when customers receive replenishment reminders, win-back flows, or loyalty bonuses vs control (where tests exist). Compute incremental repeat per 1,000 messages.
10. Financial linkage. Convert frequency improvements to value: CLV ≈ Σ(expected orders × expected margin per order). Report LTV uplift from moving T2 earlier (e.g., T2 improvement of 15 days → +X orders/year at constant IPI).
11. Prioritize actions and simulate impact. Model expected RPR/PF changes from top drivers: reduce OOS by 30%, add subscription option for top 50 replenishable SKUs, enrich loyalty benefits, or adjust email cadence. Estimate incremental orders, margin, and ROI.
12. Governance. Document metric definitions, cohorting choices, return handling, and model assumptions. Establish a quarterly cadence and ownership across Marketing, Merchandising, and Operations.

Format of the output of analysis:

• Executive summary: RPR, PF, T2, IPI by latest cohorts; top segments and drivers; projected LTV lift and ROI from recommended actions.
• Cohort charts: cumulative repeat curves (2nd/3rd purchase) by acquisition month and channel.
• Frequency distribution and IPI visuals: histograms/box plots by segment and category.
• Product lens: reorder rates and typical replenishment windows by SKU/category.
• Driver/impact slides: regression elasticities, campaign lift results, and scenario simulations.
• Operational dashboard: OOS, delivery SLA, returns, CSAT vs RPR/T2 for accountability.

How to interpret results:

• High RPR with short T2 and stable IPI: Strong habit formation and reliable experience; scale acquisition into these segments/SKUs and protect service levels.
• Low RPR but short T2 among converters: Early promise, but later churn—investigate product quality/assortment and post-purchase engagement beyond the second order.
• Long T2 and high IPI variance: Friction or irregular need states; improve onboarding, reminder timing, and cross-sell relevance.
• Promo-driven repeat spikes: If repeat occurs only during heavy discounting and margin erodes, shift to loyalty value, bundling, and convenience benefits.
• Category/channel differences: Replenishables (beauty, pet, grocery) should show higher RPR and shorter IPIs than durables (appliances, furniture). App channels typically outperform web for frequency if onboarding is effective.
• Operational impacts: Elevated OOS, slow delivery, or high return friction correlates with longer T2 and lower RPR—prioritize supply chain and service fixes.
• Benchmark lens: Favor internal cohort benchmarks; compare to prior-year cohorts and top quartile segments rather than absolute external numbers.

Steps a company can take to improve on this measure:

• Lifecycle marketing and triggers:
  • Deploy replenishment reminders based on observed IPI (e.g., day 20–25 for a 28-day SKU) via email/SMS/push.
  • Design post-first-purchase journeys (how-to content, tips, accessories) to accelerate T2.
  • Win-back flows using value-led nudges (bundles, curated picks) timed to hazard-rate peaks.
• Loyalty, subscriptions, and value:
  • Introduce subscribe-and-save for top replenishable SKUs; offer flexible cadence and easy skips.
  • Enhance loyalty earn/burn, milestone perks, and personalized offers tied to next-best purchase.
• Assortment, availability, and UX:
  • Reduce stockouts on high-repeat SKUs; expand pack sizes and bundles aligned to observed consumption.
  • Improve mobile/app UX and 1-click reordering; surface past purchases and “buy again” modules prominently.
• Experience and service quality:
  • Accelerate fulfillment and increase delivery reliability; communicate proactively on delays.
  • Simplify returns and exchanges; add satisfaction guarantees for durables to encourage repeat confidence.
• Pricing and promo discipline:
  • Shift from blanket discounts to targeted loyalty credits and bundles that protect margin.
  • Use price anchors and thresholds (free shipping at profitable baskets) to raise AOV with repeat.
• Data and modeling:
  • Adopt BG/NBD or Pareto/NBD for expected frequency and active status; feed outputs to CLV and targeting models.
  • Instrument event capture (SKU reorder flags, reminder exposure) and maintain clean cohort tracking.
• Scenario guidance:
  • If T2 is long but IPI is short thereafter, focus on post-first-order onboarding and timed reminders.
  • If RPR is high in loyalty tiers only, broaden benefits for non-members and streamline enrollment at checkout.
  • If repeat depends on heavy promos, pivot to subscriptions, bundles, and convenience to sustain frequency profitably.

Benchmark comparisons:

General benchmarks:

• 12-month Repeat Purchase Rate in transactional retail often ranges 25–60%, higher for consumables and lower for durables.
• Median Time to Second Purchase (T2) commonly 20–60 days for replenishable categories; 90–180+ days for durables.
• Healthy programs see 30–50% of purchasers place 3+ orders within 12 months in consumables; 10–25% in durables.
• Subscriptions materially lift frequency; target 60–80% on-time renewal across cycles for top SKUs.

Segment- or industry-specific benchmarks:

• Grocery/CPG/Pet/Beauty: High-frequency; IPIs of 2–6 weeks are common; repeat >50% at 12 months achievable with strong availability and reminders.
• Apparel/Footwear: Moderate frequency; 12-month repeat 30–50% for engaged bases; seasonality drives spikes.
• Electronics/Appliances/Furniture: Low frequency; focus on accessories, services, and protection plans to create repeat opportunities.
• When external norms are unreliable, build internal benchmarks: compare last 4–6 acquisition cohorts by channel/category, set top-quartile targets for RPR and T2, and manage to cohort-based improvement goals (+3–5 pts RPR, −10–20% T2) each quarter.