How we compute your checkout loss.

We aggregate public commerce signals across 1M+ stores: annual sales, payment technology stack, country, business size. Numbers are computed live per store at request time using the formulas below.

Cart abandonment is the single largest source of leak in e-commerce. Baymard Institute's meta-research puts the average cart-abandonment rate at ~70%. We apply that as a top-of-funnel loss factor on annual sales, then split 70/30 across cart and checkout stages to reflect that pre-payment exit is heavier than mid-checkout exit.

monthly_loss_cart_and_checkout = annual_sales × 0.70 / 12
cart_loss      = monthly_loss × 0.70
checkout_loss  = monthly_loss × 0.30

Payment failures are the differentiated layer. Most operators don't see these dollars: failed transactions disappear from default dashboards without instrumentation. We compute a per-store failure rate from a segment baseline, modulated by the store's payment tech stack. Rates are clamped to [4%, 15%].

base_rate = { starter: 10%, growth: 9%, scale: 7%, enterprise: 5% }

rate += 1%   if no BNPL signal (no Klarna/Affirm/Afterpay)
rate += 1%   if no multi-PSP signal (no PayPal AND Stripe/Adyen/Braintree)
rate -= 0.5% if annual_sales ≥ $5M (Plus-grade stack proxy)
rate  = clamp(rate, 0.04, 0.15)

attempted_revenue       = annual_sales / (1 - rate)
payment_failures_loss   = annual_sales × (rate / (1 - rate)) / 12

Worked example · $5M growth storeBase rate 9% + 1% (no BNPL) + 1% (no multi-PSP) − 0.5% (≥$5M Plus proxy) = 10.5%. Monthly payment-failure loss ≈ $5M × (0.105 / 0.895) / 12 ≈ $48.9K / month.

Each stage has its own close rate, drawn from production performance across our deployments.

Total monthly sales closed = cart_loss × 30% + checkout_loss × 30% + payment_failures × 50%.

• The model treats annual_sales as captured GMV. If your reported number differs, override it on your report page before generating the snapshot.
• Segment medians are aggregates, not store-specific. They're shown as peer context, not as your forecast.
• Payment-failure rates are modeled from segment baselines and tech-stack signals, not real-time auth data. The clamp to [4%, 15%] keeps the model bounded.
• Stores not in our dataset fall back to growth-segment medians and are labelled clearly in the report.