Payments are experiencing issues due to temporary restrictions in Russia. If your payment does not go through, please submit a support request.Our support team is available 24/7 — we are always here to help with hosting and server issues.We are now accepting requests for dedicated server rental and colocation services in our data center.Reminder: we recommend enabling backups for additional data protection.A new VPS/VDS lineup with NVMe storage and improved performance is now available.Maintenance work on some servers has been completed. All services are operating normally.
Article6 min readViews2

How much does an hour of checkout downtime cost a store

A training model helps estimate the direct cost of a checkout failure, account for recovered purchases, and determine when it makes sense to pay for faster response.

Comments 0

Stopped clocks, a package, and a payment terminal on the desk — the cost of checkout downtime
In this article

Checkout downtime cannot be honestly valued as the store's hourly revenue. Some customers will return later, others will contact a manager, and revenue does not equal the result after variable expenses. To decide on response speed, you need a range: expected paid orders in a comparable hour, the share of unrecovered orders, the marginal contribution of a single order, and confirmed additional expenses.

In the basic training scenario, 45 minutes of unavailable checkout yields a direct effect of 24,710 rubles: 15,390 rubles in unrecovered marginal contribution from orders, 1,320 rubles in team time, and 8,000 rubles in external emergency work. This is not a forecast for a specific store. The numbers show the order of calculation, which must be replaced with your own data.

First, define the calculation unit

Do not use the annual average; instead, take several comparable periods: the same day of the week, hour, channel, and campaign type. If only payment stopped, do not attribute all site orders to the failure. If the entire store is unavailable, do not transfer statistics from a single funnel step to it.

A convenient base unit is expected paid orders per hour. It already separates browsing from purchasing. For a seasonal promotion, a range, such as 14–22 orders per hour, is more useful than a single precise figure from a quiet week.

Direct impact formula

For a single incident, we use a training formula: direct impact = orders per hour × outage duration in hours × share of unrecovered orders × marginal contribution per order + team time × internal hourly cost + confirmed external expenses.

  • Orders per hour — the median or range for comparable periods, not the best hour of the month.

  • Duration — the downtime specifically for the tested scenario, not the time the ticket existed.

  • Share of unrecovered orders — the portion of purchases that did not occur later and were not completed through another channel.

  • Order marginal contribution — the amount after the cost of goods and variable expenses incurred with the order. Revenue cannot be substituted for marginal contribution.

  • Team time — the actual hours spent by managers, support staff, and the supervisor according to a single internal evaluation rule.

  • External costs include only confirmed payments: urgent contractor work, fees for processing payment refunds, or additional communication. Reputational damage without a measurement method is excluded from the total.

Scenario

Marginal contribution of unrecovered orders

Team time

External costs

Total

Optimistic

2 880 RUB

550 ₽

0 ₽

3 430 ₽

Basic

15 390 ₽

1 320 ₽

8 000 ₽

24 710 ₽

Stress test

79,200 ₽

3,300 ₽

18,000 ₽

100,500 ₽

The base scenario unfolds as follows: 18 paid orders per hour × 0.75 hours × 60% unrecovered × 1,900 rubles marginal contribution = 15,390 rubles. Six tickets at 12 minutes each yield 72 minutes, or 1.2 hours. At an internal cost of 1,100 rubles per hour, this adds another 1,320 rubles. After adding 8,000 rubles for external work, the total reaches 24,710 rubles.

Three training scenarios for direct outage cost: 3.43, 24.71, and 100.50 thousand rubles
The range varies due to duration, share of recovered orders, marginal contribution per purchase, and additional expenses.

Not every unavailable cart is lost

The most sensitive assumption is the recovery rate. If a customer repeats payment within an hour, the order should not be counted as lost a second time. If a manager completes the purchase manually, the order's marginal contribution is preserved, but labor effort is incurred. If a payment is deducted but the order is not created, a state reconciliation is needed first: a possible refund or reprocessing is a separate expense, not an automatic loss of the entire amount.

Therefore, I would calculate two boundaries. The upper bound assumes that no affected order is recovered. The lower bound accounts for confirmed repeat payments and manual completions. To select the required response time, use the range until observation yields a stable recovery rate.

Add only those expenses that can be explained

Basic scenario row

Calculation

Total

Margin contribution from unrecovered orders

18 × 0.75 × 0.60 × 1,900 ₽

15,390 ₽

Team work

6 × 12 min = 1.2 h; 1.2 × 1,100 ₽

1,320 ₽

Emergency work

Confirmed invoice in the sample scenario

8 000 ₽

Direct impact

15 390 + 1 320 + 8 000

24 710 ₽

Do not add total revenue, margin, and marginal contribution simultaneously: these are three representations of a single sale, not three separate losses. Do not multiply the time of all participants by the same rate if the internal valuation rule differs. And do not convert unverified complaints into a monetary line item just to achieve a larger total.

When a faster response justifies the cost

Assume a standard process restores checkout in 120 minutes, while a faster one does so in 45 minutes. The difference is 75 minutes, or 1.25 hours. In the base model, the saved marginal contribution equals 18 × 1.25 × 60% × 1,900 = 25,650 rubles per comparable incident. Team time and other consequences are intentionally excluded here to keep the boundary conservative.

Parameter

Standard process

Fast process

Difference

Time to recovery

120 min

45 min

75 min

Expected unrecovered marginal contribution

41 040 ₽

15 390 ₽

25 650 ₽

Yearly surcharge

0 ₽

96,000 ₽

96,000 ₽

With an annual surcharge of 96,000 rubles, the break-even point is 96,000 ÷ 25,650 ≈ 3.74. Rounding up, at least four comparable incidents per year are required for the saved marginal contribution to match the surcharge. This is not a promise of payback. The accelerated process must genuinely cover the required component and reduce time, while incident frequency must be confirmed by logs, not expectation.

The 96,000-ruble surcharge threshold is crossed by the saved marginal contribution at the fourth incident
In the training model, four comparable incidents yield 102,600 rubles in saved marginal contribution against a surcharge of 96,000 rubles.

Before comparing tariffs, determine who notices the outage and who fixes the online store. The stated response time may apply only to the server, while the root cause lies in the payment module or integration. A useful next step is to review the scope of website support for one specific checkout scenario and formalize handoffs between teams.

What to record in the incident card

  • start and end of unavailability based on a shopper scenario check;

  • which step failed: catalog, cart, checkout, payment, or confirmation;

  • expected paid orders for a comparable period and the source of that data.

  • how many orders were completed later, manually or through another channel;

  • the marginal contribution of a single order and the list of accounted variable expenses;

  • the actual time spent by participants, external bills, and actions taken after recovery.

Verifying the result is simple: recalculate using actual data after the incident and compare it with the preliminary model. If the duration decreased but the share of unrecovered orders remained unchanged, it is premature to draw conclusions from a single indicator: it is possible that a smaller share of customers returned later or the traffic composition changed. If the calculation constantly changes the decision, store the range and the event for review.

Boundaries of the training model

The model estimates the direct effect of one type of failure. It does not prove that every incomplete order went to a competitor, does not assess long-term trust, and does not replace accounting. For subscriptions or annual contracts, you must match the same composition of work, period, and number of incidents.

Include the surcharge for a fast process in the same period as other expenses. If you are comparing the full scope of launch and maintenance options, list emergency response as a separate line item in the overall budget and separately calculate the budget for the store's first year.

Discussion 0

Share your experience and ask questions. Comments without links appear after editorial review.

No comments yet. Start the discussion.