Load Shedding Sales Impact: How to Measure It

To measure the impact of load shedding on your sales, line up your sales by hour with the times your site was without power, then compare outage hours against similar hours with power. The difference, adjusted for day of week and season, is your estimated lost (or shifted) trade. It will not be perfectly precise, but it turns a vague feeling into a rand figure you can use to decide on backup power, trading hours or staffing.

Step 1: Get the right sales data

Daily totals are too coarse, because an outage from 18:00 to 20:30 affects only part of the day. Export transactions with a timestamp from your point-of-sale, e-commerce or ordering system and aggregate them by hour. You want, for each branch or channel:

• Date and hour (for example 2026-07-14 18:00).
• Number of transactions.
• Sales value (be consistent: all VAT-inclusive or all exclusive).
• Branch or channel, if you have more than one.

Step 2: Build your outage log

Published schedules tell you when your area was scheduled to be off, but actual outages can start late, end early or be skipped. The most accurate source is your own record:

1. Record the actual start and end time of each outage per site (a simple shared sheet works).
2. If you only have the schedule for your area and stage, use it, but label the data as ‘scheduled’ not ‘actual’.
3. Include unplanned outages too, and mark them separately.
4. Note whether backup power (generator, inverter, solar) was running, and what it covered: tills only, or lights, fridges and card machines too.

Then convert the log into the same hourly grid as your sales. An hour is ‘affected’ if power was off for, say, 30 minutes or more of it. Add a column: outage = yes/no.

Step 3: Make a fair comparison

Comparing a Tuesday outage evening with a Saturday morning tells you nothing. Compare like with like:

A worked example

Say your takeaway normally does R4,200 between 18:00 and 20:00 on weekday evenings with power, averaged over the last month. Across 12 weekday evenings with outages in that window, it averaged R2,700. The shortfall is R1,500 per outage evening, or R18,000 across the month.

Now check for shifting. If the 20:00–22:00 window on those same evenings averaged R900 more than usual (customers came back when the lights did), the net loss is closer to R600 per evening, or R7,200 for the month. That distinction matters: it is the net figure you should compare with the cost of a generator or inverter.

Step 4: Look beyond lost sales

• Costs: diesel, generator maintenance, extra staff hours when trading shifts later.
• Stock: spoilage of perishables after long outages.
• Online channels: online orders may rise when stores are dark, or fall if your own systems go down.
• Payment mix: a jump in cash share may point to card machines failing.

Our guide to spotting anomalies in business data helps separate outage effects from other unusual days such as public holidays and month-end paydays.

Common pitfalls

• Too few comparison hours: two or three outage evenings are not enough to separate a real effect from normal variation. Wait for more data or widen the window.
• Ignoring other causes: rain, school holidays, payday weekends and promotions all move sales. Note them in your log so you can exclude or explain unusual days.
• Mixing stages: short, predictable outages behave differently from long or unplanned ones. Analyse them separately where you can.
• Counting backup-powered hours as outages: if your inverter kept the tills running, those hours belong in their own category.

Step 5: Turn the analysis into a decision

Once you have a monthly net loss estimate, compare it with the monthly cost of backup options (finance repayments, fuel and maintenance) or changes like opening earlier on scheduled outage days. Repeat the analysis quarterly; outage patterns change, and a decision that made sense at one level of disruption may not at another.

If you want this without building formulas, you can upload the joined file (sales by hour plus an outage flag) to Summarix. It calculates the comparisons and charts with its own code, highlights the biggest gaps and writes a plain-English summary; on paid plans you can ask follow-ups like ‘what was the average basket during outage hours at the Durban branch?’.

Key takeaways

Measure at hourly level, use your own outage log where possible, compare the same hours on similar days, and always check whether sales were lost or merely shifted. The result is a defensible rand figure that makes investment decisions about backup power much easier. For more on reading patterns in sales, see how to find trends in sales data.