A receipt records what you paid on one trip. It becomes useful for budgeting only after you make the product, quantity, unit, store, date, and promotion status comparable. Keep your personal purchase history separate from official inflation statistics: they answer related but different questions.
The practical goal is not a perfect index. It is a small, repeatable record that tells you whether the same basket is becoming more expensive, whether pack sizes changed, and which substitutions are driving your total.
Start with a product definition
“Milk” is not one product for comparison. Record brand or type, fat level, package size, number of units, and whether the receipt shows a discount. Use a stable product name such as “semi-skimmed milk, store brand” rather than the abbreviated till code, and keep the pack size, such as 1 litre, in its own column, as in the table below.
Normalize every line before comparing prices
- Copy the date and store. Keep the purchase context.
- Record quantity and package size. A two-pack is not one unit.
- Calculate the unit price. Divide the final line amount by the weight in kilograms (grams ÷ 1,000), the volume in litres, or the number of items.
- Label the price state. Keep shelf price, loyalty discount, coupon, and final paid amount as separate fields.
- Note substitutions. A cheaper brand or smaller pack is a choice, not evidence that the same product fell in price.
Use a unit-price table
The figures below are an illustrative example, not prices copied from real receipts.
| Date | Normalized product | Pack | Final price | Unit price |
|---|---|---|---|---|
| 4 May | Oats, store brand | 1 kg | €1.80 | €1.80/kg |
| 11 May | Oats, store brand | 1 kg | €1.95 | €1.95/kg |
| 18 May | Oats, store brand | 750 g | €1.65 | €2.20/kg |
| 25 May | Oats, store brand | 1 kg | €1.70 promo | €1.70/kg |
The 18 May ticket looks cheaper than €1.95, but its normalized price is higher because the pack is smaller. Mark the 25 May promotion separately so it does not become your assumed everyday price.
Separate promotions from the underlying price
Record at least three values when the receipt allows it: regular or shelf price, discount amount, and final amount paid. For a household budget, final paid amount matters. For a price history, regular and promotional prices answer different questions. Never average them together without a label.
- Use “regular” for an ordinary shelf price.
- Use “loyalty” for a discount that requires a membership.
- Use “coupon” for a targeted or clipped discount.
- Use “clearance” for a price unlikely to repeat.
Review four weeks without overclaiming
At the end of four weeks, count comparable observations, calculate the median unit price, and write down substitutions and promotions. Four observations can reveal a pattern worth checking; they cannot describe national inflation or forecast next month’s bill.
- Compare like with like before comparing stores.
- Keep missing observations visible instead of filling them with a guess.
- Review the basket when the product, package, or recipe changes.
Personal receipts are not CPI or HICP
The U.S. Bureau of Labor Statistics publishes average prices for selected items, the USDA Food Price Outlook explains its forecasting framework, and Eurostat documents the HICP methodology. These official measures use defined samples and methods. Your receipts show your household’s choices, stores, quantities, and timing.
Know what the receipt cannot tell you
A receipt may omit stock-outs, quality differences, waste, meal substitutions, and the price you would have paid at another store. Merchant text may also be too short to identify a product. Correct the line manually and keep a note rather than pretending the data are precise.
- Do not treat a sale as a permanent price.
- Do not compare different weights as if they were identical packs.
- Do not infer a national trend from one household’s basket.
Choose the lightest tracking workflow
A paper list works for a small basket. A spreadsheet is useful when you want formulas and a visible audit trail. A receipt-scanning app can reduce typing, but recognition still needs correction and category review. Choose the workflow you will actually maintain.
How Lumy can support the record
Lumy publishes this guide, so read this section as the app maker’s description of its own product. When you scan a photo of a paper receipt or a screenshot of an e-receipt, Lumy reads the merchant, date, total and individual product lines, including discounts and deposits, and saves the purchase straight away. Open it afterwards to correct any field the scan misread, including each product line. No scanner reads every receipt correctly: faded thermal paper, folds, unusual layouts and very long receipts are where mistakes are most likely, so compare each item name, quantity, unit and price with the receipt.
With Pro, scanned product lines build a price history for each product, and Opportunities points out overpayments and spending increases measured against your own history. Lumy does not read live shop prices. Receipt scanning also works on Free: the AI assistant gets 10 requests a month, and a typed message, a voice message or a receipt photo each uses one. Pro removes the monthly limit and scans up to 5 receipts at once, and a 3-day Pro trial is available in the app on iPhone and Android. Without Pro, you can still log purchases in Lumy and keep this price record on paper or in a spreadsheet. Compare the Free and Pro plans.
Your next step: log the next four receipts
Choose ten repeat products, create seven fields for each line (date, store, normalized name, pack size, price paid, unit price, price state), and enter the next four receipts. At the end, circle one price change you can explain and one that needs another observation.
Sources and methodology
Official price programs and a personal receipt history use different samples and purposes. The workflow above is a budgeting aid, not an official inflation estimate.
- U.S. Bureau of Labor Statistics — Average Price Data
- USDA Economic Research Service — Food Price Outlook
- Eurostat — HICP methodology
Prepared and reviewed: 31 July 2026.
