A typical kirana store bill has eight to twelve line items — rice, dal, vegetables, snacks, soap, and so on. Logging that by typing each item and manually entering a total is one approach; pointing a camera at the receipt is another. Here's an honest, step-by-step look at both.
Open the app — roughly 2 seconds. Navigate to add a new expense — another 2-3 seconds. Type the amount, say ₹487 — 5-8 seconds including switching to a number keyboard. Type a description like "kirana groceries" — another 8-10 seconds. Select a category from a dropdown — 2-3 seconds. Confirm and save — 1-2 seconds. Total: roughly 20-28 seconds for a single grocery bill, assuming no typos requiring correction along the way.
Open the app and tap the scan option — roughly 2-3 seconds. Point the camera at the receipt and capture — 2-3 seconds, assuming reasonable lighting. The system reads the printed total, the merchant name if visible, and the date directly from the image — processing typically completes in 3-5 seconds. A confirmation screen shows the extracted amount and suggested category (Food or Groceries) — review and confirm, 2-3 seconds. Total: roughly 10-14 seconds, with no typing required at any point.
The raw time difference — maybe 10-15 seconds — sounds minor in isolation. The more important difference is cognitive effort, not just elapsed time. Typing requires reading the receipt, mentally extracting the total, switching keyboards, and typing accurately without typos. Scanning requires pointing a camera and confirming what's already extracted — a fundamentally lower-effort action, even when the time difference is small. Lower effort is what determines whether a habit survives a busy week; a 15-second time saving that also removes most of the mental effort tends to matter far more than the raw seconds imply.
Faded thermal-paper receipts, common at smaller kirana stores after a few hours in a pocket, can be harder to read accurately than a fresh, clearly printed one. Handwritten bills from very small vendors with no printed receipt at all aren't scannable in any meaningful sense. In both cases, the honest answer is to fall back to manual entry or voice logging — scanning isn't meant to replace every other input method, just to handle the specific case of a clear, printed receipt faster than typing would.
Picture four grocery and kirana runs across a week, each with a clear printed bill. Typed manually, each takes roughly 25 seconds, totaling about 100 seconds across the week — under two minutes, genuinely not a huge burden on its own. Scanned instead, the same four bills take roughly 12 seconds each, totaling 48 seconds — less than a minute for the same data. The difference across a single week is modest; the difference compounded across a year of regular grocery shopping is meaningfully larger, and more importantly, scanning removes enough friction that the logging is less likely to be skipped entirely on a busy day.
Scanning a receipt captures more than just the rupee amount — the merchant name and date get extracted automatically too, which means the entry is timestamped and labeled without any additional manual input. Typing the same entry manually would require either typing a merchant name as well (adding more time) or skipping it entirely (resulting in a less detailed record). Scanning tends to produce a more complete entry in less time, rather than trading completeness for speed.
Neither method is universally better — a quick chai purchase with no receipt is suited to voice logging, a clear printed kirana bill is well suited to scanning, and an unusual purchase might still need manual typing. The point of having all three available isn't to declare one the winner; it's to make sure whichever purchase is happening, there's a fast enough way to log it that the habit actually survives past the first few days.