Every accounting app looks competent in a screenshot. The screenshot is taken on a good connection, with tidy demo data, on a device the company chose. Your counter is none of those things — which is why choosing on features is how shops end up switching twice in a year.
What follows is not a comparison table. It is four tests you can run yourself in an afternoon, on any product including this one, and what each of them is really asking.
Test one: turn the internet off, then sell something
Put the phone in aeroplane mode and raise a bill. Not open the app — complete a sale, print or share it, and take the payment. This is the single most useful five seconds of testing you can do, because it separates software that treats your counter as the source of truth from software that treats a server as the source of truth.
The difference only shows up on the day it matters. A shop with a good connection six days a week will not notice until the seventh, when there is a queue and the app is showing a spinner. If the sale cannot be finished offline, everything else about the product is a secondary consideration.
Test two: check the currency before you check anything else
Sign up and look at the first bill. Is it in your currency, or in the currency of the country the software was built in? This sounds trivial and it is not: an app that quietly defaults everyone to one currency is telling you that the people who built it had one market in mind, and you will keep meeting that assumption in the reports, the number formatting and the tax fields.
Khata resolves a currency from the country you choose at signup, for every country in the world, because the country a business trades in is a fact rather than a guess. That was not always true here — it defaulted everyone to Indian rupees until a shop registered from the UAE and was created billing in INR, which is how the problem got found.
Test three: make a staff login and look at your own margins
Create a second user with a cashier's role. Then, logged in as that user, try to find what an item cost you. If you can, then everyone on your counter can see your buying price, your supplier terms and your margin on every line.
This is the test most products fail quietly, because permissions are usually built as an afterthought — a single 'staff' flag that hides a menu rather than a rule that hides a number. Check the report screens as well as the product screen; the cost price often survives in a report that nobody thought to lock.
| A cashier should be able to | A cashier should not be able to |
|---|---|
| Raise and print a bill | See purchase price or margin |
| Take a payment against a balance | Delete a completed bill |
| Look up what a customer owes | Change another person's entries |
| Add a new customer at the counter | Export the whole customer list |
Test four: get your data out before you put it in
Before you enter a single real customer, find the export. Download your products, your parties and your balances, and open the file. If the export does not exist, or arrives as a PDF, or requires a support request, you are being asked to make a decision you cannot reverse.
This is the one test where the answer matters more than the product. A shop that can leave has a supplier. A shop that cannot leave has a landlord.
The questions that are really about size
Most disagreement about which accounting software is best is actually disagreement about how big the shop is. Three counters and a storeroom is a different problem from one person behind a till, and a product that suits one will feel wrong for the other in opposite directions — too heavy, or too thin.
- One or two people, mostly cash, some credit: you need billing, a khata that keeps a running balance per person, and stock counts. Anything more is setup you will not maintain.
- A counter with hired staff: permissions stop being optional. So does knowing which person entered an amount.
- Several people entering at once, or a back office: you need somewhere that is not a phone — a screen where a hundred rows are corrected at a time rather than one bill at a time.
- Registered and filing regularly: your accountant's requirements now outrank every other consideration on this list. Ask them first.
When Khata is not the right answer
It is worth being direct about this, because a page like this one is usually written to arrive at a single conclusion. Khata is a poor fit in four specific situations, and all four are things it does not do rather than things it does badly.
- You need e-invoicing or IRN generation. Nothing in Khata produces one. If your turnover puts you in that bracket, this is a hard stop rather than an inconvenience.
- You sell things with a batch number or an expiry date — medicines, most food distribution. The product record has neither field, which is why there is no pharmacy page on this site.
- You sell the same item in sizes or colours and need to track them separately. There are no variants, so each becomes its own product.
- You run several branches and want them as one business with separate books. Khata does not do multi-branch; running them as separate accounts and adding up the exports is the honest workaround, and for some shops that is too much friction to accept.
If none of those describe you, run the four tests above on Khata and on whatever else you are considering. The tests are the point. The conclusion is yours.
A sensible order to do this in
- 1
Write down what actually goes wrong today
Not features you want — problems you had this month. Balances you could not remember, stock you ran out of, a bill someone disputed. Three or four is enough, and they become your test list.
- 2
Install two candidates, not five
Five means you compare none of them properly. Two means you can run a real day on each.
- 3
Run one real day on each
Your own products, your own customers, your own takings. Demo data is built so that nothing goes wrong, which is precisely the wrong test.
- 4
Export from both before deciding
Do this while you still have very little in there. It is the cheapest possible moment to discover the export is bad.
- 5
Keep the old system for a month
Carry over balances, not history. Run both until the numbers agree, then stop. A month of overlap costs far less than a lost ledger.