Research and sources

Pakistan SME bookkeeping: what app evidence can and cannot tell us

Public khata-app listings show substantial distribution, but downloads do not measure complete books or the number of businesses using them.

Key takeaways

Use app-store observations as evidence about particular listings, while keeping adoption, activity and accounting completeness separate.

• Several checked khata listings display cumulative download bands of five million or more.
• A historical study reported about 17.8 million downloads and 740,500 monthly active users across four apps in 2022.
• That comparison is historical and is not a measured retention rate or a 2026 market-share estimate.
• No representative primary survey was found that establishes the national split between paper, spreadsheets and accounting software.
• A small paid-app download band does not prove that subscriptions failed or that businesses will never pay for software.

Method and verification boundary

We checked public Google Play listings on 16 September 2026 and read the original published historical engagement analysis.

This is desk research, not a survey of shopkeepers or an installed-app test. The listings were fetched with the language and region parameters linked below. Ratings, review totals and availability can differ by locale, device and retrieval time. Displayed downloads are bands, not precise counts of unique Pakistani businesses. We did not create accounts, inspect private analytics, test loan or payment services, or verify advertised accounting functions inside the apps.

The historical engagement figures are attributed to Data Darbar's published analysis, which in turn used named third-party measurement providers. We did not obtain those providers' underlying datasets. Current active-user counts, paid conversion, revenue, the national paper-versus-software split and several companies' current operating status remain unverified. Uncertainty is retained rather than converted into a confident description of the whole market.

What did the public app listings show?

The checked listings show large distribution bands and different combinations of credit, cash and business-management claims.

The name Rupin, formerly Udhaar Book, is visible in the current listing. That supports the current product-name observation; the exact rebrand announcement date was not established. The separate DigiKhata Pro listing should not be confused with the much larger main DigiKhata listing. Likewise, an app update date is evidence of a displayed store update, not proof that every backend service, support channel or advertised feature is operating successfully.

Why do some figures differ from older notes?

App-store ratings and update fields change, and locale-sensitive pages can present different snapshots.

An earlier research note recorded different ratings and update dates for some apps. We used the refreshed primary listing rather than preserving the older number as current. This is why the table identifies a check date and links the exact listing. A future reader may see another value without either observation necessarily being fabricated. The correct way to report the difference is to preserve the date and scope, not average incompatible snapshots into a supposedly definitive score.

We omit review-count comparisons because the retrieved pages themselves showed differences between top-level and review-section totals. Those totals are not necessary to the main conclusion. A more rigorous longitudinal study would capture a consistent locale, device category and observation method at repeated intervals. This article does not claim to have performed that study or to rank the apps by user satisfaction.

What did the historical engagement study actually say?

The 2022 figures compare cumulative downloads with one month's estimated activity across a defined group of apps.

Data Darbar's analysis by Mutaher Khan and Natasha Uderani, published 2 November 2022 and checked 16 September 2026, reported 17.8 million combined downloads for Udhaar Book, Easy Khata, CreditBook and DigiKhata, using Appfigures. It reported 740,500 combined monthly active users for October 2022 using Apptopia. The approximate quotient is 4.2%.

That arithmetic is not a cohort-retention measurement: the numerator is one month's estimated activity and the denominator is cumulative downloads from another provider. It is also not evidence that only 4% of Pakistani businesses keep digital books. The defensible finding is narrower: large distribution figures did not translate directly into an equally large measured monthly active population in that historical comparison.

What would establish current adoption?

A current study would need a defined business population, consistent activity measures and an explicit treatment of duplicate users.

One person can install several apps, reinstall one app or use a tool for a purpose other than running a business. A business can also have multiple staff devices. These possibilities prevent a direct conversion from installations to businesses. To measure current adoption, define what counts as an active business and what activity counts as bookkeeping. Record the period, geography and sample selection. Then distinguish occasional balance checks from maintaining a complete monthly record.

A useful survey would ask about the combination of tools used, not force each business into a single category. Paper, an app, a spreadsheet and an accountant's system can all coexist in one records process. Without that detail, the survey might misclassify a digital credit notebook as full accounting or overlook digital records maintained by an external bookkeeper.

Does DigiKhata Pro prove that paid subscriptions failed?

No: the small public download band is an observation, while a commercial outcome needs revenue and conversion evidence.

The DigiKhata Pro listing, checked on 16 September 2026, displayed 100+ downloads, an upfront purchase and in-app purchases, with an August 2024 update date. The fetched locale showed a dollar price rather than a verified Pakistan offer. We therefore do not republish an older rupee subscription price as current, and we do not claim that the listing measures every way customers may pay for DigiKhata services.

A small distribution band can motivate a research question: how much demand exists for that particular paid listing? It cannot answer what revenue the vendor earned, whether sales occurred elsewhere, whether paid functions moved to another app or whether the experiment met its own goals. Describing it as the only hard test of subscriptions, or as proof that subscriptions failed across Pakistan, would go beyond the evidence.

How should willingness to pay be investigated?

Separate the price of a tool from the price of setup, support, compliance work and other services.

A business may obtain an app without an upfront charge while paying for payments, lending, support or professional bookkeeping elsewhere. Another may pay for software because it replaces a specific manual task. To study willingness to pay, identify the purchased service, the user, the buyer and the alternatives. Ask what expense or effort changed after adoption. Do not assume that a free download means the entire records process costs nothing.

This article does not publish market-wide subscription demand, revenue estimates or a recommended price. It also does not treat vendor-published fees as evidence that customers actually paid them. Such a study would need observed transactions or a clear survey method, along with an explanation of selection bias and geography. The app-store table is insufficient for that conclusion.

What is the difference between khata and complete books?

A customer balance record can be useful without explaining every asset, liability, income and expense in a business.

Khata or credit record

A record of amounts due and movements with a counterparty; its exact scope depends on the tool and how it is used.

Payment record

Evidence that money moved through a particular cash, bank or payment channel.

Complete bookkeeping process

A controlled set of records and reviews that captures the business's relevant transactions and supports its reporting obligations.

These are analytical distinctions, not a claim that every app is limited to its original name. The checked listings advertise differing combinations of cash, stock, invoicing, payments and staff functions. We did not install the apps to verify those functions, their double-entry model or their reporting completeness. A feature advertised in a store description should be tested with the business's actual required cases.

For a synthetic example, suppose a shop records a customer's unpaid purchase in a credit app and receives the payment through a wallet later. The bookkeeper still needs to avoid counting the wallet receipt as another sale, explain the original revenue and close the customer balance. If stock is involved, quantity and cost need their own records. The example demonstrates a reconciliation question; it does not assert that any named app mishandles it.

What about paper, spreadsheets and the accountant?

The evidence reviewed does not establish a representative national share for any of these methods.

Public stories can illustrate practices but cannot support a claim that a specific percentage of Pakistani SMEs uses paper, Excel or unlicensed desktop software. We found no primary survey in this research that resolves that split. We also do not infer software piracy from the use of an old product name in a training course. That would require evidence about the actual installation and licence, not a linguistic shortcut.

A more informative research question is how records move between the owner and the person who prepares the books. Identify who records sales, who checks cash, who holds supplier invoices and who prepares period reports. Ask what is lost or retyped at each handoff. A mobile entry tool can be valuable even if the final books are maintained elsewhere. A desktop ledger can be valuable even if the owner never opens it. Those possibilities should be investigated rather than treated as measured national behavior.

How should company closures and pivots be reported?

A missing news item or inaccessible product page is not enough to declare that a company has closed.

The wider research notes included B2B-commerce names such as Bazaar, Retailo, Jugnu, Dastgyr and Tajir. This publication does not provide a definitive current operating-status table because we did not establish a consistent set of primary company statements covering each one. Reports about layoffs, a closed business line or a pivot also do not necessarily describe the whole legal entity. Current status remains unverified where that primary evidence is missing.

For an operator, the useful question is continuity of the particular service being used. Can records be exported? Does a second authorized person control the account? What happens if a feature is withdrawn or an app changes ownership? These questions apply to every vendor, including small open-source projects. They can be tested without making unsupported claims about another company's finances or survival.

What should a business test before relying on an app?

Use a known transaction and its correction to test the records process from entry through independent review.

1. Enter a synthetic credit transaction with a clear source reference.
2. Record a partial payment through the intended payment channel.
3. Correct an intentional mistake and inspect the remaining history.
4. Export the records and ask another person to explain the balance.
5. Test authorized access from a replacement device or account recovery path.
6. Reconcile the app's totals to the separate bank or accounting records.

Keep personal and real customer data out of the trial. Record what the app actually does, what requires a subscription and what the operator must do manually. If the business needs a trial balance, balance sheet or local electronic invoicing, ask for those exact workflows. Do not infer them from a store title containing the word accounting. The software comparison guide helps separate these product families.

Where does PHP Ledger fit?

It is a country-neutral accounting preview, not a replacement for every mobile credit, payment or compliance tool.

PHP Ledger currently provides supported core bookkeeping and a bounded cash POS showcase. Its 0.4.0 starter provides core invoices/bills/payments, optional one-location Inventory/Purchasing and manual tax configuration. It has no FBR client or automatic country tax rules. Its party and currency foundations do not change those operational limits. It also requires an operator for hosting, updates and recovery. A business needing immediate electronic invoicing should read the FBR research scope and select an evidenced solution with qualified advice.

The hosting study explains why software availability is different from an actual compatible installation. The accounting landscape survey explains why source visibility is different from complete accounting. These distinctions keep the focus on the records process rather than on a particular product's popularity.

Questions, sources and conclusion

The strongest conclusion is that distribution, ongoing activity and complete bookkeeping require different evidence.

• The six primary Google Play listings are linked in the snapshot table; all were checked on 16 September 2026.
• Data Darbar's original 2 November 2022 analysis supplies the attributed historical engagement comparison; its underlying provider datasets were not independently obtained.

Use these observations to design better interviews and product trials. Do not quote them as national market share, a current retention rate, a declaration of company failure or an accountant's endorsement. The unresolved questions are part of the result, not gaps to fill with confident assumptions.