How-to · FundNest

How to digitise a Somiti's record keeping (without ditching the register)

If you organise a Somiti, a cooperative, or any kind of weekly savings circle, you already know the real work isn't the saving — it's the bookkeeping. Member contributions, late-payment notes, loan disbursements, refund records: every Friday meeting ends with another row in the register and another chance for the maths to go wrong.

The promise of "going digital" is that the maths goes right. The reality is messier — most Somiti organisers we've spoken to in Dhaka and the districts have tried a digital switch once, found it slower than the register, and gone back. This post is about how to do the switch without that outcome.

The mistake almost every Somiti makes first

The instinct is to put everything into the new app on day one: every member, every past loan, every receipt going back two years. Don't do this.

You will spend three Fridays typing instead of collecting, the experienced members will lose patience, and you'll quietly start keeping the register again "just in case". Six weeks in, you'll have two systems that disagree.

Instead, draw a line at today's date. Past records stay in the register. New transactions go into the app. This is the same rule accountants use during a system migration, and it's the only one that works in practice.

What to move first, in order

Here's the order that works for most savings circles:

1. The member list (one Friday)

Phone number is the unique key — not name (people share names) and not NID (people refuse to share it). Get every active member into the app with: name, phone, join date, current monthly contribution. Take 10 minutes after the Friday meeting; everyone is already there.

2. Weekly contributions (start next Friday)

Once members are in, the next Friday's collection goes into the app instead of the register. This is the smallest possible change — the workflow you'd do anyway, just with a phone instead of a pen.

Two rules at this stage: collect and mark paid in front of the member, and read the running balance back out loud. Both build trust faster than anything else you can do.

3. Loans and disbursements (week 3)

Only once weekly contributions are flowing smoothly do you start logging loans. Loans are higher-stakes — a wrong loan record loses trust permanently — so let the easy stuff (contributions) build muscle memory first.

For each loan, record: borrower, principal, interest rate (if any), disbursement date, expected return schedule. Don't try to model complex repayment plans yet; a flat monthly EMI is enough for week three.

4. The kitty balance (week 4 onwards)

By week four you have enough data to start trusting the app's central fund balance. This is the point at which you can stop reconciling against the paper register every Friday. Until then, keep doing both. Yes, it's annoying. Yes, it's necessary.

What to keep on paper (forever, probably)

Three things shouldn't move to an app, even after you're fully digital:

1. The constitutional rules of the Somiti — who can join, how loans are approved, what happens when someone leaves. A digital app is the worst place for the rules of the system that runs the app.
2. Signed loan agreements — Bangladesh courts still want a piece of paper for disputes over Tk. 5,000+. A photo of the signed page lives in the app; the original lives in a folder.
3. The annual audit — print out the year-end statement, get it signed by two members, file it. This is your insurance against "the app is wrong" arguments later.

The three mistakes that kill the migration

We've watched this go wrong enough times to name the patterns.

Mistake 1: needing internet to record a transaction

Most consumer apps assume you're online. Friday meetings happen in community halls, fields, sometimes in someone's living room where the wifi is unreliable. If an app needs a live connection to record a contribution, it will fail at the worst possible moment.

Use an offline-first app. The data should write to the phone immediately, then sync to the cloud later when convenient. This is the single most important property to look for. (Disclosure: we built FundNest specifically for this — every contribution, loan, and withdrawal is saved to the phone first and synced when there's signal.)

Mistake 2: trusting one phone

What happens when the organiser's phone breaks? If the only copy of the data is on one device, you have a single point of failure that takes the whole Somiti down.

Use an app that supports encrypted backup to Google Drive (or whatever cloud you trust). Set it to back up automatically every Friday after the meeting. Test restoring once a quarter to a second phone, just to know it works.

Mistake 3: hiding the maths from members

The register's hidden superpower is that anyone can flip back three pages and see exactly where their money went. If your app makes a member's running balance hard to see — or only the organiser can see it — you've replaced a transparent system with an opaque one. Trust goes down, drop-outs go up.

Every member should be able to see (or be shown) their own contribution history, loan status, and current balance in under 10 seconds. This is non-negotiable.

What a successful month-one looks like

If the migration is going well, by the end of month one:

• Every active member is in the app with the right phone number.
• Four Fridays of contributions are logged. The register has the same four weeks for cross-check.
• One or two small loans are recorded.
• The kitty balance in the app matches the cash you've actually counted, to the taka.
• Two or three members have been shown the app on their own phone and have nodded.

If any of these are missing, slow down — don't push to the next phase until they're true.

The deeper reason this works

A Somiti isn't really a savings vehicle. It's a trust network with money attached. The digital tool either supports that trust or erodes it; there's no neutral position.

The migration plan above is designed around that fact. We don't move past contributions until everyone can see their own balance. We don't fold the register away until the app has matched it for four straight weeks. We don't store contracts in the cloud because Bangladesh's courts don't accept screenshots.

If you respect those constraints, the app becomes a relief — no more Friday-night arithmetic, no more "Bhai, last month you said I'd paid 800". If you skip them, the app becomes a problem you didn't have before.

The Reserve Bank of India's working paper on cooperative society audits is worth reading for the formal version of this — most of the failure modes they document are versions of the three mistakes above.