Data Analytics and Business Intelligence Platforms

Three answers to the same question

It usually surfaces in a management meeting. Somebody asks how many active customers the institution has, three departments give three different figures, and the next twenty minutes go on reconciling numbers instead of deciding anything. Nobody is being dishonest. The figures were produced in different places for different purposes, and nobody ever settled which of them was the institution's.

Institutions here arrived at that position honestly. Most grew by adding systems: a core, a channel bought later, an agent network, a lending product, each acquired at a different time for a different reason and each reporting on itself. Around them sit the spreadsheets that fill the gaps, usually kept by one person who becomes a dependency nobody planned for. Digital channels then multiplied the number of things there are to count.

The cost is paid twice. Once in effort: board packs assembled by hand over several days, a month-end that consumes the finance team, returns prepared under pressure. Once in speed. A question asked on Monday comes back a fortnight later, by which time the decision has been taken on instinct or has stopped mattering.

It presses hardest on whoever is accountable for the whole picture, usually the finance or operations lead. A bank has board committees, a supervisor and a branch network all asking for different cuts. A SACCO answers to an elected board and to members at a general meeting, who ask their questions in public. A microfinance institution reports portfolio quality to funders as well as regulators, and a fintech's investors expect numbers monthly that hold up under scrutiny.

Agreeing what an active customer is

The hardest part of this work is not the reporting. It is agreeing what the words mean, which is a political exercise inside your institution rather than a technical one. Take the active customer. Sales counts everyone who has opened an account. Operations excludes anyone dormant. Finance counts those holding a balance above a threshold. Each is reasonable, and each is attached to how a department is measured and sometimes paid.

The same argument sits under most numbers worth reporting. Whether the unit is a customer or an account. When a loan is in arrears, and from which day. Which branch owns a deposit when the customer signed up in one town and transacts in another. Whether a reversal counts. None of these have technically correct answers. They are decisions, and someone has to take them.

That part belongs to you, and it usually takes longer than anything on our side. It needs one person with authority above the departments, able to hear both cases and rule, and a written record of what was ruled. Without that, the disagreement survives the project and reappears in a better looking format. Definitions also change as products launch and rules move, so settle who owns them now. We bring the questions and run the sessions. The decisions are yours.

Five numbers for the board, and their own for everyone else

A board wants very few numbers. Where the institution stands, which way each line is moving, and a short explanation of anything that moved unusually, in a form that can be read before the meeting rather than presented at it. Giving a board forty charts is a way of not answering the question.

Below that sits the depth. Deposits and lending by product and segment, the cost of funds, arrears by product, region and officer, adoption and income by channel, branch performance against target. A branch manager wants their own book, their own targets and their own exceptions, this week rather than in next month's pack. Once people can see their own position, the argument moves from whose figure is right to what to do about it.

Finance and compliance have a different requirement again: month-end close, and the returns your supervisor requires, in a form you can stand behind when asked how a figure was arrived at. Product, credit and risk teams need the follow-up question. Whether an arrears pattern is national or confined to two regions. Whether a new channel won transactions from a competitor or moved them off your own counters. The value is in that second question, asked without commissioning a piece of work.

Which comes first is decided with you. We usually argue for whatever somebody currently builds by hand every month, because that return is immediate and visible to the people who must be won over. What a SACCO owes its members and what a funder expects of a microfinance book are different starting points, and the scope is set against yours.

Where this sits beside the systems you already run

This does not replace the systems that run your business. Your core, your lending platform and your channels carry on doing their work. What it adds is the ability to answer questions that cross all of them, which no single system can do, since each sees only its own part.

Buying reporting from each system separately reproduces the original problem in a newer format: numbers shaped by the system that produced them rather than by the questions you are asking. It is also usually the first honest look at what your other investments returned. Whether the app moved customers off the counter or gave existing users another way in. Whether the agent network is earning its commission.

What has to be settled on your side

Two of those have been named already: one agreed set of definitions, and someone with the authority to settle a disagreement between departments.

Then your right to your own information. Some of what you need sits inside systems run by suppliers under agreements signed years ago, and what you may extract, use and keep is a contract question rather than a technical one. Some institutions have more room than they assumed; others need to reopen the conversation with the provider. Worth knowing before anything is designed, and often the longest lead time in the project.

Then who may see what. A branch manager seeing their own branch is uncontroversial. Seeing a neighbouring branch, or a colleague's individual figures, is a management decision with consequences for behaviour. Customer information carries obligations under Tanzania's Personal Data Protection Act, so the line between an aggregate and an identifiable record is worth settling deliberately.

Then an honest account of what is not held anywhere countable: a register kept in a branch, a spreadsheet a department has maintained for years, an approval that exists as an email. What gets corrected at source, what is taken as it stands and what you accept as unavailable for now is an operational decision with a cost attached. Expect people to check new figures against their own for a while too, which is healthy and worth managing rather than resenting.

Working with us

We start with discovery: the questions your leadership needs answered, the decisions that hang on them, and the reporting your people build by hand today. That produces a scope and a plan, and the plan is yours whether or not we go on to build it.

Through delivery you see working software in an environment you control rather than status reports, and the source code and documentation are yours from the first commit. Afterwards we support it and train your people to take on as much as they want, including producing their own reporting. There is no lock-in, by design and by contract.