How Northlea approaches a business problem
Better technology starts with a better understanding of the business.
Most technology projects begin with a solution in mind. A new platform, an AI use case, an automation opportunity or a system somebody believes the organisation should have.
We prefer to start further upstream. We want to understand what the business is trying to achieve, what is preventing that outcome today, and which decisions actually determine whether the economics improve.
Technology is downstream of the commercial problem.
A business can implement excellent technology and still fail to improve its economics. The system may work exactly as designed while solving a problem that was never commercially important enough to justify the investment.
That is why our first questions are usually about the business rather than the technology. What creates revenue? What determines gross profit? Where does cash become constrained? Which risks matter most, and what is preventing management from acting earlier?
The technology conversation comes later.
The first question is not “What can this technology do?” It is “What would materially improve the business?”
Understand what happens from demand to growth.
Business problems rarely respect organisational charts. A sales decision can affect inventory, inventory changes working capital, working capital affects cash, and cash determines what the company can invest in next.
Looking at a single function in isolation can therefore produce a locally sensible answer that creates a problem somewhere else. We prefer to follow the economics end to end.
Behind most commercial outcomes sits a decision.
Once we understand the problem, we want to identify the decision that changes it.
Sometimes that decision is obvious: how much stock to order, whether to extend credit, which customers to prioritise or where to allocate capital. Sometimes the real decision is hidden beneath several symptoms.
A forecasting problem, for example, may not really be about forecasting. It may be about deciding what to buy, where to position inventory or how much capacity to commit before demand becomes visible.
That distinction matters because a system should be designed around the decision it needs to improve.
Work backwards from the economic outcome.
Once the decision is clear, the rest of the problem becomes easier to structure. We use a simple sequence to move from the business problem to a practical intervention.
BUSINESS PROBLEM
What commercial outcome is being constrained?
We try to define the problem in economic terms rather than technology terms. “We need an AI forecasting system” is a proposed solution. “We are repeatedly carrying the wrong inventory because we cannot see changes in demand early enough” is a business problem.
DECISION
Which decision would materially change the outcome?
If the decision does not matter economically, improving it may not be worth much.
INFORMATION
What would someone need to know to make that decision better?
This is often where the real constraint appears. The information may not exist, may exist but arrive too late, or may already be present somewhere in the business without being connected to the decision.
INTERVENTION
What is the simplest practical way to improve that?
This could involve AI, automation, analytics, integration, a workflow change or a much simpler rule.
ECONOMIC OUTCOME
Did the intervention change something that matters?
Good decisions lose value when they arrive too late.
Businesses do not always struggle because management lacks intelligence or experience. Often, the problem is that information reaches the decision-maker after the useful options have already narrowed.
Something changes first. The signal appears later. The business interprets that signal, makes a decision and eventually acts.
We call the distance between those stages Decision Distance.
The objective is not always to predict the future perfectly. It is often to shorten the time between something important changing and the business becoming able to respond.
The value of information is whether you receive it while you can still do something about it.
The answer may already be somewhere inside the business.
Organisations often respond to an information problem by deciding they need more data. Sometimes they do.
But frequently, useful information is already being generated. It sits in transactions, customer behaviour, operating systems, conversations, field activity or patterns that nobody currently brings together.
The challenge is to determine what is commercially meaningful and connect it to a decision. That distinction becomes particularly important across African markets.
Useful information does not have to look like a conventional dataset.
Many systems were built around assumptions developed in highly formalised markets: long credit histories, consistent digital records, integrated systems and structured data at every stage.
That is not always how economic activity is recorded across African markets.
A customer may have limited conventional records while still leaving a meaningful economic trail through purchases, payment behaviour, mobile money, ordering frequency, trading relationships or operational history.
A business may have information spread across ERP systems, spreadsheets, point-of-sale data, WhatsApp, field teams and human knowledge.
Fragmented information is harder to use. It is not automatically useless.
A customer can be economically legible before they are institutionally legible.
Not every business needs prediction first.
There is another assumption we want to challenge: that the most advanced technology is automatically the best place to begin.
In some organisations, the biggest improvement would simply come from seeing what is happening consistently. In others, the opportunity is to anticipate what is likely to happen next or automate a decision that is currently slow and repetitive.
We think about that progression as five levels of capability.
Sophistication is not the goal. Better commercial capability is.
We are technology-positive. We are not technology-first.
AI can create genuine commercial advantage. So can automation, analytics, better integrations and properly designed software.
But adopting technology is not the same as improving a business. The intervention has to justify itself against the problem we are trying to solve, the value of solving it and the practical reality of implementing it.
That means Northlea may sometimes recommend a sophisticated solution. It may also recommend something much simpler.
Both are acceptable if they improve the economics.
The right technology is the one that materially improves the decision at a sensible economic cost.
The test we keep coming back to
Which decision would materially change the economics if we made it better, earlier?
That question brings the conversation back to the business. Once we know the answer, we can ask what information would make that possible, what intervention is justified and whether technology has a meaningful role to play.
That is where Northlea starts.
Start a conversation