Bangladesh's software industry usually exports quietly — outsourced code for foreign firms, rarely with a country's name on the deliverable. This one has a name. Dream71 Bangladesh Limited has been contracted under the United Nations Development Programme to design and implement the Strategic Planning and Investment Management Information System for Timor-Leste's Ministry of Planning: the software through which the country will manage its national development plan, public investment, budgets, programmes and international commitments on one platform.
What the system has to do
It must track progress on Timor-Leste's Strategic Development Plan 2011–2030 and its Medium-Term Plan 2026–2030, alongside commitments under the UN Sustainable Development Goals and ASEAN-related obligations. Underneath that it collects data from government agencies, validates and approves it, stores and analyses it, and presents the result through interactive dashboards: public investment projects, budgets and spending, movement on development indicators, and programme outcomes. Separate dashboards are planned for the strategic plan, government programmes, sector plans, the SDGs and an integrated overview. The target is delivery within six months.
Why a Dhaka company wins this
Dream71's managing director, Rashad Kabir, called the award a matter of pride and described it as more than a software project. The revealing detail is the track record: the company has already built three systems with the UNDP in Timor-Leste and counts 18 pieces of UNDP work in total.
That is the whole mechanism, and how this market operates deserves saying out loud. Donor-funded government software is not bought the way commercial software is. The buyer is an agency spending public money under procurement rules, which means the dominant consideration is not capability or price but risk of failure — a procurement officer is judged far more harshly for a project that collapses than for one that merely cost more than it might have.
So references dominate. A firm that has already delivered inside the same ministry has demonstrated the one thing that cannot be demonstrated in a proposal: that it can get a government department to actually use what it built. Incumbency compounds, which is why the fourth contract in a country is much easier to win than the first, and why the first is worth bidding for at a price that barely makes sense.
The structural advantage, and its ceiling
Government technology — planning systems, budget trackers, digital registries, monitoring dashboards — is a niche where Bangladeshi firms have a real edge. They spent fifteen years building exactly these systems for their own government, often to donor specifications, which produced a pool of companies that understand how ministries work, in countries with comparable constraints, at a fraction of the cost of European or Singaporean consultancies. Timor-Leste, Bhutan, Nepal, several African states and the Pacific islands are the natural market; UN agencies are the natural channel.
The ceiling is worth naming too, because it is where this kind of success usually stalls. Competing on delivery cost works until you want margin. Every contract is bespoke, every contract consumes senior people for its duration, and revenue scales only by hiring. The step that changes the economics is turning repeated builds into a configurable product sold many times — which is a different company, with different risks, and very few firms in this niche anywhere have made that transition.
Why these projects actually fail
Six months is aggressive for a system spanning a whole government, and public-sector scope grows. But schedule is not the usual cause of death, and neither is code.
A planning dashboard is a consumer of data it does not generate. Every number on it arrives because someone in a line ministry opened a form and typed it in, correctly, on time, every month, as an addition to their existing job. That person frequently gains nothing from doing so — the dashboard serves the centre, while the cost of feeding it falls on the periphery.
So the predictable failure is not a crash. It is a system that works perfectly and goes stale: launched with enthusiasm, populated during the project when consultants are present to chase people, and three-quarters empty eighteen months later, with officials maintaining the spreadsheets they always used alongside it. Anyone who has worked on government systems has seen this. The honest version of such a project spends as much effort on who enters what, why they would, and what happens if they do not, as on the software.
The things that predict survival are unglamorous: making the system the only route to something the data provider needs — a budget release, an approval, a report they must file anyway — rather than an extra duty; keeping entry burdens small enough to fit a real workload; and having a named owner inside the ministry after the vendor leaves.
What to watch
Not the launch. Launches are easy and photograph well. The measure is whether Timor-Leste's officials are still entering data two years later, and whether a minister ever makes a decision because of something the dashboard showed.
If both are true, the reference is worth considerably more than the contract — and for a Bangladeshi firm trying to build an export business on exactly this kind of work, the reference is the product.
