The verdict in three sentences
An SME commissioning its first custom software needs neither an 80-page requirements document nor an improvised backlog, but a 3 to 8 day discovery workshop that produces a product vision, 40 to 120 prioritised user stories and acceptance criteria. In 2026 such a workshop costs in the order of 3,000 to 8,000 EUR or USD, and it lets you sign a fixed price per release or capped time and materials. On comparable projects it avoids around 30 % budget overrun.
Classic requirements document or agile backlog
A classic specification describes everything upfront: screens, rules, edge cases. It feels safe, but it locks in choices made before anyone has seen a screen, and it is often written by someone who will not build the product. An agile backlog splits the need into user stories delivered in two-week sprints, and accepts that priorities shift.
The risk of poorly framed agile is real: without a vision or scope, the budget drifts sprint after sprint. The right approach combines both: a stable specification base, then agile execution.
| Criterion | Full requirements document | Agile backlog, no discovery | Discovery then agile backlog |
|---|---|---|---|
| Preparation time | 4 to 10 weeks | 0 to 1 week | 1 to 3 weeks |
| Preparation cost 2026 | 6,000 to 20,000 EUR/USD | near zero | 3,000 to 8,000 EUR/USD |
| Price commitment possible | yes, but rigid | no | yes, per release |
| Adapting to user feedback | weak, change orders | strong | strong within scope |
| Overrun risk | 20 to 40 % via change orders | 40 to 80 % | 5 to 15 % |
| Best suited to | public tender | internal prototype | SME, first custom app |
These are orders of magnitude observed on SME projects in Europe, North America and Africa.
What the discovery workshop must deliver
The workshop brings together the owner, two or three key users and the supplier's team. It lasts 3 days for a simple tool and up to 8 days for an app with several roles and integrations.
| Deliverable | Content | Typical volume |
|---|---|---|
| Product vision | problem, users, success metrics | 1 to 2 pages |
| Process map | current and target journeys | 3 to 8 diagrams |
| User stories | « as a..., I want..., so that... » | 40 to 120 |
| Acceptance criteria | testable conditions per story | 3 to 6 per story |
| Prioritisation | MoSCoW, minimum core identified | 50 to 60 % Must |
| Wireframes | clickable key screens | 10 to 30 screens |
| Release plan | releases of 4 to 8 weeks, priced | 2 to 5 releases |
| Risk log | integrations, data, compliance | 5 to 15 risks |
The most underrated deliverable is the acceptance criterion. A story saying « I want to export orders » says nothing. With three criteria (CSV format, date filter, 10,000 rows exported in under 10 seconds), it becomes testable, and therefore billable.
MoSCoW prioritisation protects the budget: the Musts form the first release, Shoulds and Coulds come later if the value justifies them. A project that starts with 100 % Musts is a poorly framed project.
Which contract to sign after discovery
Two models work well in 2026. Fixed price per release: each release has a set of stories, a price and a date; new requests go into the next release or replace a story of the same size. Capped time and materials: the supplier bills actual time, with a cap per release beyond which it works at its own cost or asks for a decision.
At Kolonell, custom applications are priced in three levels: simple, 2,500,000 to 4,000,000 FCFA (3,810 to 6,100 EUR), medium, 4,000,000 to 7,000,000 FCFA (6,100 to 10,670 EUR) and complex, 7,000,000 to 15,000,000+ FCFA (10,670 to 22,870+ EUR). Discovery is precisely what places your project in the right level before any commitment.
Two-week sprints end with a demo: the business sees the software progress, accepts stories against their criteria and adjusts the next sprint's priorities. That is where budget control is won.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Mini case study
Daniel, owner of a medical supplies distributor in Montreal with 25 staff, wants an app for orders, stock and delivery routes. Without discovery, a first quote suggested 70,000 CAD with a vague scope, and such projects commonly overrun by 30 to 50 %, so a likely final cost of about 95,000 CAD.
After a 5-day discovery costing 7,000 CAD, 72 user stories are written, 41 of them Musts. Release 1 (orders and stock) is priced at 42,000 CAD over 8 weeks, release 2 (routes) at 21,000 CAD. Total 63,000 CAD, with overrun held to 8 %, about 68,000 CAD. Adding the workshop gives 75,000 CAD against a likely 95,000 CAD: about 20,000 CAD saved, and release 1 is live 2 months earlier.
FAQ
How long does a discovery workshop take?
From 3 days for a simple tool to 8 days for an app with several roles and integrations. It is often spread over 2 to 3 weeks so users have time to validate.
How many user stories for a first app?
Between 40 and 120 depending on complexity. A useful first release usually holds 25 to 45.
Can we change our minds mid-project?
Yes, that is the point of agile: a new story replaces one of equivalent size in the release or moves to the next one. The release budget stays fixed.
Do we need a project lead on our side?
Yes, a product owner available 2 to 4 hours a week to answer questions and attend the demo every two weeks.
Is discovery wasted if we change supplier?
No, as long as the contract says you own the deliverables: stories, wireframes and release plan remain usable by any team.
Let's scope your project. We run your discovery workshop, write the user stories and split your app into priced releases, with release 1 live in 8 to 12 weeks. Detailed quote within 48 h. WhatsApp +221 77 596 93 33.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.
