Short answer. In 2027, in Algeria, an ordering site or web app is built in six to ten weeks; a mobile app with back-office in three to five months; a vertical SaaS with invoicing in four to six; a two-sided platform with escrow and logistics in five to eight; an AI agent connected to an existing back-office in six weeks to three months; a takeover of an abandoned app in six weeks. What turns a three-month project into six is almost never the code: it is the merchant file started too late, store accounts opened in the wrong name, data to import delivered last, and sign-offs that wait. An honest calendar has five lines, not one, and it gives a go-live date because it shows what runs in parallel.
This guide complements can you know the total price before starting: price and timeline are written together, in the same scoping. The timelines below are the ones we write into our calendars; they assume a signed scope and a counterpart who signs off within forty-eight hours.
Real timelines by project type
Six project types, two columns. The first is build time. The second is what gets added when it was not started on day one.
The ordering site or web app, appointment booking, catalogue, payment, tracking, is built in six to ten weeks; what makes it longer is online-merchant status, obtained through the bank and GIE Monétique in weeks to months, which starts on day one or delays everything. The mobile app with back-office, Android and iOS from one codebase, offline, payment, takes three to five months; store accounts in your name, offline mode tested in the field and data import are the three lines that stretch it, detailed in the seven decisions of a mobile app in 2027. The vertical SaaS with invoicing and subscription takes four to six months, and the scope of the trade and the ten pilot customers decide, as in launching a SaaS in Algeria in 2027. The two-sided platform, escrow, logistics, takes five to eight months, because sellers are recruited before the code and carriers are contracted before the pilot. The AI agent on an existing back-office takes six weeks to three months depending on the state of the data and written rules, and one more month if nothing exists in a system. The takeover of an abandoned app takes six weeks, two to three of them to recover accounts and code that are not under your control, as described in my app is stuck with a developer who vanished.
«What turns a three-month project into six is almost never the code.»
The five phases, and who holds each
A project is not measured in months of code. It is measured in five phases, three of which depend on you as much as on us.
Written scoping, one to two weeks: you decide, we write the screens, the rules, the total price and the five-line calendar. It is the phase six-week quotes skip, which is why they do not hold. The tested prototype, one to two weeks, with five real users, not relatives: the cheapest moment to change your mind, and every change requested afterwards costs ten times more. Development, the bulk of the time, delivered every two weeks in versions you can open and test; a project you see nothing of for two months is a late project that does not say so. Data import and field tests, two to four weeks: your real files, your phones, your network, your users; an app opened with wrong data loses the team's trust within a week. Go-live and fixes, two weeks: stores, production, first corrective versions included in the project, not in an extra quote.
What runs in parallel, and depends on nobody
The difference between a quote and a calendar is that the calendar has five lines. Here are the five.
You: decisions and sign-offs, forty-eight hours each or the project waits; and the data to import, delivered clean as early as possible, not the day before go-live. It is the line that most often makes the difference between three months and six. Banks and GIE Monétique: online-merchant status takes weeks to months, depending on the bank and the state of your file; it starts on day one of scoping, and no code can speed it up. The stores: developer accounts in your name, with identity verification, then app review at every publication, a few days, sometimes back and forth. Administration: depending on the project, the company, the startup label, the platform registration with the tax administration planned by the 2027 finance bill; in parallel, never in series. Us: construction, the only line the studio fully controls, and the only one it can commit to without condition.
A quote that shows only one line promises a date it will not keep, because it ignores the other four. A calendar that shows them gives you a go-live date, and tells you exactly what moves it if it moves.
«A quote that shows only one line promises a date it will not keep.»
A real calendar: sixteen weeks for an app with back-office
Here is what the calendar of a mobile app with back-office, payment and data import looks like, for real, when everything starts on day one.
Weeks 1 and 2, scoping; and from week 1, the merchant file and the store accounts, which run until week 12 with nobody able to speed them up. Weeks 3 and 4, the prototype, tested with five users. Weeks 5 to 12, development, delivered every two weeks. Weeks 12 to 14, data import and field tests, with your files received in week 8, not week 13. Weeks 15 and 16, go-live, store review included, and the first fixes.
The same project, with merchant status started in week 8 and data delivered in week 13, goes live in week 24. Nothing changed in the code. That is why we refuse to write a timeline without writing the five lines, and why scoping always starts with the list of what begins on day one.
What turns a three-month project into six
- Merchant status started after the product. The app is ready, cards are not accepted. It starts on day one of scoping.
- Store accounts in the vendor's name. Transfer, verification, weeks lost. In your name, from week 1.
- Data delivered the day before. The import reveals duplicates and empty fields, and everything stops. Clean files mid-project.
- Sign-offs that take three weeks. Every pending screen blocks the next. One counterpart, forty-eight hours.
- No prototype. Changes arrive in development and cost ten times more. Five real users in week 3.
- Nothing visible for two months. The delay is discovered at the end. Deliveries every two weeks, or change vendor.
- A scope that grows along the way. Every “while we're at it” adds a week. The written scope is the list; the rest goes to version 2.
Frequently asked questions
- How long does it take to build a mobile app in Algeria?
- Three to five months for an Android and iOS app with back-office, payment and data import, when the merchant file and store accounts start on day one. An ordering site or web app takes six to ten weeks; a two-sided platform five to eight months.
- Why do quotes promise six weeks?
- Because they count only the construction line and ignore the other four: your sign-offs and data, online-merchant status, store accounts and review, administration. An honest calendar has five lines and gives a go-live date it can keep.
- What depends on nobody?
- Online-merchant status, obtained through the bank and GIE Monétique in weeks to months; developer account verification and store review; and administrative steps depending on the project. These lines start on day one and run in parallel; no code speeds them up.
- How long does scoping take?
- One to two weeks, to write the screens, the rules, the total price and the five-line calendar. It is the phase fast quotes skip, and the reason they do not hold.
- What can I do to make the project faster?
- Three things: sign off within forty-eight hours with a single counterpart, deliver your clean data mid-project rather than at the end, and start merchant status and store accounts in week 1. Those three lines make the difference between three months and six.
- Are fixes after go-live included in the timeline?
- Yes. The last two weeks of the calendar cover go-live, store review and the first corrective versions, which are part of the project and not of an extra quote. Maintenance for the following year is decided separately, before launch.