How to create a matching platform in Algeria (2027): an app for tradespeople and home services, requests with no answer, cash that does not go through you, customers who bypass the platform; density, vetting, revenue model, and what we put in place

“I want to launch an app to find a plumber, an electrician or home help in Algiers. Today people search on social networks, call three numbers, nobody comes or the price changes on site. How do I create a matching platform? I do not know where to start, how to vet tradespeople, how to get paid if everything is settled in cash, or how to stop the customer and the tradesperson doing without me from the second time on.” A matching platform is never short of ideas: it is short of requests that find a taker, and of money that comes back to it.

Symloop13 min read
A matching platform: one neighbourhood, three trades firstA request in three screens, dispatch to nearby tradespeople, vetting, reviews, cash, prepaid creditMATCHING PLATFORM · ALGERIA · 2027A matching platform:one neighbourhood, three trades firstA request in three screens, dispatch to nearby tradespeople,vetting, reviews, cash, prepaid creditACROSS THE WILAYAno answerIN THREE COMMUNESaccepted in 4 minutessymloop.com

Short answer. A matching platform rests on four decisions. Density: a single zone and three trades, until every request finds a tradesperson within minutes. Trust: vetted tradespeople, reviews given after a real job. Revenue: when everything is paid in cash, the platform's money has to arrive before the job, not after. A reason to come back: the work redone if needed, a receipt, a history. In this page's example, the same twelve tradespeople answer in 4 minutes across three communes, and do not answer across a whole wilaya.

This page is for founders, in Algeria or in the diaspora, who want to launch an app for tradespeople, repairs or home services, and for companies that want to organise their own network of field workers. The definitions come from public sources: two-sided market, network effect, online marketplace, disintermediation, lead generation, reputation system. The maps, the distances, the amounts and the rates are a constructed example, with fictitious figures. The page covers neither the legal status of the platform and of the tradespeople, nor tax, nor insurance, and quotes no hardware price.

A request, in six steps

The figure follows a request, from the need to the review.

Figure 1 · Six steps. The accented step is dispatch: a request is only worth something if it reaches, within the minute, tradespeople who are near and free.

The market. A two-sided market is an intermediary platform that brings together two distinct user groups; here, customers and tradespeople. Cross-group network effects arise when two groups depend on each other: customers come if there are tradespeople, tradespeople stay if there are requests. It is the chicken-and-egg problem, and it is solved zone by zone. The need. Three screens: the trade, the problem with photos, the place and the time slot. The place is a point on the map, not an address to dictate, see delivering without an address.

Dispatch. That is the accented step: the request goes to tradespeople of the right trade who are near and free, through the channel that arrives, see notifications. Acceptance. For a routine job, the first to accept; for a job that needs estimating, two or three quotes. The job. The arrival time, the work, the price confirmed before starting. Payment. In cash to the tradesperson, most of the time; the app keeps a receipt. The review. Only after a real job, otherwise it is worth nothing.

01

Density before numbers

The figure places the same twelve tradespeople on two maps.

Figure 2 · The same twelve tradespeople: in three neighbouring communes the request is accepted in 4 minutes; spread over the whole wilaya it gets no answer. Fictitious positions.

In the example. Twelve tradespeople of the same trade, one request, a dispatch radius of 3 km. In three neighbouring communes, seven tradespeople are within the radius and five are free: the request is accepted in 4 minutes. Spread across the wilaya, only one is within the radius, and they are busy: after 15 minutes nobody has answered. What that says. The number of tradespeople signed up measures nothing; what counts is how many are free ten minutes from a request.

Where to start. A zone you know, three trades where urgency is real, plumbing, electrics and locks for example, and fifty tradespeople met one by one. What is measured. The share of requests accepted in under fifteen minutes; as long as it is not high in the first zone, you do not open a second. Widening the radius. It saves nothing: nobody accepts a small job twenty kilometres away. The side to help. To attract one group, a platform may subsidise the other: at launch it is the tradespeople you go and find and equip, because customers are already looking.

«The number of tradespeople signed up measures nothing; what counts is how many are free ten minutes from a request.»
02

Five ways of earning

The figure gives five models, and what can go wrong in each.

Figure 3 · Five ways of earning. None is right everywhere: each fails in its own way, and the choice is tested on one zone before spreading.

Commission. A percentage of the price, paid over by the tradesperson after the job: natural on a marketplace where payment goes through online, fragile when it is made in cash. A fee per request. A fixed amount, taken from credit bought in advance, when the tradesperson accepts: it is a form of lead generation paid by the unit. Subscription. A monthly flat fee: simple, but unsellable while requests are rare.

A fixed price online. The customer pays by card and the platform pays out: the cleanest, when the customer has a card and the price is known in advance, see online payment. Paid promotion. Tradespeople pay to appear first: a side income, to be handled with care. Disintermediation. It is the removal of the intermediary. From the second time on, the customer has the tradesperson's number: no model prevents it. What brings them back is what the phone does not give: a tradesperson found in four minutes when their own does not answer, a receipt, the work redone if the problem returns.

03

Getting paid before, not after

The figure follows the platform's money, in days around the job.

Figure 4 · On 1,000 jobs paid in cash: 280,000 dinars collected of 400,000 due with a commission, 250,000 collected in advance with prepaid credit. Fictitious amounts.

In the example. 1,000 jobs at 4,000 dinars, paid in cash to the tradesperson. With a 10% commission, 400,000 dinars are due to the platform: 220,000 arrive on time, 60,000 after a reminder, 120,000 never. With a fixed fee of 250 dinars taken at acceptance from prepaid credit, 250,000 dinars are collected, before the job. What that says. The two amounts look alike; one means chasing hundreds of people, the other does not.

Why commission leaks. It depends on what the tradesperson declares: the real price, and even the fact that the job took place. What prepaid credit requires. Serious requests: a fake or cancelled request is refunded automatically, otherwise tradespeople leave. And a simple way to buy credit: card, approved transfer, or a top-up point. The accounts. Every purchase, every deduction and every refund is written to a journal, see roles and the audit journal. Nothing is final. The model is tested on one zone, measured, and changed before the next is opened.

«What is taken at acceptance never needs chasing.»
04

Twelve weeks, for one zone and three trades

Twelve weeks in the example, before a closed launch.

Figure 5 · Twelve weeks in the example. The accented period is the first; the launch is closed, with fifty vetted tradespeople in a single zone. Indicative durations of a constructed example.

Scoping. One week: the zone, the three trades, reference prices for routine jobs, the revenue model to test. The tradesperson app first. Sign-up, documents, working zone, availability: it is the one shown to the first fifty. Sign-in is by phone number and code, see sign-in by code. Vetting. Identity, number, trade, and a first job observed: it is done by hand, at the start.

The customer app. It comes next, and fits in three screens. The closed launch. One zone, real requests, corrections every week. What the platform does not do. It makes nobody competent and does not replace meeting people: it organises the request, the answer, the record and the trust. The durations are those of a constructed example; the starting decisions are described in creating a mobile application.

05

What is inside a matching platform

The figure stacks five layers; the third is the difference between a listing and a platform.

  • What you receive: the customer app and the tradesperson app; dispatch by trade, zone and availability; vetting of tradespeople; prepaid credit with deduction at acceptance and refund; receipts and reviews; the dashboard per zone.
  • What the platform avoids: the request with no answer, the signed-up profile that does not exist, the review written by a friend, the commission nobody pays over, the second zone opened before the first one works.
  • What is measured: requests accepted in under fifteen minutes; time to arrive; cancellations and no-shows; requests per customer; credit bought and used; average rating; per zone and per trade.
  • What Symloop does: scoping the model, the two apps, dispatch, vetting, credit and the movement journal, notifications, the dashboard, in Arabic and in French.
Figure 6 · Five layers. The accented layer is dispatch: it decides whether a request finds a taker within minutes.

Two apps. The customer's, to ask without phoning; the tradesperson's, to answer in one tap from a job site. Dispatch. That is the accented layer: trade, zone, availability, radius, time limit. A classifieds site leaves the customer to call; a platform sends the request and knows whether it found a taker, see the classifieds site.

Trust. A reputation system lets the members of a community rate each other: documents checked, reviews after a job, disputes traced. Steering. Requests served, time to accept, cancellations, per zone and per trade, on a dashboard. Ready-made or custom. Templates for matching apps exist and are enough for a demonstration. Custom work is justified for dispatch by zone and availability, prepaid credit and its refunds, vetting, and everything that makes a request find a taker; for a marketplace of products, see creating a marketplace.

How do you create a matching platform?
By starting small: one zone, three trades, fifty vetted tradespeople. You build the tradesperson app and the dispatch of requests first, then the customer app, and you measure the share of requests accepted within minutes.
What comes first: customers or tradespeople?
The tradespeople of a single zone. Customers are already looking; what is missing is someone who answers fast. With no free tradespeople nearby, the first request is also the last.
How do you vet tradespeople?
By hand at first: identity, phone number, trade, and a first job observed. Then through reviews, given only after a real job.
How do you earn money if everything is paid in cash?
By being paid before the job: a fixed amount taken from credit bought in advance when the tradesperson accepts a request. A commission claimed afterwards depends on what the tradesperson declares.
How do you stop customer and tradesperson bypassing the platform?
You do not stop it: you give a reason to come back. A tradesperson found within minutes when their own does not answer, a receipt, the work redone if the problem returns.
Are two apps needed?
Yes: the customer's and the tradesperson's do not have the same screens. They share the same server, and can be developed from a common code base.
How much does a matching platform cost?
It depends on the scope. The heaviest item is not the customer's screen but the dispatch of requests, the credit and the vetting. A week of scoping gives a costing.
Should job prices be fixed?
For routine jobs, reference prices reassure the customer and avoid the price that changes on site. For the rest, a quote after diagnosis.
What do you do with a request that gets no answer?
Re-send it to other tradespeople, warn the customer after a few minutes, and record it: that is the figure that says whether the zone is dense enough.
When do you open a second zone?
When the first one answers: most requests accepted in under fifteen minutes, customers who come back, tradespeople who buy credit again.
Tell us which zone, which trades, and how you plan to earn: we start with a week of scoping

An app for tradespeople or home services to launch, requests with no answer, cash that does not go through you, customers who bypass the platform, or a project you are setting up in Algeria from France

Symloop builds the platform that fits your project: customer app, tradesperson app, dispatch, vetting, prepaid credit, dashboard. Describe your zone to us, your trades and your model.