Short answer. A B2B ordering portal fills up when three conditions are met: the retailer sees their own price and credit ceiling, not a public price list; they see a cut-off time that promises next-day delivery before confirming; and the pre-seller opens the accounts during visits and earns a bonus on orders placed through the portal by their route. The rest, catalogue, one-tap reorder, WhatsApp read by an AI agent, Android app, card payment, is added over twelve months. Ten to fourteen weeks for the first column, a pilot on one route after Eid al-Fitr, one route a week afterwards.
This guide is for the food distributor or beverage wholesaler who has equipped their pre-sellers, or is about to, and wants the retailers who know what they want to order without a visit; the importer selling to wholesalers; the pharma wholesaler; the supplier to restaurants and cafés; and the local brand that wants a portal for its distributors. It is the sequel to the pre-sale app described in the previous guide: same master data, same depot, one more channel. It covers neither tax nor wholesale trade regulation.
What the figures say
B2B self-service is measured worldwide. In the ninth edition of its B2B Pulse survey, run in 2024 among 3,942 decision makers in the United States and twelve other countries, McKinsey finds that digital self-service and remote interactions with a sales rep account for 34 % of B2B revenue, and describes a “rule of thirds”: one third self-service, one third remote with a rep, one third in person; 39 % of buyers, against 28 % two years earlier, say they are willing to place orders above 500,000 dollars through these channels. It is a survey of large buyers, to be read as a direction, not as an Algerian figure.
The case closest to Algeria is in Cairo. MaxAB, the Egyptian ordering platform for grocery shops, merged with Kenya’s Wasoko on 27 August 2024: more than 450,000 merchants in five countries, 4,000 employees, Egypt as the largest market with 180 million dollars of annualised sales, and more than 20 million dollars of credit extended to retailers in the year, repaid at over 99 %; before the merger, MaxAB served more than 150,000 shopkeepers in Egypt and Morocco. On 22 May 2025 the group acquired Fatura, whose network brings together 626 wholesalers and 25,000 retailers in sixteen Egyptian cities, with orders placed in an app and delivered by the wholesalers’ own fleets. In France, the portal has settled among restaurateurs: Choco connects 15,000 restaurants and 10,000 suppliers in six countries, including France, for more than 100,000 orders a month according to a May 2025 profile, free for the restaurant, paid by the supplier.
In Algeria the ground is the same and the portal is missing. The figures presented by the minister of domestic trade on 10 December 2025 count 1,092,176 retail traders and 115,423 wholesalers in the trade register, one wholesaler for roughly ten retailers; 51.63 million 4G and 5G subscriptions in the second quarter of 2026 according to the ARPCE put a capable phone in every shop; and we found no dominant B2B ordering platform between wholesalers and retailers, orders going through the pre-seller, the depot’s phone and WhatsApp. That is the gap this guide describes, wholesaler by wholesaler, before any marketplace.
A retailer’s order, in six moments
The portal is not described by its screens but by what happens to an order, from account opening to reorder, and by what must be true at each moment for the retailer to come back.
Opening the account does not happen online: nobody signs up alone on a wholesaler’s portal. The pre-seller creates the account during a visit, from the master data: the customer’s price list, credit ceiling, route day, and sends them a link by SMS that opens the portal with nothing to install. The order is placed in the catalogue at their price, with their last orders replayable in one tap and the month’s promotion; or by a WhatsApp message, text or voice, in Arabic or French, which the AI agent turns into an order to confirm, never validated on its own. The cut-off is the decisive moment: before 2 pm, the order goes on tomorrow’s route; after, on the day after’s; the retailer sees it before confirming, and it is that kept promise, not the app, that they adopt.
Picking receives the orders of the three channels in a single list per route; on a stock-out, a substitution is offered to the retailer and confirmed, rather than discovered at delivery. Delivery and payment produce a numbered note, a collection in cash, cheque or CIB or Edahabia card, the balance updated and an SMS receipt, as for an order taken by the pre-seller. Reordering is what makes the difference at six months: the reminder on the usual day with the typical basket, and a follow-up when three weeks pass without an order, because a retailer who stops ordering has often changed supplier without saying so. The depot and collection mechanisms are those of the pre-sale guide; the portal adds none of them, it adds a door.
Three ordering channels, one depot
The portal is not a second circuit; it is a third channel sitting on the same master data and ending in the same picking list.
The master data is shared by the three channels and that is what holds the whole thing together: one price per customer, one ceiling, one cut-off per route, available stock. A portal with its own price file diverges from the pre-seller within a month and the retailer picks the lower price of the two. The pre-seller keeps the app described in the previous guide, plus two roles: opening accounts and ordering for those who will never order alone, who remain many. The portal is first a web page opened from an SMS link, with nothing to install; the Android app comes in the second quarter for those who order every week, because it is what allows reminders and one-tap reorder.
WhatsApp is the channel retailers already use: a voice message saying “send me the same as last week plus two cases of oil”. The AI agent reads it, matches it to the typical basket and the catalogue, and produces an order the retailer confirms with one word; it never validates alone, because a wrong reference costs a route. The guide on AI agents in companies describes what this agent can do and what it is not entrusted with. The depot receives everything in one list and the dashboard says, per route, what share of orders comes from each channel: that is the figure that drives the pre-seller bonus and the pace of extension. If the depot already has a delivery app or stock management, they plug in here.
«The retailer does not adopt an app; they adopt a cut-off time and a next-day delivery.»
A worked example
An example on assumptions, to redo with your figures: a food distributor, 1,100 retailers on file, twelve pre-sellers, one depot, twelve months after the portal launch.
The self-service share. At twelve months we assume 38 % of orders go through the portal or WhatsApp and 62 % through the pre-seller. That figure does not rise by wilaya but by route: the route whose pre-seller opens accounts and earns their bonus places 60 % of orders in self-service, the one whose pre-seller fears the portal stays at 10 %. Next-day delivery. With the 2 pm cut-off and a picking list at 2:05 pm, 95 % of orders leave the next morning, against D+2 delivery when the notebook was re-entered in the evening; that is the figure the retailer sees, and it is for that they come back.
Coverage. The visits freed by self-service are not removed: a pre-seller who no longer has to take fifteen orders a day visits the small outlets at the end of the route, and the share of retailers visited in the month rises from 75 % to 90 % in our assumption, 165 more shops served. Reordering. On 400 retailers active on the portal or the app, the usual-day reminder with the typical basket adds one order a month each in our assumption: 400 more orders a month, at equal average basket, without a visit. The assumptions to check before believing any of it: the number of retailers who already order by phone or WhatsApp without a visit, which is the first reservoir, the current delivery lead time and the order rate per visit, which one month of notebooks lets you count. What depends on no assumption: an order that exists once, whatever the channel, is no longer re-entered.
Six reasons retailers do not order by themselves
The B2B portals that stay empty in Algeria do not stay empty for lack of technology; they stay empty for six reasons retailers readily give when asked.
The first three are removed by construction. A web link with nothing to install, WhatsApp accepted and an app reserved for those who order every week settle “I don’t install apps”. One price per customer shown to them alone settles “I don’t know my prices”; it assumes the price master data exists, which is the real work of the first four weeks. A small free discount range and the visible promotion of the month settle “I want to haggle” for most; the rest goes to the sales manager for approval, with the history in front of them. The last three are removed by rules written before the pilot. The ceiling and balance shown to the retailer, with a rule saying when the order is accepted, blocked or sent for approval, settle “I buy on credit”; the cut-off and route day shown before confirming, the confirmation then the delivery notice settle “I don’t know when it arrives”.
“I prefer the pre-seller” is the reason that decides, and it is legitimate: the relationship is worth more than the tool. The portal that works does not remove the pre-seller; it pays them to open accounts and for every order placed through the portal by their route, at a rate that makes a portal order at least as worthwhile for them as one taken in the notebook. The retailer keeps their pre-seller, who spends less time entering and more time selling new products and collecting. A seventh obstacle is not in the table because it belongs to the wholesaler: launching the portal on a wrong price file, where the retailer sees a higher price than the one the pre-seller gave them; they do not come back.
Distributor, wholesaler, importer, pharma, restaurants, brand: who builds what first
Six situations, and in each what gets built in ten to fourteen weeks, then what comes next.
The food distributor builds the web portal, the WhatsApp agent, the price per customer, the cut-off and the single picking list; the app, card payment, the reorder reminder and the bonus come next. The beverage wholesaler lives on reorders: the typical basket replayed in one tap, returnable packaging tracked per customer and the cut-off per route are their first column. The importer selling to wholesalers has a two-level portal, the wholesaler then their own retailers, with tiered discounts and a balance per wholesaler; it is the only one of the six cases where multi-level is justified from the start. The pharma or parapharma wholesaler adds batches, dates and temperature-controlled delivery, described in the cold-chain guide.
The supplier to restaurants and cafés has an inverted cut-off, the evening order for morning delivery, voice messages in the middle of service and a delivery note checked in the kitchen; it is the case closest to what Choco does in France. The local brand selling through distributors builds a portal for them, with allocation when the product is short and sales reporting per zone. In all six cases, the multi-supplier portal, where several wholesalers sell to the same retailers, is another project, described in the guide on B2B platforms and marketplaces; it comes after one wholesaler has proven their own. The first column is specified in nine pages as the guide on the specification shows, whose worked example is precisely an ordering platform for retailers.
«The pre-seller is not replaced by the portal; they are paid to fill it.»
The portal and the pre-seller: not one against the other
The question every wholesaler asks is about the pre-sellers: will they fight it? Yes, if the portal is presented as what replaces them, and they will win, because they are the ones who open the accounts. No, if the rule is written before the pilot: the pre-seller opens the accounts of their route, earns a bonus on every order placed through the portal or WhatsApp by their retailers, and keeps the coverage target. Their day changes in content, not in value: less data entry, more visits to small outlets, more time for new products and collection.
The figure that settles the discussion is the self-service share per route, shown on the dashboard and known to everyone. A route at 10 % after three months says the pre-seller is not opening accounts; a route at 60 % says they understood. The bonus is computed on that figure and on collection, not on volume alone, otherwise the pre-seller fills the portal with orders they would have taken anyway. The pre-sale guide describes the time-stamped visit and coverage; the portal adds a column to the same table.
Seen in France and in Cairo: Choco, MaxAB, and what does not transpose
In France the restaurateur orders from suppliers in the evening, after service, in Choco: an app free for them, paid by suppliers, where their usual suppliers have their catalogue and where they replay last week’s order; 15,000 restaurants, 10,000 suppliers, more than 100,000 orders a month according to the May 2025 profile. What transposes: free for the buyer, the replayed order, the voice message turned into an order, and the evening cut-off for the morning. What does not: the multi-supplier model from day one. Choco aggregates a restaurant’s suppliers; an Algerian wholesaler starts with their own retailers, because they are the one holding the master data and the depot.
In Cairo, MaxAB-Wasoko built what many Algerian wholesalers imagine: an app where the shopkeeper orders, credit granted on history and repaid at over 99 %, private labels for more than 10 % of volumes. What transposes: ordering by the shopkeeper, credit tied to order history, the app as the small shop’s channel. What does not: nationwide in-house logistics funded by fundraising rounds. The Fatura acquisition points to the Algerian way: 626 wholesalers delivering themselves the orders placed in the app. An Algerian wholesaler does not need to become MaxAB; they need their retailers to order from them before 2 pm, and CIB or Edahabia card payment comes when the retailer has it.
The 2027 calendar
A portal’s calendar is dictated by Ramadan, expected from 8 February to 9 March 2027, whose three preceding weeks are the food distributor’s season, and by summer, the beverage season.
January is for master data and rules: one price per customer, a ceiling, a cut-off per route, the free discount range, the pre-seller bonus; all of it written with the sales manager and the pre-sellers, because they are the ones who will open the accounts. Ramadan is for building the web portal and the WhatsApp agent while the field runs at full speed; nothing is launched. After Eid al-Fitr, expected around 9 or 10 March, the pilot starts on one route of 80 retailers whose pre-seller has signed the bonus: three written criteria, the self-service share at eight weeks, the share delivered next day, zero price gap between portal and notebook. April to May, the Android app and card payment arrive for those who order every week. May to June, extension happens one route a week, before summer. Summer is the beverage season: the reorder reminder, the three-week follow-up and volume promotions run without a visit. September reviews per channel and per route, and takes the decision on multi-supplier. Ramadan dates are forecasts. What a year of portal then costs, servers and stores included, is detailed in the guide on keeping an app alive.
What makes a B2B ordering portal fail
The ordering portals we see fail almost always fail for one of these reasons.
- Launching an app instead of a cut-off time. The retailer does not adopt a screen; they adopt a next-day delivery they see before confirming.
- A public price list. A portal where everyone sees the same price contradicts what the pre-seller did; the retailer does not order and the pre-seller says nothing.
- A separate price file. Portal and pre-seller diverge within a month; the retailer picks the lower price and the depot no longer knows which to apply.
- The pre-seller paid against the portal. Without a bonus on their route’s portal orders, they open no accounts, and nobody signs up alone.
- The AI agent validating alone. A WhatsApp order validated without confirmation produces one wrong reference a week and a lost route.
- Multi-supplier first. A marketplace without one wholesaler who has proven their portal has neither master data nor depot; it has a shop window.
- Launching before Ramadan. The three weeks before it are the season; a new channel that day costs orders.
Frequently asked questions
The questions we receive from wholesalers, distributors and importers.
- Will grocers really order on a portal?
- Some, and the share grows by route: those who already order by phone or WhatsApp without waiting for the visit are the first reservoir, often a quarter of retailers. They adopt a cut-off time and next-day delivery, not an app; the web link with nothing to install and WhatsApp read by the AI agent are there for them. The others continue with the pre-seller, who opens the accounts of those who switch.
- Will the portal replace my pre-sellers?
- No, and a portal presented that way fails, because the pre-sellers are the ones who open the accounts. The portal changes the content of their day: less data entry, more visits to small outlets, more collection and new products. They earn a bonus on their route’s portal orders, computed on the self-service share and collection, not on volume alone.
- How do we handle credit on the portal?
- With each retailer’s ceiling and balance, shown to them alone, and a written rule: order accepted under the ceiling, sent to the sales manager for approval above it, blocked beyond a second threshold. The rule is replayed on one month of history before being switched on, to see how many orders it would have blocked.
- Do we need an app or a website?
- A website opened from an SMS link first, with nothing to install, because that is the retailer’s first objection. The Android app comes in the second quarter for those who order every week: it allows reminders, one-tap reorder and offline work.
- What exactly does the AI agent do on WhatsApp?
- It reads a text or voice message in Arabic or French, matches it to the catalogue and the retailer’s typical basket, and produces an order the retailer confirms with one word. It never validates alone, it does not invent references, and it hands over to the pre-seller when it does not understand.
- How long before it shows?
- Ten to fourteen weeks for the master data, the web portal, the WhatsApp agent and the single picking list, then a pilot on one route after Eid al-Fitr. The pilot route’s self-service share is read at eight weeks; next-day delivery shows on day one.