Where a software company's clients actually come from

Ask a software company in Cairo where its last five projects came from and the answer is usually a list of names: a former colleague who moved into distribution, a cousin's employer, a client from two years ago who opened a second business. That is a list of accidents, not a pipeline. The accidents are real revenue and can carry a firm for two or three years, but they arrive at a rate nobody controls, and when the rate drops there is no lever to pull.

Work reaches a software company through a limited set of channels. Each one has a price in money and in founder-hours, a lag before it produces a signed contract, a ceiling it cannot pass, and a characteristic kind of client. Most of the wasted years in small firms come from judging one channel by another's rules: waiting for referrals to scale, or abandoning content in month four because it has produced nothing.

Referral dominates early and then stops without warning

Referral costs no money and a moderate amount of attention. Someone who has worked with you, or knows you socially, hands your name to a person with a problem. The first call is short because the trust is already established, and the project is often signed two to four weeks after it. Nothing else moves that fast.

The ceiling is the problem. A network produces referrals in proportion to its size and its overlap with people who buy software, and both of those are close to fixed within any given year. In the firms we have worked with, a strong professional network in Egyptian retail or manufacturing produces a handful of qualified referrals a year, and that number barely responds to effort. Doubling the hours spent on lunches and phone calls does not double it.

Referred clients also negotiate differently. The relationship behind the introduction makes it harder to hold a price, harder to enforce a scope boundary, and harder to leave a project that has gone wrong. The channel that is easiest to close is the one where commercial discipline is weakest.

The client you have already delivered for is the cheapest revenue available

A system running in production generates its own demand. The finance module makes it obvious that purchasing is still on spreadsheets. The commerce platform works, so the same company wants the warehouse side built the same way. The second project usually appears six to eighteen months after the first goes live, needs no formal proposal, and closes at a better margin because the discovery work was done a year ago.

This channel exists only if you are still in the room. A firm that hands over a repository and disappears has no returning clients, because by the time the next need forms the client's memory of you is a zip file and a password. A firm that operates the system, answers the phone when a report breaks, and reviews the thing once a quarter is present at the moment the need appears. The mechanics of that are the subject of after delivery: support, recurring revenue and the second project.

The ceiling here is set by how many systems you have delivered and how many of them you still operate. It grows as the firm grows, slowly and predictably. It is also the channel most easily destroyed: one abandoned or badly run project removes a client from the list permanently.

Partnerships borrow trust that someone else built

Design agencies, management consultants, accounting firms, and hosting or infrastructure providers all sit next to clients who need software and cannot supply it. An agency wins a brand project and the client asks who will build the ordering system behind it. An accountant sees a manufacturing client running the whole business on three spreadsheets. Introductions from these people arrive pre-qualified: the budget is roughly known, the problem has been articulated, and someone the client trusts has vouched for you.

The cost is mostly time before revenue. A partnership takes three to nine months to produce a first project, because the partner has to watch you handle something before attaching their name to you. Some partners want a share of the contract, commonly ten to twenty per cent; others want technical help they cannot justify hiring for. Both are cheaper than the equivalent volume of cold outreach.

Two cautions. A partner can only introduce you for work they can describe in a sentence, so this channel rewards a narrow, sayable specialism and mostly ignores the general software house. And a partner who owns the client relationship can end your access to it with a single decision, which is concentration risk in miniature.

Outbound and inbound are the two channels you can actually build

These are the only channels whose output responds directly to what you put into them. They behave in opposite ways, and a firm intending to grow past its network needs both.

Outbound

Outbound is deliberate approach to companies you have identified as having the problem you solve. The money cost is small: list research, a mailbox, sometimes a data source, under 10,000 EGP a month for a small operation. The real cost is attention. Fifteen to twenty researched approaches a week is about half a day of a founder or a competent salesperson, every week, without gaps.

The first signed project usually lands two to four months after the first message, and only if the approaches were specific enough to be worth answering. The ceiling is high but strictly linear: twice the volume produces roughly twice the meetings, and the channel stops the day you do. Client quality is the widest of any channel, which makes qualification the entire game. What separates outbound that works from outbound that irritates people is covered in outbound for a software company.

Inbound

Inbound is the slowest channel to start and the only one that compounds. An article explaining what a warehouse system actually has to reconcile, or why two estimates for the same project can land three to four times apart, keeps arriving in front of buyers for years after it is written. The cost is writing time, and it is not delegable: a substantial technical piece takes a day of attention from someone who has actually done the work.

Expect six to twelve months before a search-driven inquiry becomes a contract, and the first half of that period to look like failure. The compensation is quality. A buyer who arrives having read your reasoning has already accepted your approach and often your price range, so the conversation starts several steps ahead of a cold one. Being found: content and search for a technical company covers how a small firm sustains this without a content team.

Marketplaces, events and tenders suit particular firms

Freelance marketplaces

Upwork, Mostaql and Khamsat produce work faster than any other cold channel, sometimes within days. The cost is platform commission, usually ten to twenty per cent, plus bidding time that is significant and unpaid. The ceiling is low. Budgets on the Arabic platforms are small, every proposal is compared against the cheapest bid on the page, and the profile you build is rented rather than owned. A narrow, well-reviewed Upwork profile reaches mid-size international budgets, but none of that reputation transfers elsewhere. Marketplaces suit a small team that needs cash flow this quarter, or a firm testing whether a new specialism has demand.

Events, chambers and trade associations

In sectors where buyers cluster — manufacturing, pharmaceutical distribution, logistics, contracting — the industry exhibition, the chamber of commerce committee and the trade association meeting put you in front of decision-makers with no intermediary. The costs are real: membership fees, a modest stand costing more than a month of a mid-level engineer, travel, and several full days of founder time per event. Nothing closes at the event itself. Relationships mature over three to twelve months. The channel is strong for a firm that serves one sector and close to worthless for a firm that serves everyone.

Job boards, RFPs and tenders

A company advertising for two in-house developers has both a budget and an admitted software problem, a buying signal free to anyone who reads job boards deliberately. Tenders are a different business. A serious response to a government or large-corporate RFP consumes one to two weeks of senior time, win rates are low by design, and payment terms run long after award. Contracts are large and the documentation burden is heavy. This channel suits firms with the cash reserves to wait months for a decision and months more for payment, and it quietly damages everyone else.

Concentration is what ends firms, not competition

A company where most of the year's revenue arrived through a single channel is one change away from a bad year. A partner hires their own development team. A founder's network moves on. A platform adjusts how bids are ranked. The failure is rarely dramatic: the pipeline thins over two quarters and nobody can name the cause, because nobody recorded where the work came from in the first place.

The defence is arithmetic rather than courage. Record the origin of every project you sign, in one line, and read the list at the end of each quarter. If one channel is above roughly two-thirds of signed work, the next quarter's development effort belongs to a second channel, whatever the first one is currently producing.

The time to build the second channel is while the first one is still working.

Every channel except referral has a lag measured in months, so one started after the pipeline empties arrives too late to help with the emergency that prompted it. Two working channels is stability; three is a firm that can plan a year.

If you want the next system scoped as a piece of work rather than an intention, talk to us. Softwiro scopes projects the same way whether they arrive through a referral or a search result.