A look back, written in 2026. I started taking client work on Elance in 2013 and bought numanhussain.com in 2015. This is what the years in between taught me, and what changed when the work stopped coming through a platform.
The platform years
Elance was a marketplace: a client posted a job, freelancers bid, and a rating followed you everywhere. At the end of 2013 it announced a merger with oDesk, and by 2015 the combined business had become Upwork, with Elance itself wound down soon after. For anyone building a reputation there, that meant moving it — and learning that a profile on someone else's platform is rented, not owned.
My first jobs were pixel-perfect WordPress themes built from PSDs, for clients I never met. The pay was low and the competition global. What I got in return was volume: dozens of different clients, codebases and briefs, far more than most developers see early on.
What bidding taught me
Read the brief twice and answer the problem, not the job title. The proposals that won were short, specific and showed I had understood what the client was actually worried about. The ones that lost listed my skills.
Ask the question nobody else asked. A client who had posted “build me a website” often needed something smaller, or different, and saying so was the fastest way to earn trust — even when it made the job smaller.
The rating system was harsh but useful. It made every project a reference, which is a good habit to keep long after a platform stops enforcing it.
What I had to unlearn
Competing on price. Platforms push every freelancer towards the cheapest credible bid, and it trains you to undervalue judgement — the part of the work that saves a client from building the wrong thing. That judgement is what clients now pay for.
Saying yes to everything. On a platform, a gap in work feels dangerous. Directly, the projects that went badly were almost always ones I should have turned down or reshaped at the start.
What changed when I went direct
When clients started coming to me directly rather than through a platform, the work moved from tasks to outcomes: not “build this page” but “we need more enquiries” or “this system has to be maintainable by someone else”.
That shift changes everything upstream. Discovery becomes paid work. Proposals describe a problem and a plan, not hours. And the relationship lasts longer, because a client who trusts your judgement brings the next problem to you before they have written a brief.
What did not change
Reply quickly. Write things down. Say what will be done, by when, and what is not included. Tell the client early when something is going wrong. None of that is sophisticated, and it is still the difference between the clients who come back and the ones who do not.
The platforms have changed names and rules several times since 2013. Those habits have not, and they are most of what I would tell anyone starting out now.
