How to Learn Coding Online (Without Quitting in Week Three)

A woman emailed me in March with a list. She'd signed up for four platforms in eleven weeks. Two free, two paid. She sent me screenshots of all four dashboards, and every one of them showed the same thing: somewhere between 8% and 14% complete.

She wanted to know which platform was the problem.

None of them were the problem. Every one of those courses would have worked. She'd been doing the hardest version of learning online, which is doing it with nobody expecting anything from you, and she'd been blaming the software for a structure problem.

That's the part almost nobody writes about. The material is not the hard part anymore. It hasn't been for a decade. You can learn this entire craft online for free right now, today, and the reason most people don't isn't the curriculum.

What learning online actually removes

When you take a class in a room, a lot of invisible machinery comes free with it. Someone decides what you learn this week. Someone notices when you don't show up. There are other people beside you who are also confused, which quietly tells you that confusion is normal and not a sign you're too dumb for this.

Take all that away and hand someone a video library, and the library is not the thing that broke.

Online learning strips out three supports and leaves the content. So when you build your own version of this, you're not looking for better videos. You're rebuilding the missing parts: a sequence, a reason to show up, and someone to tell you your code is a mess before the habit sets.

Everything below is about rebuilding those.

Pick one platform and close the other tabs

The single biggest tax on self-taught beginners is platform shopping. It feels like research. It's procrastination in a lab coat, and it can eat a genuinely shocking number of weeks.

So here's a shortcut. Pick based on how much structure you personally need, not on which one has the best landing page.

If you're disciplined and broke, start with freeCodeCamp or The Odin Project today. Both are free, both are enormous, and people get hired off them every year. The catch is that nobody is coming to check on you, so they reward the self-starters and quietly punish everyone else.

If you know you stall without a path, pay for structure. Not because paid content is better, exactly, but because a sequence someone else designed removes the daily question of "what now," and that question is what kills momentum.

If you want to touch code immediately rather than watch it, look for interactive lessons where you type into the browser and something happens. For a lot of beginners that loop is the difference between finishing week one and not.

I broke all of this down properly, with honest cons for each option, in the best affordable coding courses. If money is the blocker right now, free coding bootcamps covers the genuinely free routes.

Whatever you choose, choose once. You can always switch in a month. You cannot switch every Tuesday and also make progress.

Build the week before you build the app

Here's the part that actually separates the people who finish from the people with four dashboards at 11%.

Decide when you're doing this, in advance, on specific days. Not "evenings." Not "when I get a chance." Tuesday and Thursday 8 to 10pm, and Sunday morning. Whatever it is, it has to survive contact with a normal bad week.

Three focused sessions you actually keep beats seven you plan and skip. I'd rather you commit to four hours a week for a year than twenty hours a week for a month, because the second one ends in March and the first one ends with a job.

Some things that make the sessions hold:

Protect a small, ugly, repeatable slot. The people who make it are almost never the ones with ideal conditions. They're the ones who found 45 minutes before the house wakes up and defended it for eight months.

Start the session by writing code, not by watching. Open yesterday's project first. Read your own last few lines. You'll be back in the problem within two minutes, and you'll skip the warm-up scroll that turns an hour into fifteen minutes.

Stop mid-problem. Hemingway's trick works on code. Quit while you still know the next step, and starting tomorrow costs you nothing.

Track days, not hours. A calendar with an X on it is a dumb little motivator that works better than it has any right to.

The accountability problem, which is the real problem

Nobody online notices when you disappear. That's the whole difficulty, and it's worth solving deliberately instead of hoping motivation covers it.

Motivation will not cover it. Motivation is a mood, and you're going to hit stretches where the thing you're building doesn't work for four days and you feel stupid. That happens to everyone. What carries people through it is some external thing that expects them back.

A few that work, in rough order of effectiveness:

A person. A friend doing the same thing, a study partner, a group chat where you post what you built. Even a weekly text to someone who'll ask. This is by far the strongest one and the most skipped.

A community you actually post in. Lurking does nothing. Posting a broken bit of code and getting three replies does a lot, partly for the answer and partly because now people know you're there.

A public log. Build in public, post your ugly progress. It's uncomfortable for about two weeks and then it becomes the reason you keep going.

A deadline with a person on the other end. Tell someone you'll show them a working thing on the 30th.

If you take one thing from this post, take this: find something outside your own head that expects you back. Willpower is a terrible plan for a nine-month project.

What to actually do in the sessions

Build things. That's the short answer, and it's covered in depth in how to learn coding from absolute zero, so I won't rebuild that whole argument here.

The online-specific version is this. Because nobody is assigning you work, you have to assign it to yourself, and beginners consistently assign themselves too much watching and not enough building.

A rough ratio that works: for every hour of instruction you consume, spend two hours writing code that is not the instructor's code. Follow the tutorial, then close it and rebuild the thing from a blank file. You will fail the first time. That failure is the actual lesson, and it's the one you skipped when the tutorial felt so smooth.

Watch out for the specific online trap of the perfect course. There's always a newer, better-reviewed, more current one. Switching to it resets you to zero while feeling like an upgrade.

How long this takes, honestly

Longer than the ads say. Shorter than you fear when you're three weeks in and nothing makes sense.

Somewhere in the range of six months to a year of consistent part-time work to get to job-ready, and the range is wide because it bends almost entirely around how much you build versus how much you watch. Someone doing four honest hours a week for a year is in better shape than someone who did twenty hours a week for six weeks and stopped.

Anyone stamping an exact number of weeks on it is selling you something. I've never once seen that number hold up, and I've been teaching people online since 2016.

There's a real shortcut though, and it isn't about speed of learning. It's about which door you walk toward.

The part that changes your timeline more than your study habits

Most people learning online aim at the same target: front-end developer, full-stack, the jobs every bootcamp ad shows you. Then they apply into a pile of four thousand other applicants for one junior role and conclude they didn't learn hard enough.

They learned fine. They aimed at the most crowded door in the building.

There are entry-level lanes that hire constantly, don't ask for a degree, and have a fraction of the competition, because they're unglamorous enough that nobody makes ads about them. Email development, martech, AI automation, Shopify and ecommerce, web content management, PHP. Boring on a billboard. Very good on a first payslip.

That's the whole reason CodingPhase exists, and I'll be straight since it's mine. We've been teaching this since 2016, mostly to people who were never supposed to get into tech: parents fitting it around nap time, people who'd had enough of driving for apps at 1am. The membership is built to fix exactly the three things online learning strips out. Guided paths so you're never guessing what's next, interactive lessons where you write real code in the browser, an 80,000+ member community for the 10pm bug, portfolio templates and the projects to fill them, a résumé builder made to survive ATS screening, and a job board at the end of it.

It lists at $60 a month or $365 a year, and there's always at least 30% off, so it comes to about $42 a month. Less than $2 a day. The annual plan lands near $255. Check the pricing page for today's number, and there's a 7-day refund if it isn't for you.

The honest con: free options cost nothing, and if you're genuinely self-directed you may not need us at all. Go take freeCodeCamp with my blessing. We're for the people who've already proven to themselves that unstructured doesn't stick.

FAQ

Can I really learn coding online for free? Yes, and plenty of people do every year. freeCodeCamp, The Odin Project, and CS50 are all free and genuinely good. What free doesn't give you is sequencing, feedback, and anyone noticing when you stop. If you can supply those three yourself, free is enough. Most people who stall didn't hit a content wall, they hit a structure wall.

How many hours a week do I need? Four to six focused hours, kept consistently, will get you there. Ten distracted ones won't. Consistency beats intensity on a project measured in months, and the people who burn twenty-hour weeks in month one are usually gone by month three.

Is learning to code online as good as a bootcamp? For the learning itself, yes, and for a tiny fraction of the price. What a bootcamp actually sells is enforced structure, a cohort, and career services. If you can rebuild those with a path, a community, and your own calendar, you can get the same outcome without the $15,000. I went through the real numbers in what a coding bootcamp costs.

Do I need to be good at math? For most web, ecommerce, email, and automation work, no. You need to be comfortable with logic and patient with things that break. The heavy math lives in specific corners like graphics, machine learning, and some backend work, and you'll know if you're heading there.

What should I build first? Something small enough to finish this week. A page with a button that changes something. A to-do list. A script that renames a folder of files. Finishing a tiny thing teaches you more than abandoning an ambitious one, and the finishing is a skill you're also practicing.

Four dashboards at 11% is not a talent problem, and it's not a platform problem. It's what happens when you take a nine-month project and run it on motivation alone.

Pick one thing. Put three sessions in your calendar. Tell one human you're doing it. Then build something small and slightly beyond you, over and over, for longer than feels reasonable.

That's the whole method. If you want the sequence handled for you so you can spend your energy building instead of guessing, come see the career paths we've built. And if you'd rather do it free and alone, genuinely, go do that. I'd just rather you finish.