Founding an online startup. Do I need to learn programming?

[whimsyjim]

So I have what I would consider a fantastic business plan with the online experience to back it up. I've done my research and have 3 very serious investors that want to invest as soon as they can.

The problem: The idea centers around marketing (I'm one of the best at SEO. I work for banks and fortune 500 companies) AND a machine learning mechanism to decide what content to deliver to the visitor.

The question is should I try to learn php and take the free Machine Learning course Stanford is offering right now? The other option is keep sub-contracting to development agencies, collecting my paycheck after all 10 hours of work, and hiring a ninja-programmer.

Another small issue, the tech talent in my area, Wilmington, NC, isn't the greatest.

I'd love to hear everyone's thoughts on this.

[achompas]

AND a machine learning mechanism to decide what content to deliver to the visitor.

This sounds like a pie-in-the-sky idea. How can you develop a startup around a ML algorithm when you can't code and have never studied ML? Do you have a technical co-founder?

What kind of "mechanism" (these are actually called algorithms or machines) will you use? What frameworks or languages (NumPy/SciPy, pandas, Weka, Mahout; Java, Python, R) will you use to implement this algo? Do you know what data you'll analyze? Or how you'll normalize that data?

Get to the point where you can answer these questions and you'll be on your way. Until then, you'll struggle to find a technical cofounder (or, if you have one, you'll struggle to communicate with them and evaluate whether they are good at what they do).

tl;dr: unless you have a technical co-founder, you'll need much more than a Stanford ML course and some PHP to make this work.

[mjs00]

If you weren't already intrigued about learning to program prior to this startup (and already doing something about it), the answer is 'no'.

You need a technical co-founder, or a technical partner/adviser you trust that can help guide development. What you miss with a development agency/contractor-only approach is the critical technical feedback loop and iterations that strengthen and develop your idea that you'd typically do with that co-founder. Agencies/contractors will say 'yes' to what you want (and are happy to take your money), absent the really rich and valuable feedback and collaboration you get from a technical partner who has a vested interest in a shared vision.

You are only 2 hours away from Raleigh which I understand is pretty good for for tech talent. (http://www.fastcompany.com/1699323/why-you-should-start-a-co...). Maybe set up shop there?

[oscarleung]

Personally, I think you need a committed partner who can code. Like Steve Blank/ Eric Ries like to put it, a Startup is not a company - it is a hypothesis.

It's great that you have a business plan, but a business plan is a remnant of the 80s and 90s. What you really have is a hypothesis about how you believe the world is. You need to get your product out there to see if your hypothesis holds true. If you've done your research, it will be close to the mark. From there, you'll have to iterate your product until you've reached a product-market fit.

And to iterate rapidly, you cannot keep subcontracting coding. You're going to need a committed CTO that will be able to deploy code rapidly.

That's my two cents.

[whimsyjim]

It's funny you say that because I am on your side here, but a close mentor of mine who has successfully founded, raised venture funding, and sold 3 tech companies disagrees. He recommends doing as much as I can for now and then get funding rather quickly. I'm not saying he's right but he has some pretty good points to back up his recommendations.

The only problem I foresee with getting a committed CTO is finding one around my area. I had one incredible programmer who had just quit his job at Match.com, but he got offered a job he couldn't refuse with equity in the venture funded startup in the valley.

I guess I just need to start the process of finding the perfect fit. I'm probably going to reach out to Raleigh for this.

[oscarleung]

It can definitely be done like your mentor said. As a Rails developer myself, I'm kind of biased to coders being part of the DNA of the company and having a strong stake in it.

This is a great article on humbdledmba for non-tech co-founders: http://www.humbledmba.com/please-please-please-stop-asking-h...

They make a good point. The coder's biggest is fear is wasting his/her time. Do as much as you can to alleviate their fear. For example - start a blog, get a following.

I hope this is helpful for you.

[revorad]

You're unlikely to learn php AND machine learning in a short amount of time. You should find someone to work with.

Move to a tech hub. If you're not committed enough to do that, you will have a tough time finding someone who'll build a product for you.

And learn to code anyway. You won't turn into a ninja any time soon, but it's another way to prove you're committed and technical partners can't bullshit you.

Most importantly, build your product without the machine learning. Do it manually. Do it in wordpress. It's just marketing content, right? If you're one of the best, you can prove the concept without code, then get someone to automate that.

[glimcat]

It takes a lot of time to become good at this.

HOWEVER:

I strongly feel that anyone who is going to be working with or managing programmers should spend at least some time learning to code. You will not be able to do their job, but you will have a much better understanding of the basic issues involved.

As a technical person, I also have much more respect for someone who wants to talk to me about an idea they have after they have taken the effort to make some sort of rough mock-up or prototype. I may have to throw it out and start over, but it shows commitment and it shows that you have been doing some serious thinking about the problem.

[JoachimSchipper]

Machine learning is hard. Depending on the quality you need, you may need to find an expert. (Also, I'd be surprised if you end up with calculating recommendations in PHP on every page load.)

[asto]

Dude! You're like that guy from that "how to spot a bad non-technical co-founder" article. Pretty sure this is a troll.

[whimsyjim]

Not sure who "that guy" is, but I can assure you I'm not a troll. I've been following ycombinator for a long time, I am just new to the Hacker News Section.

-Jim

[asto]

Okay, I'm pretty new here too. Here's the article I'm referring to. It's quite interesting, give it a read

http://blog.hirelite.com/warning-signs-in-a-non-technical-co...

[willpower101]

can I email you? I'm actually interested in discussing your work with seo.

[coryl]

Banks and Fortune500's use SEO?