--
“So my friend tells me you do programming…could you build a site like YouTube?”
This is the big opportunity that my friend told me about, the phone call that I needed to drop everything for and take.
Sigh.
I get this call once or twice a year. Or it’s an email. Or someone sees me coding in a coffee shop and says Hello. If you’re an experienced engineer or architect with a decent network of business types in or around a startup community, I would imagine you do too.
The answer to the question, by the way, is Yes.
And No.
And You don’t understand.
This weekend I sat down and actually thought about what I mean by all that. What is it that two talented people — for surely we all know that a programmer isn’t going to build anything working by himself, that she/he needs a partner in crime, to help design, to guide, to inform, to shape and steer what’s being built — what is it that two people could realistically build in a brief amount of time?
Here’s my theory:
Two talented people working semi-full-time-ish by themselves, with minimal funding, can in 3–9 months build 80% of (almost) any app.
80% of Facebook in four months, no prob. It’s CRUD for people with connections between people and message boarding with likes.
80% of YouTube in three months, sure. I did that one time.
80% of Uber in nine months, yup.
80% of Google… no. You couldn’t build 80% of Google in any time. Google is special. Google is all algorithm.
Not Amazon either. Amazon is all logistics and distribution network. You could build 80% of the web app, but that’s not ‘Amazon’.
80% of Twitter. Instagram. Github. Pinterest. Reddit. Tumblr. Yes to all of the above.
What is 80% of YouTube worth?
Not what you think it might be.
In my experience, 80% of the way there in a web or mobile app is either
A) a half baked contraption that kinda-sorta works if you feed it exactly the right inputs… IOW it works great when the founders demo it, but falls flat on its face if a real user shows up
OR
B) a bright, shiny facade over a totally faked backend… “don’t look behind the curtain!”
So wait, am I telling you that what we’d end up with - after four months of sweat and tears working on our YouTube clone - would be merde, would have no value?
Not at all. It would definitely have value. As a prototype, something to prove product/market fit; as a demo to support a pitch to raise money, get interest in an idea, get into a startup accelerator.
To raise the $$$ we’d need to take the next step.
It would not be anything that people would want to use. This is not going to be the successful commercial product. Sorry.
That’s a source of misunderstanding to a lot of would-be founders, and I wish it was better known. That first thing you build is not going to be The Thing. It’s going to be the bridge that gets you the funding to build the team to build The Thing. And that takes commitment, guts, moxie, a network, a whole host of skills and talents and assets that makes the Business Founder the valuable person he/she is.
How to build something that’s actually useful to someone
Because after building that initial 2 Person Prototype, you’d need another 80 / 20 iteration to achieve an application product that people would actually use and potentially pay — with money or with their time or oh-my-goodness do-we-dare-dream with BOTH — to use.
This next 80 / 20 iteration is usually done with roughly three to eight people. Three to eight people, within 2 years, can build a decent product in a single vertical. I’ve seen this. The group I work with right now did this. The thing can make money… lots of money.
The thing that this group would build will, if product-market fit is there, make $$$.
It will also not be scalable, and at some point it will hit a brick wall of technical debt due to the shortcuts the team has to take to get there.
To expand, to hook up an advertiser app, or go after another market, or add customization for big clients, guess what? You need another 80 / 20.
The Third 80 / 20
This next 80 / 20 iteration takes, in my experience, 20–100 people, three to five more years, and might involve rewriting a lot of what the group of three to eight wrote to make it scalable. This is to achieve a second and third vertical…. maybe an advertiser application and accompanying BI systems that works with the original application. Or maybe a direct-to-consumer piece to disrupt the distribution channel. You are still not at Facebook yet, but you might be Pinterest.
Three iterations of 80 / 20 is 99.2% of the way there. Maybe that’s enough.
The Fourth and Fifth 80 / 20
The next 80 / 20 iteration takes hundreds into the thousands of people. I think. This is where my experience wanes… I’ve spent my entire career in the first three iterations. I’ve never made it this far.
I would imagine that this iteration might get you in the ballpark of what YouTube has in terms of performance and reliability. After four 80/20 iterations you are, if I still know how to do math, 99.84% of the way there. A billion users might even be doable.
The final 80 / 20 iteration would then get you in the ballpark of Facebook. Signup the whole planet and leverage an incredible network of network of opportunities. This takes thousands of people and a decade.
I’m not going to even try to touch on the why of all this, why it seems to require such an exponential increase in effort to get to those last fractions of a percent… that’s a whole subject matter of its own. I’m simply presenting this to have something to point someone to when I politely decline their offer of helping them build the YouTube Of Portland.
(YouTube, of course, is the YouTube Of Portland… that’s a whole ‘nother post too)
This post was partially inspired by an annual call with a frustrated would-be founder, and partially inspired by bumping against Quora posts like this https://www.quora.com/Why-does-Facebook-have-so-many-employees-Are-they-all-doing-customer-support-I-know-by-experience-that-you-don%E2%80%99t-need-more-than-100-developers-to-take-care-of-a-big-site-like-Facebook-Is-this-just-a-farce-to-keep-the-value-of-the-company-high
