Human Intervention as a Competitive Advantage | Hacker News Reader

Human Intervention as a Competitive Advantage

sivers.org

140 pointssinak34 comments

This is fantastic advice. Being small can be an asset, yet it's often treated as a liability. Sure, there are some things your company can't do because you don't have as much cash in the bank as Google, but there are some things your company can do that Google can't compete with because it's too big.

The size is neither an asset nor a liability - its just an attribute.

However size can determine whether you can do certain things. Unfortunately, small size prevents you from making profit from economy of scale, and so you must somehow differentiate with other attributes (such as personalized services).

But i think the market has spoken - large size business is more valuable and thus make more profit than a small sized business.

Googling "what percent of economy is small business" yields

http://usgovinfo.about.com/cs/businessfinance/a/sbatopten.ht...

10. Small businesses make up more than 99.7% of all employers.

9. Small businesses create more than 50 percent of the nonfarm private gross domestic product (GDP).

From this it follows that "not-so-small businesses" create the other 50% of nonfarm private GDP, despite being only 0.03% of employers right? So what is the conclusion exactly?

The conclusion is simple - economies of scale for larger businesses are incredibly powerful.

To me it says that small businesses are the long tail, but large businesses are also interesting and valuable in their own right. Both are going to be around, and a "healthy" distribution is probably good for us.

That small businesses do not control resources (natural or otherwise) from which production can be extracted at a high per-employee ratio.

6. Small businesses employ about 50 percent of all private sector workers.

Not-so-small businesses do have slightly higher production per employee, but not that much more (assuming that "more than 50%" implies "less than the next round percent after 50").

What it really means is that 99.7small-business-size == 0.3large-business-size ==> not-so-small businesses average 330x as big as small businesses.

Which is perfectly understandable, for example my employer has something like 10k employees.

The conclusion is that you're misinterpreting that number. Small businesses are 99.7% of employers, but do not hire 99.7% of employees and do not represent that fraction of business activity. "An employer" is not a meaningful unit of economy because it varies so wildly in size.

This argument is like comparing the United States to the other 199 countries in the world and concluding that the US should represent 0.5% of world economic activity. It's mistakenly treating entities of vastly different sizes as equivalent in the same population set.

>But i think the market has spoken - large size business is more valuable and thus make more profit than a small sized business.

The market has spoken? Do you actually know anything about "the markets"? The first thing a company does when it can is start investing in politics so they can erect barriers against competition. The market hasn't said shit because it wasn't asked.

I completely agree. When you're a small startup, few things are in your favor. It's worth every minute to pour over creating a great experience, even if it's done by hand.

@MassageJoy (http://www.massagejoyspa.com), we're currently going through Pinterest and the wider web, and pulling things together to create thoughtful Care Packages for our clients.

Is it hella expensive? Yes. But that doesn't matter. The basic point is to make sure the experience that we deliver is unmatched. And we're either going to do that or die trying.

Are you a sexual service? Maybe you should consider posting on backpage.. "Book Online in Secs" sounds like "Book Online in Sex"...

We are NOT a sexual service - I'm not sure if you were joking, but I'm honestly offended. At the same time, if you were being serious, I'm happy that you asked instead of making an assumption.

One more time: We ONLY do professional massage therapy.

I think he is being genuine but a bit brash. The advice is sound and I second it.

You should consider replacing the stock images by your own photographs! I think that would add to the authenticity and professional appearance of your service. It might even help against first impressions like verbalist had ;-)

That's definitely on our to-do list.

When people ask me "So what technology are you using for transcription and sound editing?" I answer "Humans. The end product is to be used for marketing. Computer technology is not good enough today. That is also why my price per transaction is so high".

Them: "But that will not scale, your costs will be too high!" Me: "I really want that problem. I also want to have the problem of having to pay too much in taxes"

Love how the acronym for human intervention is "Hi".

:-) Thanks for catching that.

Very good article (as always from Derek). I'd just like to point out something he didn't mention in the article: Derek was coming at this mostly from a "User Experience" angle. But there's a very clear business angle to this as well.

While most startups gloss over this fact, programming time costs money. Sure, if it's you doing the programming you're not actually paying yourself money, but that's a fiction - you're costing yourself a ton of money on opportunity costs.

So when you're trying to make a decision on whether to automate or not, you should also take into consideration how much money it costs to automate, versus how much it costs to program. A good programmer costs thousands of dollars a month, so spending a month building an automated billing system (for example) isn't a good business decisions if the alternative is paying someone $100 a month for going over 30 customers and manually doing the billing.

Not to mention the fact that a month spent building things that can be cheaply automate is a month farther from getting market validation for your product.

NYTimes article on a similar theme: http://www.nytimes.com/2013/03/11/technology/computer-algori...

I think this is good advice but I have to nitpick one thing: as a former music director for a popular indie rock station, I have to disagree that listening to new music all day is "The coolest full-time job in the world."

Joking aside, we adhere to his/pg's do more work principle almost to a fault. I constantly try to make it such that my customers have to do as little as possible to get our products to work for them and so far that has proven valuable.

This sounded less like the traditional editor's job of "wading through the slush pile", and more like "clustering great albums against other great albums from a pre-screened pool." Must be nice ;)

His advice really is 'good user experience is good; bad user experience is bad.' Human intervention is not essential to good user experience in most of his examples; he compares the human versions to poor implementations of automation. A form with some validation and normalization is probably a better user experience than having a human review the fields.

I wouldn't say that was his advice at all. I read it as being something more like "there isn't always a (known, or sensible-ROI to research) way to make an automated user experience anywhere near as good as what you get by just having a human do it."

Great advice. And quite a cool story!

It's interesting to note that Derek mentions about doing this (sort-of) front-line work yourself, but then outsource it when it grows so much that it would take too much of your time that you put on your core activities. Some entrepreneurs stay in the first phase or move from it to automation too quickly.

It's interesting reading this story, if I were in the author's place, I would have felt embarrassed to admit this point. Like our company isn't technically adept enough to implement a recommender.

Technically adept or not, his human-based recommender was so good that the CEO and VP of a 'hugely funded' online music company were willing to fly all the way from SV to NYC to hear about it.

Very few companies are adept enough to do so. Sure, implementing a similarity metric is easy. Scaling is a bit harder.

How do you optimize? Are people clicking on the links? Are they buying? How many even know they should be optimizing for those metrics, let alone consciously addressing more advanced trade-offs of LTV vs short-term sales?

Even worse, with less data there are more problems with every class of algorithm.

The point is not be technically proficient, but to create a good business. The best business is not always the best technically. (This point is often painful for programmers with a limited world view).

Is it too pessimistic to expect not to scale? How well has Apple's App Store reviews scaled? Aren't Apple still using humans to test apps before letting them on their market?

Any tips on how to hire/contract a team with some reputable local manager from China/India/a third world country to do this at reasonable prices, while staying physically in the U.S. or Europe?

Because obviously this is not feasible at the uncompetitive wage levels + confiscatory taxation in the "first" world.

This is what I do. Please feel free to contact me (details in my profile).

Isn't that how Yahoo worked in the early days? They had humans categorize web sites? Well, the only way to make that scale massively is exactly what amazon (mentioned in the article) currently does...have human reviewers work (but for free.) Google essentially did the same thing with pagerank.

Reply on news.ycombinator.com