Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

It’s Never Going to Be Perfect, So Just Get It Done

nytimes.com

Post reply on HN
Print view View original

101–110 of 135 posts

Re: It’s Never Going to Be Perfect, So Just Get It Done

#101

Post by nicklasss »

andrewstuart wrote: post #78

I still build out everything required to realise the vision, and that usually takes at least months to build. There's an analogy in startups that you should build a great skateboard first if your vision is to build a better car, but I just don't agree. If you have a vision for a better car then you need to deliver a better car, not a skateboard, because, well a skateboard just isn't a car. https://blog.crisp.se/2016/…

I think you are just using the wrong analogy, because the vision is not usually to build something but to achieve something. So the analogy should be something like:

If you have a vision for getting people to move from one place to another in a better way, you should build whatever achieve moving people from one place to another in that better way. Getting rid of everything that is not necessary for you to test if you are really getting people from one place to another in a better way.

update: typo

Re: It’s Never Going to Be Perfect, So Just Get It Done

#102

Post by nostrademons »

andrewstuart wrote: post #93
nostrademons wrote: post #90

Earlier quoted context omitted.

In practical startup terms, if your idea is to deliver either a better car or a better skateboard, you've already failed. Both these industries have a century of optimization behind them; even if you did manage to catch up, any useful innovations would be copied by incumbents. Successful startups usually come out of new markets: areas where the environment has changed such that existing product categories or delivery…

The real question is "what, exactly is an MVP?" Some people interpret an MVP almost literally as almost nothing - just a web page that describes the idea and asks for email addresses. Others interpret the MVP as almost nothing, but a "workable" almost nothing - i.e. some code that users can find. After users have found it, the code will be iterated upon until the original vision is realised. The skateboard people int…

The vision is irrelevant to the customer.

There's an easy way to tell in hindsight whether something was an MVP. Did it get beaten to market by someone who came out with something worse? Then it's not Minimal. Does nobody use it? Then it's not Viable. Do they get bored and move on? Then it's not a Product.

Doing this with foresight, so you're the one who actually puts out the MVP for the market and captures it, is significantly harder. Whoever manages to do that ends up rich, so if it were easy, we'd all be rich. But a general rule of thumb is that if the world is moving on and other people are getting rich while you sit there refining your idea, your idea of "Minimal" isn't actually minimal, while if you're continually pumping out ideas that nobody likes, your idea of "Viable" isn't actually viable.

Re: It’s Never Going to Be Perfect, So Just Get It Done

#103

Post by celticninja »

heavenlyblue wrote: post #85
andrewstuart wrote: post #78

I still build out everything required to realise the vision, and that usually takes at least months to build. There's an analogy in startups that you should build a great skateboard first if your vision is to build a better car, but I just don't agree. If you have a vision for a better car then you need to deliver a better car, not a skateboard, because, well a skateboard just isn't a car. https://blog.crisp.se/2016/…

>> They had tried to build similar systems in the past and failed miserably, mainly because of Big Bang thinking. They told me that one of their previous attempts took 7 years from inception to first release. SEVEN YEARS! By then of course everything had changed and the project was a total failure. So this time they wanted to do it differently. Or well... Maybe it's just that they have failed to build a system becaus…

At least agile, whatever it may be to people, has worked well in government/public sector IT projects. If ever a project was going to require many changes in its development lifecycle it's government. No more 7 year projects with no working product at the end. It was common practice to underbid the price to win the contract and rinse it on the change requests.

Re: It’s Never Going to Be Perfect, So Just Get It Done

#104

Post by nostrademons »

echelon wrote: post #98
nostrademons wrote: post #90

Earlier quoted context omitted.

In practical startup terms, if your idea is to deliver either a better car or a better skateboard, you've already failed. Both these industries have a century of optimization behind them; even if you did manage to catch up, any useful innovations would be copied by incumbents. Successful startups usually come out of new markets: areas where the environment has changed such that existing product categories or delivery…

> even if you did manage to catch up, any useful innovations would be copied by incumbents. If your optimization is sufficiently novel, can't you use intellectual property to protect yourself? If that isn't an option, how do you provide yourself enough runway to scale fast enough to escape the incumbents?

Only if you can a.) detect that the incumbents are copying you (in ways that can be legally protected - in general you can't protect what a product does, only how it does it) and b.) have the money for a legal battle against a well-funded incumbent and c.) know that you aren't infringing on any of their patents or IPs. Most startups don't fit all 3 of these categories.

The way most startups manage to get big is to fight on ground that big incumbents are unwilling to occupy, while their attention is occupied elsewhere. IBM thought that software was a loss-leader and the money was all in hardware: they came out swinging hard with the IBM PC and eventually dominated the market, but didn't realize that leaving the IP for MS-DOS in Microsoft's hands would be a crucial error when the PC BIOS was cloned. Microsoft knew how to fight on platforms and ended up crushing Netscape, but they had no idea that Google (ostensibly "just a search company") would end up occupying the crucial real estate on the web.

Re: It’s Never Going to Be Perfect, So Just Get It Done

#105

Post by ta1234567890 »

nostrademons wrote: post #99
Godel_unicode wrote: post #97

Earlier quoted context omitted.

This sounds an awful lot like the famous HN criticism of Dropbox; don't make a better version of something which already exists. Is there any existing data about this (whether new markets or established ones are more fertile ground for startups)? It seems to me that GoPro did a pretty great job innovating in a highly established market, for example. PayPal is another good case study here.

In the eyes of its users DropBox is a fundamentally different product from the filesharing sites like YouSendIt that already existed, or the techie utilities like rsync that HN commenters compared it to. There's an easy way to see this: people talk about dropping a file in their DropBox, they don't talk about this new filesharing program they're using called DropBox. Similarly, people talk about buying a GoPro and pu…

Your explanation is very clear and really makes a lot of sense.

> The customer's mental model is the only thing that matters here.

Totally right

> If you're building a car or a skateboard, you're seeking to fit into an existing category in their head. Even if you win (which is unlikely, because incumbents optimize very heavily with lots of resources), you lose, because you're a commodity in a market with lots of substitutes.

> If you want to actually win - in the sense of create a monopoly product that defines its own market - you, as an entrepreneur, need to be thinking in terms of the overall problem that the user wants solved (transportation, or information access, or answers) so that you can define the terms of the market in a way that you're the only reasonable solution.

At what point should you start doing that?

Using one of your examples: I don't think the Google founders had a vision of defining a new category of service. And they didn't even have a business model for several years. So what actually made them "win"?

Re: It’s Never Going to Be Perfect, So Just Get It Done

#106

Post by taormina »

paddlepop wrote: post #70
bluGill wrote: post #60

Earlier quoted context omitted.

For software you should make something - any part - work and release it. Then do regular feature releases after that. If you are not working alone you can ask your marketing people for help, sometimes it is worth waiting for a big bang release, sometimes not (For some things if you miss Christmas you should delay until next Christmas). I recommend you err on the side of release too soon - customers are the ultimate a…

This doesn't work so well for video games. Reviews come out saying its thin on content with a score to match and can kill a game before its had a chance to expand. Developers have started trying to do this more often lately with mixed results, DICE and Blizzard are high profile examples of this. Blizzards latest World of Warcraft expansion was heavily criticized for the lack of content on release day and has yet to s…

I'm trying it with my indie game. We'll see it works, but I realized Danger World (https://danger.world) at the end of January this year, and multiplayer is still under active development.

Re: It’s Never Going to Be Perfect, So Just Get It Done

#107

Post by nostrademons »

ta1234567890 wrote: post #105
nostrademons wrote: post #99

Earlier quoted context omitted.

In the eyes of its users DropBox is a fundamentally different product from the filesharing sites like YouSendIt that already existed, or the techie utilities like rsync that HN commenters compared it to. There's an easy way to see this: people talk about dropping a file in their DropBox, they don't talk about this new filesharing program they're using called DropBox. Similarly, people talk about buying a GoPro and pu…

Your explanation is very clear and really makes a lot of sense. > The customer's mental model is the only thing that matters here. Totally right > If you're building a car or a skateboard, you're seeking to fit into an existing category in their head. Even if you win (which is unlikely, because incumbents optimize very heavily with lots of resources), you lose, because you're a commodity in a market with lots of subs…

I think the Google founders had a vision of solving a problem. They wanted to make search work right.

(Well, technically, they wanted to get their Ph.Ds - the actual history of Google was that Larry had the idea to download the web and throw away everything except the links, so he did that. Then he had the idea to visualize it as a big eigenvalue calculation on the probability of a random surfer ending up at a given page, and that's where he got the idea for PageRank. Then someone - might have been him, Sergey, or Scott Hassan - got the idea to sort webpages by PageRank, and the results were better than existing search engines, and so they launched BackRub as a search engine. Then Andy Bechtolsteim thought it could be a really big company and gave them $100K to make it one, and the rest is history.)

If you steer towards problems, and do so without fear of the unknown, you'll naturally move towards ideas where you get to define the category. They wouldn't be problems if someone had already solved them! The rest - building a business, making money, marketing it, defining it in customers' heads, etc. - can all follow later.

Re: It’s Never Going to Be Perfect, So Just Get It Done

#108

Post by mamon »

nostrademons wrote: post #102
andrewstuart wrote: post #93

Earlier quoted context omitted.

The real question is "what, exactly is an MVP?" Some people interpret an MVP almost literally as almost nothing - just a web page that describes the idea and asks for email addresses. Others interpret the MVP as almost nothing, but a "workable" almost nothing - i.e. some code that users can find. After users have found it, the code will be iterated upon until the original vision is realised. The skateboard people int…

The vision is irrelevant to the customer. There's an easy way to tell in hindsight whether something was an MVP. Did it get beaten to market by someone who came out with something worse? Then it's not Minimal. Does nobody use it? Then it's not Viable. Do they get bored and move on? Then it's not a Product. Doing this with foresight, so you're the one who actually puts out the MVP for the market and captures it, is si…

>> The vision is irrelevant to the customer.

Having vision for a product means correctly identifying customer's problem and proposing a solution for it. Dropbox vision is: "Make user files accessible anywhere, by storing it in the cloud synching them across devices". Google's vision is: "Make all information in the world easy to find and access, by having search engine know you better than you know yourself" :)

Customers either buy into your vision, or ignore it, and the latter means that the problem that you stated does not exist or is not important enough to your users.

For you, having clear vision helps you create MVP, because you can easily identify which features are crucial, and which are nice to have.

Re: It’s Never Going to Be Perfect, So Just Get It Done

#109

Post by james_s_tayler »

celticninja wrote: post #103
heavenlyblue wrote: post #85

Earlier quoted context omitted.

>> They had tried to build similar systems in the past and failed miserably, mainly because of Big Bang thinking. They told me that one of their previous attempts took 7 years from inception to first release. SEVEN YEARS! By then of course everything had changed and the project was a total failure. So this time they wanted to do it differently. Or well... Maybe it's just that they have failed to build a system becaus…

At least agile, whatever it may be to people, has worked well in government/public sector IT projects. If ever a project was going to require many changes in its development lifecycle it's government. No more 7 year projects with no working product at the end. It was common practice to underbid the price to win the contract and rinse it on the change requests.

It makes perfect sense Agile works a lot better for Government projects. Also probably for agencies too.

I'm not 100% on SaaS. Need something more fine-tuned. Call it SaaSGile.

Re: It’s Never Going to Be Perfect, So Just Get It Done

#110

Post by sleepydog »

nostrademons wrote: post #91
paddlepop wrote: post #70

Earlier quoted context omitted.

This doesn't work so well for video games. Reviews come out saying its thin on content with a score to match and can kill a game before its had a chance to expand. Developers have started trying to do this more often lately with mixed results, DICE and Blizzard are high profile examples of this. Blizzards latest World of Warcraft expansion was heavily criticized for the lack of content on release day and has yet to s…

It doesn't work for AAA games, but it can work really well for indie games. Think of something like Factorio, where version 0.1 looked like this: https://www.factorio.com/blog/post/fff-184 Then they just kept releasing every week for 7 years, and now you have people building CPUs and explaining Apache Kafka with it. Ostriv is another recent game that comes to mind as having a similar development cycle. Also many F2P…

While we're on the topic of Factorio, I find that even from a Factorio player's perspective, the OP is good advice. I've wasted so much time trying to build the perfect base and starting over after building myself into a corner, when really I should have just built something that works "right now" and make iterative improvements over time.
Post reply on HN