Is releasing early ever a bad idea?
14 comments
working on a startup is like working on a recipe for cake. bake your cakes, let people try them, adjust and add to the recipe, and then bake another cake. the point isn't the cake itself, the point is the recipe.
so, get your recipe together and bake a cake. maybe to start its just a plain, vanilla, one-layer, only bread cake. you bake it and give it to people to try. if you give it to people to try before you bake it, they're eating raw batter and will get salmonella poisoning and die. if the people like the plain cake, now add some frosting or some berries or another layer. if they don't like it, fix it.
i'm hungry
so, get your recipe together and bake a cake. maybe to start its just a plain, vanilla, one-layer, only bread cake. you bake it and give it to people to try. if you give it to people to try before you bake it, they're eating raw batter and will get salmonella poisoning and die. if the people like the plain cake, now add some frosting or some berries or another layer. if they don't like it, fix it.
i'm hungry
Great metaphor!
Releasing early doesn't mean releasing crappy software-- it means trimming your feature-set down to be as lean as possible while still offering some value.
That's a starting point that can help you learn about what makes your users tick.
For a game, I don't know how this applies. If you know your game isn't fun, I don't think you should release it. If you aren't sure if your game is fun ENOUGH, I think you should.
FWIW, "releasing" isn't all-or-nothing. Start by releasing to 10 friends. Then 100 strangers. Then 1000 strangers.
That's a starting point that can help you learn about what makes your users tick.
For a game, I don't know how this applies. If you know your game isn't fun, I don't think you should release it. If you aren't sure if your game is fun ENOUGH, I think you should.
FWIW, "releasing" isn't all-or-nothing. Start by releasing to 10 friends. Then 100 strangers. Then 1000 strangers.
For a game, the fundamentals have to be right, and you can usually tell that early on, but the 'fun' can take a long to get right.
I'm going to polish to a point where I can take feedback from friends and iterate from there.
We're launching a slinkset site next week, thought I'd post to HN and ask people the important question: amidoinitrite?
We looked at just putting it out but thought we'd try and build up a community first through a beta programme. We've had some good feedback but to be fair our choice of beta users wasn't very good. We could've easily have ditched 25 of our 30 beta users and had the same amount of activity.
What we did get from those remaining 5 was all the delicious feedback we could lay our hands on, some of which we've passed on to slinkset.
I've also written to a number of key bloggers in the niche we're working in to promote the site, hopefully you'll see some noise. Slinkset have also been very supportive.
I think if we'd have just thrown it out then it'd have been stillborn. I think the fundamental thing is to have something functional 'enough' and to have some users when you start, then go for the launch. What do others think?
We looked at just putting it out but thought we'd try and build up a community first through a beta programme. We've had some good feedback but to be fair our choice of beta users wasn't very good. We could've easily have ditched 25 of our 30 beta users and had the same amount of activity.
What we did get from those remaining 5 was all the delicious feedback we could lay our hands on, some of which we've passed on to slinkset.
I've also written to a number of key bloggers in the niche we're working in to promote the site, hopefully you'll see some noise. Slinkset have also been very supportive.
I think if we'd have just thrown it out then it'd have been stillborn. I think the fundamental thing is to have something functional 'enough' and to have some users when you start, then go for the launch. What do others think?
In one word: Cuil.
How would you rate their chances now that they established themselves as a high-profile failure?
How would you rate their chances now that they established themselves as a high-profile failure?
Like others have said, Cuil's failure was a huge PR push, not the release. I'm sure that they did actually get a ton of useful feedback from it that they can use to improve their service. In this case, though, it came at the cost of being branded a failure by many people.
If they had avoided the boastful claims and stuck to low profile releases, they could have gathered useful feedback without the expectation of being a great search engine. Perhaps if they had tried this they actually would have managed to get themselves up to google's level by the time the general public became aware of them.
If they had avoided the boastful claims and stuck to low profile releases, they could have gathered useful feedback without the expectation of being a great search engine. Perhaps if they had tried this they actually would have managed to get themselves up to google's level by the time the general public became aware of them.
Cuil screwed up by doing the full-court PR press before the product was ready. They could have launched the product in a more low-ley fashion without the backlash.
I think it was a good idea. Now they can move on to more productive enterprises instead of working longer on a bad idea.
The idea isn't a bad idea; Google isn't going to be the best search engine forever. (there are friends who swear that Yahoo has better results for them); Eventually they'll be unseated. just not by Cuil, because they did it so badly, and then killed themselves with bad publicity
Damn you beat me to it. They may have gotten away with it too if they'd laid low for a bit before to let the dust settle before the PR push.
Absolutely. First looks do matter and people will be less inclined to check your site out again if their first experience was awful.
An even bigger issue is that if you're entering a crowded market, you'll be signaling to your competition what your approach and features will be. If you've developed something that's not magnitude better, you will be copied and if they have resources, they will outmaneuver you.
I'd suggest people read some of the Michael Porter books on this.
An even bigger issue is that if you're entering a crowded market, you'll be signaling to your competition what your approach and features will be. If you've developed something that's not magnitude better, you will be copied and if they have resources, they will outmaneuver you.
I'd suggest people read some of the Michael Porter books on this.
Good point, thanks for the reference.
Releasing early needn't be the same thing as promoting early.
But every case is a special case. In the case of a game, you might want to get it in front of someone who isn't a developer as early as you can, but that doesn't mean you want to spoil your big marketing splash with a long, slow public beta. This is what roommates are for!
But every case is a special case. In the case of a game, you might want to get it in front of someone who isn't a developer as early as you can, but that doesn't mean you want to spoil your big marketing splash with a long, slow public beta. This is what roommates are for!
The idea of "release early" is to get early feedback from a customer. In the context of games, you may want to present the idea of the game or some screenshots to a small circle of gamers to get their feedback about if the idea is funny enough. It probably does not make sense to release more than one version of a game, except your game is evolving permantently with new levels, new creatures, new ideas.
I agree. Games tend to be much less malleable than web applications. For one, web apps are much easier to change after release than other products because there is no distribution problem - just push the code and you're done. And in general, users have a more well-defined set of expectations from a game than from a web app - if you miss those expectations, you can be toast.
I guess one way to push the release date up would be to cut some of the content and pack it in an expansion pack or later update. But the 1.0 version has to be solid, otherwise no one will stick around long enough for the expansion.
I guess one way to push the release date up would be to cut some of the content and pack it in an expansion pack or later update. But the 1.0 version has to be solid, otherwise no one will stick around long enough for the expansion.
I'm making web based games so it is very easy to tweak and re-release (people don't even have to know you've changed it) but there has to be a minimum level of quality and fun to begin with.
Starting with a single, well polished level, and going from there sounds like the best strategy. I can add more or change it based on feedback.
Whether you're a startup or established company, I would say "no" for games-related products, but that hasn't seemed to stop Microsoft.
Game-related services are a different subject btw. Developing games for advertising, doing consulting for game dev companies, etc, is a much more safer bet than making game-based products. However, the game industry doesn't pay as much as other industries, whatever you do in there (at least that was my experience).
Or Sony Online Entertainment.
They have many years and many games under their belts where people pay to beta test their software for them.
They have many years and many games under their belts where people pay to beta test their software for them.
It always takes them three goes!
If it's a web site, I'd suggest that you launch the smallest high-quality product you can, and then tweak or add something every day until it's where you want it to be. You'll change track along the way, but that's OK because you'll do it based on real user behavior.
I launched Planaroo.com on June 30, and I wish I had launched it two months earlier. Many of my assumptions about how people would use the site were just plain wrong. I spent a long time working on features that few people use, and not enough time on features that a lot of people use.
If your site involves user-generated content, you have no idea how long it will take to ramp up. If your has lots of text, you just won't know how it will do in search for weeks after launch. Planaroo did very well in Google search almost immediately, but it still doesn't show up in Yahoo Search after 10 weeks.
I launched Planaroo.com on June 30, and I wish I had launched it two months earlier. Many of my assumptions about how people would use the site were just plain wrong. I spent a long time working on features that few people use, and not enough time on features that a lot of people use.
If your site involves user-generated content, you have no idea how long it will take to ramp up. If your has lots of text, you just won't know how it will do in search for weeks after launch. Planaroo did very well in Google search almost immediately, but it still doesn't show up in Yahoo Search after 10 weeks.
The "release early" people are advising a black/white solution to something that is very 'gray'. What they mean is don't spend years on your app and then only release it when it's 100% perfect.
My advice? Don't let perfect ruin good (see DNF).
My advice? Don't let perfect ruin good (see DNF).
> So, if your startup is games based should you still release early?
No. DAOC had a great launch and continues to be fairly popular. Vanguard had an AWFUL launch and will likely never really recover.
No. DAOC had a great launch and continues to be fairly popular. Vanguard had an AWFUL launch and will likely never really recover.
Don't release early unless the features you release are really polished. I never used Flock because it crashed when I tried it first.
Google didn't release Chrome till it was good enough to be released. This, from a company which completely believes in releasing early! Of course Chrome is still beta, but there's a world of difference between "Google releases another mediocre product" and "Chrome is amazing!"
One less cool feature doesn't mattter. Take it from Chrome: "we want this product to be ROCK SOLID"
EDIT: Several commenter have pointed out that I shouldn't have said 'Get something up and running ASAP', instead it should be the 'minimum value adding set of features at good quality'. It's a bit less pithy though.