Entrepreneurship · Startups

At what point do you sacrifice perfection in order to release a product?

David Feldman Data Analyst at Scribd

June 12th, 2015

It difficult to determine when you should sacrifice form (if ever) over function for a first-time entrepreneur. Is it better to make a product perfect, even if it means waiting to release past the set launch date? Or should you release it, even if it’s not the best it can be yet?

A great idea is 1% of the work. Execution is the other 99%. In this course, we’ll teach you how to conduct market analysis, create an MVP and pivot (if needed), launch your business, survey customers, iterate your product/service based on feedback, and gain traction quickly.

Dan Maccarone Co-Founder/CEO at Charming Robot

June 12th, 2015

No product will ever be perfect. Founders who think this end up in launch paralysis because there will always be tweaks that can be made. Big bugs that are glaring like major functionality not working are things to fix, but I'm a fan of getting stuff out there in the world, letting people use it, learning from it and making it better. You'll find out a lot of what you did wrong (and right) when you launch your product. Don't be afraid of being wrong because, inevitably, you cannot predict how people will use the product as you are not necessarily the user (even if you are A user). When you focus on what the product needs to do in order for users to (easily) accomplish their goals, you'll find that you can strip SO much out of your product so that any glaring major issues are obvious. Otherwise, get it out there and get people using it.

Axel Schultze Founder Society3 Accelerator & Fundraising market place

June 14th, 2015

Get it out as soon as

1) Your MVP does what you have defined as the most differentiating thing to others
2) Publish a list of things that it doesn't do yet (known issue list) so people know what to expect
3) Let people know it is an early prototype / beta product

Get it out as quickly as you can and give your users a chance to become major influencer in your product development. Then you build essentially a product the market is asking for. Hard to go wrong with it.


John Anderson

June 12th, 2015

The key is making sure you have enough functionality to be of users to your clients/customers, and make sure the functionality that is present is of high quality.  You don't really need to "sacrifice" features, just have a good roadmap that can show you and everyone inside/outside of the company when certain features are scheduled to be included throughout the up and coming development cycles.  It's much better to have a product with 7 well thought out and executed features, than 20 features that feel like they were thrown in.

As long as you have a good roadmap, you can mitigate responses about this or that feature not being in it.  When people see the good features you have currently, and see the schedule for when other features are coming in, they will be more likely to wait.

David Kurtz Chief Product Officer at Opera Mediaworks

June 12th, 2015

Not to be flippant, but my answer is: Immediately. Perfect is the enemy of done.

Perfection never occurs, so waiting for it is a fool's errand to start with; someone will always come up with something else you can add that will make it "perfect.".

Getting close to perfection only occurs because you get feedback from your customers on how they want to use your application. It's hubris to believe you can achieve perfection before customers touch your product and teach you how they want to use it.

That said:
  1. A release has to have a purpose, it has to meet a need. You can't release until a user can succeed at the set of use cases that go into that need.
  2. It has to have a level of usability to insure that users can actually succeed at using it without frustration setting in. 
  3. It has to have a level of polish so users who encounter it don't think: "it looks like crap so it must work like crap." But remember that doesn't always mean massive effort on UI: just look at Google's home page.

Bill Kelley

June 12th, 2015

We've established the term MVP for just this reason. We are ratcheting down expectations. 

I would generally suggest the MVP priorities are 1.) reliability, 2.) demonstrate key benefits, 3.) cosmetics. 

Reading between the lines, there is another factor that coule be playing behind your question: the infamous "feature creep." The leadership of the company needs to decide what is necessary and what is 'nice to have.' And they need to stick to it. Note: this is far easier to say than to do.

Dimitry Rotstein Founder at Miranor

June 12th, 2015

"If you're not embarrassed by the first version of your product, you've launched too late" (Reid Hoffman, Founder of LinkedIn).

Seriously, perfection doesn't matter. I would go even further - in many cases, at least when it comes to websites, you can launch a product that barely works, if at all (except maybe analytics and "contact us" form). Come on, what's the worst that can happen? A swarm of angry users will flood your mailbox? Good - it means your product is needed and you can go on and fix it, and now you will have plenty of angry user feedback that will help you concentrate on the stuff that matters. But a far more likely scenario is that almost no one will use your website, and no one will ever find out that it doesn't work as it should. Think how stupid you would feel to have perfected your product for nothing (I know this feeling all too well). With a total junk release at least you saved your resources for the next round.
So why not launch right now? Are you afraid that the users will be so angry they will never come back? Doesn't matter - others will come. If there are no others, if you exhausted your customer pool on the first launch, then you have no market anyway - move on to something better. Besides, human memory is short - in a few months these first users will forget all about your fiasco and come back to the new and improved product.

Paul Self

June 12th, 2015

Bill is dead on right. I would not start development without an MVP definition. Otherwise it is pure feature creep and you will have to "shoot the engineer and ship the product". 

I would further suggest creating a Customer Requirements Document and breaking that into various development phase with the first being the MVP. When you start with Customer Requirements, then you have to talk to customers and find out what they will pay for. You are not the customer and what you like is not important.

How do you eat an elephant? One bit at a time. 

Steve Everhard All Things Startup

June 16th, 2015

>And yet you do think like a Toyota right now :-)

Dimitry I have launched HW products and youclearly haven't. The worst situation is that you sell one and get sued. If you don't sell anything you are in prototype mode and not MVP.Prototypes have to be handled carefully and recovered in a time limited way. It is a different proposition to software. We were talking about a commercialrelease and not prototypes here.

And you'll never get good feedback from friends and family - you need a broader audience that merely those people who want to encourage your efforts. That is true for hardware and software. A EULA will not keep you out of trouble if your device is faulty from a safety perspective. Giving away hardwarewill not save you, unlike software, and marking it as experimental will bring the wrath of a government down on you.

I would stick to advice on software if I were you.

Steve Everhard All Things Startup

June 16th, 2015

 >Who cares if you have a bug somewhere if no one will ever discover it?

Interesting. How do you ensure that all your bugs are undiscoverable? The idea that you can abuse your market and merely start-up again is not borne out by experience. If you launch a product with poor initial reception you better have deep pockets to try and pivot that into something good. Reserving functionality for later is good product planning as long as what you have provides utility. There are very few markets where a dysfunctional product will leave your users gagging for your next product - I can't think of one.

In crowded markets the tolerance for bad design, or bad implementation is zero. Perfection is trapping all major bugs that cause crashes, drop outs or prevent use of the core functions. What you add, as I said, depends on the sophistication of that market. How can you do effective user testing if the users can't get to the good stuff? You have to respect your users, whatever your solution.

Brian S. Reed

June 12th, 2015

Without knowing all the details it's tough to give accurate advice. So in general terms startup projects are three phases - there's usually a time period where ideas and failure / rewrites are acceptable, and then a term of moving through 'we will do it right' and perfection, and then the period of 'we have to get this done now'. All three are important, and for large projects can last about a month each. If a product isn't ready to start deploying in three months, that's a bad sign in my experience, and likely to never ship. If you are in last phase - cut out all unnecessary feature, focus on minimally viable product,  and trust the work in 'do it right phase' will shine through. If the second phase hasn't been done long enough - push back the date. Again, hard to know without more info.