There exists a wide-ranging spectrum of ‘finished’, and I think it starts at Apple Maps (not finished on launch) and ends somewhere at General Magic (too perfected before launch). Where you set your sights will make or break the product or your timelines. I like to think of the different levels of execution within the context of the Kano Model, so before explaining the ‘spectrum of perfection’, let's go to design school and learn about Kano.
A reminder of the Kano Model.
The Kano model was developed in the 1980s by Japanese professor Noriaki Kano as a way of classifying a product’s features. Baseline expectations, performance features and delighters work in tandem to create the product’s experience. Baseline expectations are the qualifier for entry into the market. Let’s consider a simple example, a coffee cup. The baseline expectation is that it holds coffee and doesn’t burn your hand.
Performance features increase the functionality of the product in a one-dimensional way. The more of these, the better the product. Think of it like cores in a CPU, or pockets in a backpack. In our coffee cup example, maybe it's a flap to cover the mouthpiece so the liquid doesn’t spill on the move.
Delighters are features which increase customer satisfaction but where a lack of would not be noticeable. The textbook example is Innocent Drinks' quirky text they hide around their bottles - a fun little reward for observant customers. In our coffee cup, it could be the little love heart your extra friendly barista draws on.
You can apply Kano to the whole user experience with the product and evaluate the wider service, but I tend to keep it low-level and use it as a way to analyse the individual products. It's also worth mentioning that features move down over time. What was once an impressive performance feature or quirky delighter, like the iPhone's camera shutter or multitouch screen is now a basic expectation.
The Unfortunate Reality
Here's the harsh reality though; the further to the top right of the Kano Model you aim, the higher the likelihood your customer satisfaction will be and the higher the likelihood you'll never ship is. This is where perfectionism gets its bad reputation. Great product teams will always advocate for pursuing extra features and niceties but oftentimes the backlog is limitless. Compromise is best here; cull the features that can be rolled out at a later date or in a different product version and stand with conviction next to those which you believe add extreme value to the product. Unless you have hundreds of designers and engineers at hand, you'll want to keep the performance features and delighters to those which you absolutely believe in.
The spectrum of perfection
In pretty much all cases where 'done is better than perfect' is the design ethos, it's only the basic expectations which are met. Let's run through some examples, starting at 'done' and working our way up to 'perfect'.
Done: Instagram Reels
There is only one reason for choosing to aim for 'done' and that is when speed to market is more important than user satisfaction. This is particularly true in software, where you might be able to increase satisfaction after launch. This worked well for Facebook when launching Instagram Reels because all you need to do is focus on the market qualifiers (i.e; Tiktok's features) and copy. The advantage for Facebook is that less time can be spent designing new features and time to launch can be shorter as a result. The disadvantage is that the product is a mediocre clone of the original with few distinguishing features.
When the iPhone launched, 3G was already widely talked about and it was peculiar that the device shipped without 3G support. The app store was years away, as was copying and pasting and location triangulation, features that are now basic expectations of any smartphone. None of that mattered because the performance features and delighters were abundant. Performance features like a wide, multi-touch screen and an accelerometer to rotate your screen were unheard of. The delighters were plenty too; you could pan through your album covers in the iPod app with a swipe, scroll through lists and then let go to watch the momentum slowly fade and even the camera app had a pseudo-shutter animation when it was opened. Delightful!
You should remember if you're a startup, Apple is very good at product design and engineering and they can do a lot more in their project timelines than you are likely able to. Here's where perfectionism in a startup kills...
Perfection: General Magic
General Magic was a consumer electronics company spun out of Apple in 1989. If you don't know the story of their product and gruelling journey to market, I highly recommend watching the documentary. They were effectively building the iPhone ahead of its time. Apple's strong approach to building great customer experiences was rife within the company, so much so that it became a huge problem. Their founding team had a strong vision that was filled out with ambitious projects that would culminate in the Magic Link. They spent so much time building the best product they could imagine that constant delays occurred. Some of the delayed projects were delighters that were overblown, like early emoji-like graphics, that just didn't need to exist. The reality was that the market wasn't ready for their invention. The kicker is that they could've found that out years earlier by shipping something that was just a little bit more scaled back. It wouldn't have made it a success, but they'd have concluded a lot sooner.
Next time someone says 'done is better than perfect', have a think about what they actually mean. Maybe they're advocating for cloning a competitor to get the product to market faster, but chances are they mean somewhere in between 'done' and 'perfect', and that's worth thinking about.
Side note: I don't think Facebook employs 'done is better than perfect' throughout its departments, see the Oculus Quest 2 for reference and try and place it on the spectrum. Also, if you're in a design team, you should consider writing the Kano feature type next to every point on your design specification. I don't have a magical perfect ratio for basic /performance /delighters, but if you're top-heavy on the delighters, you'll know you need to scale back.