Complexity causes 50% of product returns Jason 06 Mar 2006

36 comments Latest by Sebhelyesfarku

Complexity is bad for business.

Half of all malfunctioning products returned to stores by consumers are in full working order, but customers can’t figure out how to operate the devices, a scientist said on Monday.
She also gave new products to a group of managers from consumer electronics company Philips (PHG.AS), asking them to use them over the weekend. The managers returned frustrated because they could not get the devices to work properly.

And my favorite part of the article:

Most of the flaws found their origin in the first phase of the design process: product definition.

That’s the money quote. Defining the product by defining the vision. Knowing where the boundaries are is as important as anything. Not what should you do, but what shouldn’t you do. Know your product’s limits. This is much easier with hardware than it is with software, but, as this article shows, it’s plenty hard even with hardware.

What does your product stand for? What does it mean? Why does it exist? And how are you going to explain this to customers? Key questions that need answers up front, not later on.

36 comments so far (Jump to latest)

Eloy Anzola 06 Mar 06

A great article, that clearly makes the point.

I am always amazed by how little some companies understand their own products.

You guys at 37 Signals has been fantastic in explaining *how* to really listen (or not) to costumers, and how essential features make themselves known.

Keep it up & thanks.

benny 06 Mar 06

“Complexity is bad for business.”

Not to be snarky, but is complexity ever _good_ for anything? I hope nobody would be surprised by this news.

Rabbit 06 Mar 06

Heh… I think people ARE surprised by this news.

I just read the article outloud to my co-worker (thank god it was short).

Unfortunately, he’s an apathetic brick wall when it comes to new ideas and change in general, so all he said was “okay.” (Man I’d like to knock him out just once…)

ANYWAY! What was I saying? Oh yeah, it IS new, just maybe not to us posting on this blog.

Snarky. I dunno what it means but I like it. I should name my first child Snarky. Bwa ha ha ha! :)

Oooh… insight. Benny said “is complexity ever good for anything?”

Note the word “complexity” versus “features” or “cool new features” or “features that will make you look younger and get hot girls.”

All those phrases are masks for what is essentially complexity. Interesting. The lies and double-speak we live with.

We parade around in a country we think is a democracy when really… oh, sorry. :)

niblettes 06 Mar 06

I suspect you think it�s a “great article” because it simply reaffirms your already held beliefs. That unfortunately doesn�t make a great article.

Indeed the entire thing is predicated on the opinions of “a scientist;” a scientist we learn in paragraph 2 is actually just a grad student (I assume this based on the thesis reference).

I remember reading a thesis paper from a design school graduate. Her logic went something like this: designers can draw, drawing is important in visualizing, and visualizing the future is important to leading a corporation, therefore designers should be CEOs. This level of reasoning was pretty common in the dozens of graduate level design thesis papers I read. So how much weight does a grad student’s thesis paper deserve?

Furthermore, what degree is she working towards? Design? Anthropology? Engineering? Physics? And what qualifies her to be called a “scientist” as a grad school student? Where can we find her work to see for ourselves? What was her thesis? Were these findings central or peripheral to ither thesis work?

Come on, use some critical thinking here. The article could be right, it could be wrong�there is no way of knowing based on what’s there. So aside from the preaching-to-the-choir warm fuzzy it provides, it is pretty much worthless.

and to Benny, complexity does seem to be pretty good for ecosystems—simplifying the complex diversity of interrelated elements kills ecosystems. So yeah, sometimes it is good, sometimes it is necessary.

mrG 06 Mar 06

What ‘benny’ may not realize is how “complexity” means different things to different people, and it is my bet that the designers of all those devices given to Philips managers had been created by people who truly believed they were deploying the minimum complexity for the necessary feature set.

Take, for example, this browser you are using right now. Is it complicated or simple? Have you ever added a plugin? Has your mother modified her browser? In my experience, the average non-IT-worker almost never customizes anything about their desktop, their applications or even the computer itself. It’s all just the way it came from the shop. At the early Sympatico we had to deal with fantastic traffic to the ISP homepage because the vast majority of our 1 million subscribers did not care to change the “default homepage”. Oh, they knew how, alright, but it was too much bother, far easier to just tap the bookmark key as the browser loads, leaving us with this massive server being used to pump out pages nearly no one actually read!

So is “set your homepage” a feature? Or unnecessary complexity?

Darrel 06 Mar 06

�Complexity is bad for business.�

You’d think that. However, much of the software industry, it seems, is counter to most any other industry. Possibly because those that purchase software often aren’t the ones that have to use it. Therefore, complexity is sold as ‘extensibile/feature rich/scalable’ to the suits and then it’s dumped upon the person that actually has to use the product.

So, my advice:

complexity is great in software if your goal is to sell product to the suits.

complexity is awful in software if your goal is to seel product to the actual users.

“but is complexity ever _good_ for anything”


Rabbit 06 Mar 06

Diverse doesn’t necessarily mean complex.

When you break down an ecosystem I think you get some pretty simple, similar patterns. You’re born, you eat, you mate (hopefully) and you die. Given that general pattern we get a very complex system, yeah.

So the question is at what field of depth are you looking?

Didn’t they post about this a while ago? Emergence? (Or was that it the book? I think it was in the book…)

Give a flock of birds 4 or 5 simple rules and they can create complex behaviors from it.

I think a great example is Backpack. In fact I have a dirty little secret about Backpack - I use it just like a Writeboard without the versioning.

For me it’s basically an online notepad. I don’t use the lists or the image uploading or any of the other tabs in it. I do use the actual Writeboards when I’m writing web site copy. I also use the SMS reminders - those are fun. And yes, I have a paid subscription to BP.

Anyway, what I’m saying is, Backpack has just enough rules, just enough complexity for me to use it how *I* want to use it. I’m almost certain the Signals would have never thought their product would be used the way I am.

Gary R Boodhoo 06 Mar 06

The words “complex” and “difficult” aren’t synonmymous. Neither are “simple” and “easy”. Simplicity encapsulates complexity, whereas difficulty is a sure sign of confusion, lack of focus, etc…

Complexity is not only necessary, but anything of substance will necessarily be complex - collaboration, software engineering, physics, chemistry, biology, music, art, etc…

The challenge as a designer or developer is to enable an intuitive understanding of complex systems. You know a tool works well when it provides insight into what was always there in front of you and allows you to “think different”

I wouldn’t say complexity is bad for business. Business is naturally complex. I would prefer to state that lack of focus, lack of direction, lack of vision and lack of empathy are bad for business.

street 06 Mar 06

50% ? I just bought a magic statistics determiner accessory for my Ipod but for the life of me it always reads 50%. Maybe its because it is predicting that I’m going to return it.

benny 06 Mar 06


Agreed, life is complex… but it just IS… that doesn’t make it the better alternative to some other thing… the elusive ‘simple life’ has plenty of appeal =)

Regarding ecosystems and complexity.. thats very interesting - simplifying them and thusly killing them. Very cool. I do not know ecosystems.

Dave Hoff 06 Mar 06

Nice to see the quality posts back on svn.

Kinda missing those weird pictures though.

a scientist 06 Mar 06

@niblettes: And what qualifies her to be called a �scientist� as a grad school student?

i’m a scientist absolutely with no letters. that’s probably what makes me a better scientist, the freedom from historical perspective. what qualifies anyone to be called a scientist? is there a board certification? is there some sort of badge?

i think if you’re truly a scientist it doesn’t matter if anyone else acknowledges the designation. real scientists don’t care what you think. they’re too busy doing stuff.

niblettes 06 Mar 06

Again, I’m not saying the conclusions are either right or wrong since I have no way of knowing, because all we have to go on is your professional credibility (in other words, because you said so). And if that is all we as readers are given, then we as readers are perfectly justified in questioning your professional credibility.

And from my experience with grad school thesis work (if this is indeed grad school thesis work) that isn�t much credibility.

No according to your logic I’m a scientist too! We’re all scientists, and artists and designers and fire trucks and…

What troubles me about your response is that you call yourself a �scientist� while at the same time suggesting that there are no actual qualifications for that title. In other words, �scientist� is utterly meaningless. So why call yourself one? Why not use the more truthful and appropriately contextualized (I think) term �subject-X grad student�?

Unless the term �scientist� is being used to artificially inflate your credibility? Why does your credibility need to be artificially inflated if you methods are sound, your data are good, and your conclusions well founded?

If you are indeed the �scientist� in the article, rather than whine about my comments and fling red herrings, why not share links to your work so that we can all understand your methods, see the data and decide for ourselves how much credence to give your conclusions? If they are well founded we will all thank you dearly.

Well, if since I�m a scientist too I say the figures are more like 37.5% (I aim for higher precision). Who is right?

Ian Waring 06 Mar 06

Complexity per se isn’t a problem, it’s more about how much is exposed to the products end user. From a use point of view, an Engine Control Unit in a car is pretty complex, but for most people it just gets on with the job.

The thing that’s bugging me at the moment is that you have another thread on the SvN blog where someone who is visually impaired is getting no help buying a copy of your (excellent) book, because his screen reader can’t make simple sense of your on-screen order form. Is there no-one at 37S that can help him?

Ian W.

gwg 06 Mar 06

A scientist is someone who uses scientific principals and methods to learn from others and draw conclusions. Additionally, scientists allow for and seek out peer review and scrutiny of their work to better ensure accuracy.

Let’s not start talking about science the same way we talk about art.

ML 06 Mar 06

Is there no-one at 37S that can help him?

Ian, we don’t provide tech support for our products on SvN. If anyone is having a problem, please email gettingreal at 37signals dot com. Thanks.

Steven Romej 06 Mar 06

When I saw this story I immediately thought about svn and felt certain you’d have a post up about it.

a scientist 06 Mar 06

@niblettes: “No according to your logic I�m a scientist too! We�re all scientists, and artists and designers and fire trucks and�”

i knew you’d catch on. congratulations, you’ve just earned a phd.

and i never said i was the scientist in the article, i’m just defending the craft.

@gwg: “A scientist is someone who uses scientific principals and methods to learn from others and draw conclusions. Additionally, scientists allow for and seek out peer review and scrutiny of their work to better ensure accuracy.”

precisely. using scientific methods (aka “being a scientist”) doesn’t require a degree. well, i’ve got to stop typing and start doing some stuff now, cheers.

niblettes 06 Mar 06

I just got a PhD? Wow that was a hell of lot easier than my master’s! I’ll bet everyone can earn one these days.

And still no links or pointers to any of her data opening it up for public scrutiny. So much for the scientific method. I suppose this is the new faith-based science I’ve heard so much about.

I notice this little article has already hit a number of blogs—almost entirely without criticism or question (kinda how Jakob got to be a guru).

Imagine quoting this 50% number in a sales meeting with a potential client. Client looks back and says “how do you know its that high?” How far would you get by answering “cause a press release quotes a dutch grad student’s thesis that says so.” ??

While poster “a scientist” was probably just a troll s/he illustrates the need we have to be a whole lot more critical of the “information” that gets passed around the net and of the people who pass it. This is especially true when we already believe in what the information supposedly tells us.

What we got isn’t science folks—it’s religion.

nate 06 Mar 06

First of all, complexity is an inherently relative concept.
Anything human made is fairly simple. If you want complex, look at DNA.

As far as complexity being bad, absolutely. Life is all about battling complexity. It’s like my dad always says: true genuis is finding a simple solution to a complex problem.

Saul Weiner 06 Mar 06

I think the following holds true:
Iinnovation happens you don’t listen to your customer and focus on your vision.
Sometimes you get it right and sometimes you get it wrong, but it’s rare that a customer has the vision for your product that isn’t self serving.

Mukul Gupta 06 Mar 06

Last month’s harvard business review brought an interestng point forward. Customers buying decisions are influenced by number of features or “expected utility” or a product but customer’s lifetime value is dependent on how easy they find to use the product or its “actual utility”.

So a business faces a dilemma, they must add a lot of features in order to make the first sale but they also need to seek a balance betwen features vs usability in order retain a customer. It quoted philips electronics recent drive for “Sense and Simplicity”. A good example of unnecesary complexity was the iDrive of BMW 7 Series which can manage 700 functions and thus always required the driver to keep the manual within reach. It quoted that sales dropped 10% between 2004 and 2005.

The point remains that to most new ISVs the first sale is most crucial and hence the common wisdom is to continue to add new features.

Mark 06 Mar 06

Meanwhile, simplicity is good for business. Those iPods are selling like hotcakes.

Absconditus 07 Mar 06

It should be noted that BMW also redesigned the body of the 7 series and the backend looks horrible now.

some joe 07 Mar 06

the article got quoted in most blogs not because of the numbers it presents (or the lack, thereof), but because of the statement that it makes. it states an obvious point that though as simple as it sounds, is always forgotten or if not, hard to achieve. it does not tell us of something new, but instead, encourages us to continue pursuing a line of thinking that’s already been presented.

your comments are hardly constructive anymore, and have just turned this comment list into a debate on syntax and semantics.

some joe 07 Mar 06

… and in this blog’s case, it is being discussed in the context of web applications design and dev’t.

Henk Kleynhans 07 Mar 06

This is so true!

Many years ago I worked in tech support at a large computer retailer. My favorite was an elderly couple who brought back a ‘faulty’ scanner. It couldn’t scan their 50 years of negatives as stated on the box. (There was a big button on the software that was supposed to be “intuitve”)

Today a customer called because she couldn’t connect to one of our hotspots. I sent out a technician who determined she hadn’t pushed the Wi-Fi device button on her laptop.

For years we called these problems “user related” as in “99% of computer problems are user related”.

If only more companies would realise these problems are “engineer related” or “developer related” or maybe even “drug related”?

Jake 07 Mar 06

niblettes: wow, i’ve never seen anyone get so hung up on credentials, titles, etc as you… unbelievable.

you do realize that 70% of all statistics are made up right?

Lynn Fredricks 08 Mar 06

I agree with Mukul on expected utility - or perceived utility. Simplicity as a goal makes a nice tag line, but its an oversimplification ;-)

A great many software products sell because, even though they are extremely complex, the perceived utility is high for the investment in the actual work. Same idea here applies to both the iPod and developing procedural textures in Alias Maya. More smiles per second is not the same as simplicity.

Ive gone into more detail on my own blog - Technology Tribe, including what I call the Five Minutes to Joy on demo software.

Alex Benenson 08 Mar 06

The authors of the article say that “definition” at the beginning of the design process is the primary cause of problems.

Therefore I think it’s safe to translate “definition” to “requirements”.

Off the top of my head I can think of several ways that bad requirements can lead to bad products:

1. The product has too many features (complexity)

2. The product uses immature technology (bleeding edge)

3. The product has the wrong audience (marketing)

4. The product was designed in the requirements phase instead of afterwards (marketers/managers trying to be designers)

flipdoubt 09 Mar 06

English major turned coder here. In a “17th Century Metaphysical Poets” class in college, we made a big deal about the difference between complex and complicated. Complex connotes layers of something — meaning in poetry, indirection or abstraction or security in computers — but is not necessarily the opposite of simple. A haiku can convey a complex message using simple imagery and verse. You may not notice how complex something is, i.e. an iPod. Complicated connotes something more noticeable and, thus, more negative, such as features or meaning all jumbled up together. Like a non-iPod music player.

The magic is in providing features that address complex issues without making the solution too complicated. I guess “too” is a key word.

jocabola 09 Mar 06

always keep it simple…

Sebhelyesfarku 10 Mar 06

most people are simpleton so obviously they need something similar