Why do Microwave oven UIs fail?

Question

I’m not a professional chef, but I do know my way around the kitchen and its machines. The dishwasher, the fridge, the freezer, the stove and the oven all works as expected in many homes. Could you start the dishwasher, my friends might say, but they never say – heat this in the microwave oven. Probably it’s because they are programmable these days, and have plenty of buttons with arbitrary non-standard, non-conventional icons and no explanatory text to help. Fortunately there is a trend among microwave oven manufacturers to add a panic-button at the lower left of the control panel. It’s the button that starts the microwave oven at max effect and runs for 30 seconds.

But I wonder why that is. All other kitchen machines could be operated fairly well if you give it a little time, but the microwave just doesn’t. There can’t be different designers, but maybe more options where microwave oven UI designers had to invent new symbols for new features and add a note in the manual?

Micah Wittman has even made an effort to start a “Microwave Oven UI Standard Project” to overcome this problem. I’d be happy to join if it is still running. So far this is his suggestion:

Still one wonders, why do Microwave UI fail?

Answer 1

One UI is analogic, the other (the failing one) is digital, that is, the appliance is a microwave-computer mongrel.
Alan Cooper, in his book "The inmates are running the asylum", teaches us that when you mix a computer with anything else, you always get a computer no matter what the anything else was.
Thus, mixing an oven with a computer, the outcome is a computer.
If you don't happen to have the Inmates book, you can read the chapter online opening the Amazon link above and following the Look inside the book link. After scrolling a lot the chapter is available, it's a highly recommended reading, somehow humorous, enlightening.

Then, back to the reasoning that tends to answer the question, IMO there are two main issues:

The first issue is that digital controls are less intuitive than analog controls. For example when you operate your mother's microwave, even for the first time, you know that turning the knobs clockwise you set a higher power level or a longer time. This is not so with the digital UI.

The other issue, an important one, is feature creep. The designers (usually engineers) try to give us more, because it's rather inexpensive, so they can solve us problems that we don't actually have.

Consider a digital UI made up of two button pairs, one pair to control power and the other to control time.
The effect of the buttons appears in a display, changing the power level gauge or the displayed time.
The UI completes with a [start] and a [stop/reset] button.
A digital UI such as this would be perfectly usable, even for everybody's grandmother (given that the labels are printed in a suitable font size).
This UI is usable by anyone because such an UI is aligned with the user's mental model.

Because, ultimately, the usable UI is the one that's aligned with the user's mental model (this concept is explained in another book by Alan Cooper, "About Face").

And an anecdotal story. I work for a software development factory, in locations populated by hundreds of people in their twenties, mostly of the nerd type.
At lunch time they avoid those microwaves of the digital type. Some kids really loathe them. The analog ones have a greater audience.

Answer 2

You made me think about it. My mental model is:

^{download bmml source – Wireframes created with Balsamiq Mockups}

Please remember, it is not even UI wireframe, so forget the button sizes, fit for user expectations and all such UX stuff. Just shared a MM.

Oh, and another thing. No weighting information is intentional.

Answer 3

To answer your question directly, I believe microwave UX is bad because people buy microwaves based on the feature list on the box (and "one-touch cooking of up to 99 different types of food!" sounds like a good thing while you're in an appliances store), and microwave computers can be engineered most cheaply if you ignore good design. Basically: as much as it drives me mental, I'm unlikely to throw away my microwave because its UX is terrible, and I'm not generally able to review the UX of a new microwave before I buy it.

---

More generally, microwave UX is a topic I spent some time discussing in my UX presentation Don't Cook Babies. You can watch the relevant section on YouTube. Tog also spends a little bit of time making a great point about microwave UI efficiency here.

To summarise my complaints with microwave UI:

• The main way I wish to use my microwave is to put some food in it, and then follow the instructions on the food packaging itself to ascertain how long it needs to be in for. That means I generally want to enter a specific amount of time.
• Manufacturers have all decided to put a bunch of pre-defined presets on the control panel instead of numbers, which I guess is great if you confine your cooking to those types of food, but means you're completely lost when you heat something that is either not one of the presets, or is a combination of multiple presets (e.g. reheating a plate with both meat and vegetables).
• Most microwaves now have some kind of quick cook button, but they all tend to behave differently from microwave to microwave, so you can't ever learn how to use it reliably (as shown in the video, mine goes up by 30s increments at first, then spontaneously switches to 1m increments, then 10m increments at some point, and begins cooking immediately after first pressing the button).

To me, there are two main functions I wish to perform on a microwave:

1. Cook food for the time indicated on the packaging. The hardest part about this is that the packet generally says something like "cook for 3 minutes at 1,000W", and while conversion tables exist, I have no idea what power my microwave is running at, so they don't help me at all.
2. Defrost a certain amount of meat before cooking it. This is actually generally pretty well supported by most microwaves (although I still think they give me too many options). On my microwave I have to press "Defrost", then choose "Defrost by weight" (by moving my finger to the OK button), then type the weight in grams (moving my finger again), then press "Start" (moving my finger again).

What I literally never need to do is take a bunch of unpackaged food, e.g. "meat" and cook it for some unknown-to-me, but known-to-my-microwave amount of time (which microwaves seem to be optimising for nowadays).

To me, the ideal microwave UI would work as follows:

• Provide a keypad to type in the time you wish to cook for (in minutes and seconds). If I press "Start" or just stop typing for more than, say, 5 seconds, start cooking on high for that long.
• Put a weight sensor inside the microwave that can measure the weight of the food I put in it. Put a button on the front called "Defrost", that when pressed weighs the food and tells you how much it weighs in grams (e.g. 502g). You could type a different amount over the top of it if you need to (e.g. if the food is on a heavy plate), or you can just press "Start" (or walk away) and it would defrost.

Some things that wouldn't be possible on my microwave:

• Manually specifying a heat level (always either defrost or high).
• One-touch cooking of a variety of crazy things, including (but not limited to) healthy babies.

I think I would be able to preserve the ability to set the time but I don't value that functionality particularly highly myself (though I understand I'm apparently in the minority on that point).

Answer 4

I worked on the UI of the Microwave in the photo.

The problem is in order to sell product the Sales & Marketing are convinced you need features. More features = good. (Increased selling price).

So as an Engineer how do you add Auto cook, auto defrost & selection of other features whilst keeping thing simple. Some Models you also have sell worldwide so no English text is allowed. You have to use symbols.

For basic cooking on the Oven shown, you rotate the dial to the time you want & hit the start button. Or keep hitting the Start+30 until you get the time you want.

If you want more advanced features you can start to use other keys. Belive it or not there are some people out their who use their microwaves for serious cooking & want the extras.

We have tried to keep the basic controls, seperate & simple to use, But in the end the best you get is a control panel like edgarator's.

A couple of buttons people use and lots of them they don't.

Answer 5

How about:

Originally for temperature cooking, I put that a temp probe might be useful, but it would be even better if the microwave could automatically detect the temperature of the food without you needing to hook up sensors to it. Maybe it could somehow measure the thermal radiation of the food as it cooks to get an accurate reading... Also, it would be nice to have an estimated time till the food is completed, so that you can bump up/down the power level if you want your food to be ready at a certain time, or maybe reverse that and tell the microwave "I want my chicken to be 165 degrees by 6:30pm, now make it happen microwave!".

Answer 6

The big problem with the dial approach suggested by Micah Wittman and quoted by you is that the dials can be inaccurate - mainly because they were originally a mechanical device. There are a number of issues that this raises:

1. You wind the dial to 30 seconds (say) but you might not get it exactly right and it could be anything from 27 to 33 seconds (for example).
2. Then as the dial winds down to zero it might actually take more or less time than it was set to.
3. When it hits 0 a mechanical switch has to trip to turn off the oven. This takes time.

As the dial wears through use the potential error can build up, and you end up with food that's either undercooked (hasn't killed any bacteria, has cold spots) or overcooked (impaired flavour, hot spots that can burn).

The digital timers are a response to this, whether they are ones where you enter the times directly or turn a dial to update the time.

This isn't to defend the current designs - they could be better - but it's an explanation of where we are today.

What's required is to translate the physical act of turning a dial into something that sets a digital timer to an exact value and rather than controlling the countdown responds to it. Something like a touch screen that mimics the dial might be a workable solution.

Answer 7

I don't understand what's wrong with the number pad and a power setting. Sure it doesn't cook turkey breast with a button, but neither does the turkey button. Anyone who has used a mid-low range microwave knows those buttons are nearly worthless. Why not satisfy peoples need for marketing with an "AUTO" button, which tries and fails to cook EVERYTHING. Rid the micro of all the extra buttons, give the sales a marketing gimmick, and the customer peace of mind.

Is it unethical? Probably. Any more than multiple buttons that don't work? no.

Answer 8

I agree with the sentiment of this question. I own a combi microwave oven that can also function as a hot-air oven and as a grill, and the operation is not easy at all.

But even if I limit myself to the microwave functions, there are many issues. There are modes for defrosting stuff, but the selection is using some arbitrary number that I have to look up (the food-types corresponding to the numbers are printed on the edge of the oven door...) and I always have to put in the weight I put in. Then, if I just want to operate manually, I actually have a dial available for the time and power, but it is the same dial. I have to use a button to switch if the dial operates the power or the time, and I can't see both at the same time. So, plenty of problems there...

---

My ideal microwave

I think my ideal microwave would support these use cases:

1. Follow the recipe kind of operation, so set a time and power accurately. I don't want to mess with conversion tables. If the device physically can't handle the given power, let it do the adjustment for me, but don't make me do the math. Use separate inputs for time and power, and show both.
2. Heat to a specific temperature. For instance, I have a baby, and I'd like to have the formula heated to 37 degrees C, no matter how cold it was when I put it in or how much is in the bottle.

That's really it. Defrosting is just heating to a set temperature. For quick operation, perhaps a set of pre-set, often used temperatures would be useful. I'd like the oven to do the work as much as possible, including measuring the temperature of whatever I put in as well as measuring the rate and evenness of the heating so it can adjust the power it uses or instruct the user to stir the food if needed. That should avoid needing to input the type of food I put in.

Answer 9

I really like microwaves with an interface like this, I remember when I started looking at these in houses, I got amazed with the wittiness of the designers...

^{download bmml source – Wireframes created with Balsamiq Mockups}

Need 2 minutes -> press +1 Minute, +1 Minute Start or Start 4 times...

I've never come across something like, "Microwave for 2:48" so pretty much, no one really ask you for fractions, it's either, 1 minute, 2 minutes, 3 minutes, 5 minutes or 10 minutes...

The logarithmic approach works best for me, and the quick start button makes wonders...

Answer 10

I am astounded no one has mentioned the most useful button (and the only one I use besides opening the door) on the microwave, "add one minute and start" button, everything else is supplementary.

If a microwave UI has this button it gets a pass, if it has a button to read my mind and also analyse what substance has been loaded into it to calculate the time and power to run at but doesn't have this button, then it's a bad UI.

Answer 11

Top 4 reasons why microwave UIs fail:

1.) They are not designed to be USABLE, they are designed to be MARKETABLE based on comparison of feature lists.

2.) The people who design products like this A.) almost NEVER watch real users try to use the product in the environment / scenarios the product will be used in, and B.) on the rare occasions UX observations ARE part of the design process programmers and designers frequently have no empathy for users and tend to "explain away" the user's struggles.

3.) Products with digital controls and UIs are "computers" and run on software. Software is designed for the convenience of programmers, not for the convenience of users. (Read "The Inmates are Running the Asylum" by Alan Cooper. It's still the rule rather than the exception).

4.) Good UX is difficult and expensive, ESPECIALLY for embedded system designers. Who on the team of microwave designers is an expert in human cognitive processes? In my experience it's highly unlikely that even ONE person on such a team has ANY training, interest or experience in human cognition.

Read "Plans and Situated Actions" by Lucy Suchman. It's a hard read, but she nails the disconnect between programmers and UX.

<><>

I found this article while looking for a decent quality microwave with simple controls. I've searched for 40 minutes. I think no such thing exists. It should be so easy:

• Two "soft" dials: • Time, with a digital display to indicate exactly how much time is set/remaining. • Power, with a digital display to indicate the % power in 10% increments. • Variable ratio: turn slowly for small, faster for large increments
• One lighted button: hit Green to go, Red to stop

If the product manager wants to add a bunch of garbage buttons, presets, programs and sensor controls go ahead if you must. But if the simple controls above are included (and visibly grouped on the panel) the microwave will be easy to use.

Answer 12

Personally, I am a fan of the two-dial variant, mainly because most microwaves I had were still stupid machines where a complex UI would just fake a complex device, that they are not. So what about newer and upcoming microwaves that are really "high-tech"? Would they demand a complete UI makeover?

I recently saw this microwave at an electronic store (which represents my personal worst case in microwave UIs). Apart from the membrane buttons, which by the way had no tactile feedback at all, we are presented a grid of seemingly unarranged control elements. Above all, the user is deprived of a conventional time and power selector. Instead he gets three different "Auto" setting and the possibility to enter the weight (?) of the meal.

While the microwaving system itself might be cleverly designed and the auto features are working as they should, the poor UI makes it very hard to trust the backend. And when you have no control of the cooking parameters, you HAVE to have faith in the automatic.

Maybe the two-knob version is already obsolete, but modern devcies that could use more suitable UIs should still communicate the feeling of the simple machines that microwaves once were.

Answer 13

The real problems are:

1. All the automated modes that don't work. Ovens don't have them and people don't get confused.

2. The microwaves with analog wheels that have more than one function or some non-standard function. - The wheel, if any, should control either the time or the power, not both and not the mode/function.

3. Small or hidden start/stop buttons - these should be the largest buttons and next to each other, either at the top, or better yet - at the bottom (since they are already there in many usable models).

A microwave should have a clean, simple UI:

1. Optional - For combined devices (e.g. MW+Grill+Oven) - Select mode (if not default = Microwave).
2. Optional - Select power (if not default = MAX).
3. Select duration.
4. Start

Also, microwaves should have a pause/cancel button, enabling the user to stop the cooking and then continue with the start button or the cancel the cooking.

An additional useful feature could be increase/reduce time (without waiting for operation to end).

The duration has to support fine resolution for short durations e.g. 10sec, 15sec, 20sec and etc. and long durations. This is hard to implement with an analog dial, especially since the user needs accurate feedback.

If the duration is not set with a dial, it should have a numberpad (assuming physical buttons are used), not +/- buttons which can take a lot of presses to get right.

The mode, power and duration should not use the same controls.

The controls should have clear labels (e.g. no icons that no one understands and no text that is poorly translated).

If the microwave has different modes, then the power should have separate controls per mode (e.g. 250W, 500W, 750W, 1000W for MW vs 100-250 degrees for oven vs upper/lower/both for grill).

Answer 14

The reason why microwave oven are so complicated is due to the competitive features that each manufacturer puts out to out-do one another. More features means more buttons and options.

It doesn't matter how useful or user-friendly that feature is, as long as it is unique and adds an additional tick on the box.

I doubt if anyone ever used the "Cook frozen chicken" or "Boil corn soup" feature (or any other advanced options like that) with their microwave ovens.

In my opinion, a microwave that offers basic Timer + Heat Level setting is the best. Anything more is just useless and makes the interface harder to use.