Fog Creek Software
Discussion Board

Bram's Law: What Customers Want

I copied this from his blog (he's the BitTorrent guy):

"The things which will make people love your software, by rapidly plummeting order of importance, are:

1. ease of use
2. stability
3. performance
4. features

The order of priority many people use when writing software, and, unfortunately, what users generally say they want when asked, are:

1. features
2. performance
3. stability
4. ease of use

This is a siginificant discrepancy."

Now this is interesting, any comments?

Green Pajamas
Saturday, July 10, 2004

Focus on performance and stability, and you will not be too off the mark?

Saturday, July 10, 2004

I think a related concept is the "prettiness" of a UI.  Features will get someone interested, but frankly, it's the UI that gets the signature on the check.  I'm constantly surprised at the number of software company web sites that don't have screenshots.  If the application has a GUI, when all is said and done, it's going to make or break the sale.

Saturday, July 10, 2004

Stability is simply a must-have.  If your app hasn't got that at least 99% right, you shouldn't be shipping it.  What good is an app that crashes and eats your data all the time?

Performance is slightly more arguable, as it depends on the context.  There are probably things your customer wants more than improving the speed of that algorithm from 500 records/second to 1000 r/s.  But if the whole UI in general is sluggish, they're definately gonna make that a priority.

Ease of use is definately important, but also fairly subjective.  What one user loves another will inevitably hate.  Also, unless your product, or at least type of product, has been on the market a while, it's really hard to know how people will interact with it.  People seem to have lost the stigma against v1.0 apps these days, and therefore have pretty high expectations.  But as long as you actively solicit feedback and act on it, you've at least got some wiggle room.

I think features do matter more than a pretty UI -- but only the major ones.  If you're building a CRM app, and it doesn't provide accounting integration or reporting, you're probably not going to win based on your slick UI.  But, on the other hand, you can probably wait on that nifty PDA hotsync feature. 

Overall, if your UI doesn't look professional, people will generally assume that your app isn't either.  But "professional" and "snazzy" are two different things.

Saturday, July 10, 2004

Ease of use will get me to try it in the first place. Any software that I load up that I can't figure out - WITHOUT READING THE F'N MANUAL - in 5 minutes gets dropped.

Stability will get me to keep using the software - when running 2 similar apps, if one crashes it's out.

Performance will get me to recommend it to my friends becuse I won't be afraid of them abusing the software & it crashin on them.

Features... Yeah, they're nice too, and I believe this is one of the reasons MS Office has such high market penetration - there's a feeling that if you buy it for your company, they're never going to run in to a situation where Word or Excel won't do this or that that another program can do. And if there is, then odds are that other program can only do that and nothing else.
Saturday, July 10, 2004

> Ease of use will get me to try it in the first place. Any
> software that I load up that I can't figure out - WITHOUT
> READING THE F'N MANUAL - in 5 minutes gets dropped.

While I agree that most consumer-targeted software should not require in-depth instruction, I have to say that there is an appropriate time to RTFM.  Let's see if I can make an analogy relating cars and software without getting thrashed for it :)

When you buy a new car, yes, you expect to be able to drive it off the lot without having to spend 3 hours reading the manual.  However, if it's a new brand of car for you, and you want to know how to work the CD changer, set the clock, etc you're probably going to have to at least thumb through the quick help guide, right?  Similarly, while you can probably figure out how to change the oil and fill the gas tank, you might wanna read up on the specs before attempting to install your own stereo system.

Using new software should be the same.  You should be able to figure out the most essential functions without any help.  But depending on what kind of program you're using, how complex your task is, and your previous level of experience, it isn't unreasonable to expect that you may need to refer to the manual or online help every now and then.

Saturday, July 10, 2004

Oh, I agree, but if the basic functionality is beyond me, then there's a problem, and I'm very much tempted to "blame" the designer.

I'm just talking about frustration levels in general when using the software.
Saturday, July 10, 2004

Okay, what if you will be more efficient after reading the manual.

Try for example `grep` command  against easy of use search-and-replace tool.

You can easily use the second but the first is more reusable (you can use it in many contexts and automate your work).

Max Belugin (
Saturday, July 10, 2004

Obviously it depends on what my goals are. If I will be using Search & Replace often & I need a powerful tool because it's not a straight search & replace, then sure GREP would be the way to go.

But if I have 30 documents in front of me & a simple search & replace will do, then I'll do that.

In each case, the right tool is needed for the right job, and there are times when GREP - even if I had to learn it - would be the simpler way to go.

I am a big fan of regular expression, though I don't use it often enough to really remember it well, so I have to make use of reference materials whenever I do.

I guess one could posit a certain ratio of:

difficulty : goal

To use the car example, I'd be willing to learn to drive because it's a lifelong skill & will serve me well in a great many situations. I wouldn't learn to drive because I ran out of milk & needed to go to the store. Unless, of course, that was the last straw because it was the 1,000th time I needed something from the store and couldn't get it.
Saturday, July 10, 2004

When considering features a good way to evaluate them is with Kano Analysis:

Some features are must haves, the product is incomplete without them. Some are nice but not required, some are of no interest, and some are delightful surprises. Unfortunately too many features added fall into the "no interest" category.

Anony Coward
Saturday, July 10, 2004

Is setting a clock such a difficult and complex operation compared to driving a car that it requires special instruction?

Is using a CD player?

Frankly, if you have to read the manual to use a CD player, it seems obvious to me that that CD player has no business being installed in a car where a few moments off the road and several people are dead or crippled for life.

Automobile extras such as clocks and radios must be designed to be so intuitive and obvious that it should be illegal to even ship them with a manual. Anything more complicated is a serious threat to public health and safety.

Dennis Atkins
Saturday, July 10, 2004

Oh come on, you'll need a manual for SOMETHING. Even if it's only programming the radio's presets, or setting up the CD player's programmed track playback. You know, the kind of things you'd just never do when actually driving the car.

(And anyway, lack of manual doesn't mean it's safe to do when driving. I didn't need a manual to work out how to use my ipod, but it's pretty dangerous to use when you're on the road. Fortunately, there was noone else around at the time.)

Saturday, July 10, 2004

The dark side of "competition."

Tayssir John Gabbour
Saturday, July 10, 2004

It's not black or white.

I think the key is to figure out the minimum necessary capability in each area: useability, features, speed, etc.

I think that you should be able to operate the "top level" (most fundamental) features without a manual.

I.e., I should be able to enter a new contact in the CRM software, search for a contact, FIND the query form, etc.

BTW, GREP is a good example of (IMHO) overkill:
A lot of software these days seems to have either simple string search or GREP. But, GREP is useful to MAYBE 1% of computer users.

What would be INCREDIBLY useful to 90% of user is a simple BOOLEAN SEARCH. Just look how useful Google is. People do Boolean Searches WITHOUT EVEN KNOWING they're using it.

That's power: making sophisticated features easy enough for grandpa to use.

Mr. Analogy
Saturday, July 10, 2004

Re: Car Manuals

I hate it when I can't figure out how to set the clock in the car, but I don't mind not knowing so much because I don't have to do it very often. The strangest clock radio I ever had to set was actually set with those little indented buttons you need a pen to push. Who wants to get pen marks on their brand new car? Sheesh.

Setting radio presets should be pretty intuitive. Remember when the world switched from PULL to PUSH & HOLD to set their presets? (and no, I'm not dating myself, I didn't actually drive at the time when PULL was the norm).

It's intuitive, though, because every other radio in the world works that way. What about radios with 5 buttons & 10 possible presets though. how do you figure that one out? Or 5 buttons & 9 presets? Yes 5 buttons & 9 presets. There's a raised dot connecting 2 buttons and but pushing it - i.e. 2 buttons at once - you get another preset.

These are things you might need to look in the manual for, but you really shouldn't have to. You should be able to figure out the radio by glancing at it for half a second, because if you do a lot of driving, you WILL want to change the presets mid-drive as radio stations fade in & out.

Here's a good example. I have a lot of data stored in an Excel spreadsheet right now. When I try to insert a function, I got to Insert -> Function, and next to it is a little *f symbol. I was using one function frequently enough that I wished it was on the toolbar, so I tried to figure out how to add buttons to the toolbar. But I stopped mid search because I saw a little *f symbol on the toolbar. Now my function was 2 clicks away - the *f button & since my last used function was highlighted already, OK.

There's an advanced feature that I did not need to whip out the manual for. Maybe I could've added that function to the toolbar, but in the toss up between 2 clicks every time & (unknown quantity) minutes of searching to make it 1 click, 2 clicks every time won. Maybe if I was using 2 functions regularly (increasing my number of clicks by 1, not to mention having to THINK each time I clicked to see which was selected), it would've been worth it, but the path of least resistance won out.

Re: Kano Analysis Article

Interesting article, but it's flawed. I don't know much about six sigma, but this article completely lacks the "time / features / cost" aspect of doing business. Adding features can give you a competitive advantage, but so can having a product to market before anyone else. So can having a *stable* product to market, even if it has fewer features, when your competitor's product is feature-filled & buggy.

Re: It's not black or white.

Exactly. Give me a tool that can do search & replace that any moron could figure out, boolean that someone with half a brain can figure out, and regular expression if it's needed.

Case in point: Sometime I need to rename dozens of files. I found 2 tools for Windows that does this. One has lots of advanced features, but I couldn't figure it out, I had to configure every little thing I wanted it to do. Another opened up the titles of files in a folder as if it was notepad. It also has a macro recorder to record keystrokes if you need it.

Guess which one I ended up using?

BTW, the two programs in question were The Rename

and Oscar's File Renamer (the one I used)

It should be noted that the complex one had an extensive help section, but Oscar's had a single HTML page that prints to just over 2 pages.
Sunday, July 11, 2004

Re: The dark side of "competition."

I hated MS Money. I liked an early copy of it I had, but I upgraded because I think I wanted it to somehow help me with my taxes & I was afraid it would be out of date.

My biggest gripe was - I couldn't balance my checkbook with it. The checkbook feature only showed you your current balance. You couldn't get a historical balance from, say 2 or 3 transactions ago, so I couldn't reconcile it with the monthly statements I got, and half the time with the internet either.

It's not exactly rocket science, but I have yet to see a company get this right. And now MS Money is the focus of the damned Inductive UI. "Hi, let's make everything a wizard and treat you like it's your first time using the product every time you use it." AKA "Let's pretend everything is the web." UGH.
Sunday, July 11, 2004

"My biggest gripe was - I couldn't balance my checkbook with it"

Perfect example of Kano Analysis. Checkbook reconcilliation is a Must Have feature, you don't consider the application worth buying without it no matter how stable, first to market, or whatever.

OTOH, forcing you to use wizards is a negative feature; again it doesn't matter how stable or fast to market they made it, the product would be better without it.

Anony Coward
Sunday, July 11, 2004

>et's make everything a wizard and treat you
>like it's your first time using the product every
>time you use it

Got a copy of Microsoft Money. I could not figure out how to use it. All I wanted was to help keep the checkbook balanced (now I just use a spreadsheet) and keep track of household budget (like how much at the supermarket is really food, how much vitamins etc.)

finally just gave up

dot for this one
Sunday, July 11, 2004

I know. What you want to do really isn't that complex. The software should be like MS Word - open it up & start typing. Lots more under the hood, but only if you go looking for it.
Sunday, July 11, 2004

When I look for software to download, I always, always, look at the screenshots.  If it has a pretty GUI, as someone mentioned early, then I will check the features page to see if it does what I want, if I can't tell from the screenshot itself.

I'm an expert user and employed as a QA tester at a software company, but if the program looks ugly and hard to use, I figure it will take too long to learn and look at your competition to see if they have a better offering.

Reply to "Bram's Law: What Customers Want "
Sunday, July 11, 2004

Screen shots are the first thing I look for when downloading software & if it doesn't have one, depending on my need, I may or may not decide to download it. If I can't get a sense of how to use it from the screenshots, I won't use it. But if there's nothing else in the category, or everything in the category is equally miserable, I'll do it.
Sunday, July 11, 2004

" If I can't get a sense of how to use it from the screenshots, I won't use it"

Well put. To *me*, the screen shot indicates, at a glance, general useability. I'm not looking for PRETTY, I'm looking for functional. I.e., it could be monochrome and flat, but if they manage to squeeze the features in that I want, it's obvious.

It's like a newpaper: you can have a well-designed, easy-to-follow newspaper that features 200 year old times new roman text all in monochrome. It's the LAYOUT that impresses me.

Mr. Analogy
Sunday, July 11, 2004

I think Linux and Windows NT based systems are still

1. features
2. performance
3. stability
4. ease of use

there's a lot of room for improvements in terms of usability.

Li-fan Chen
Sunday, July 11, 2004

The trouble is what people say they want isn't always what they will in fact buy and more to the point the poor sod using the damn thing isn't always the one who bought it. 

Rational spending decisions, well thought out and made with an eye on business objectives are rarer than we like to think.    Hence the number of products (not just software) sold on the basis of features used once and then forgotten about.

a cynic writes...
Monday, July 12, 2004

I agree that car radios *should* be more intuitive to use.  But sadly, the norm across manufacturers seems to be push & hold these two buttons, push that one, push the other one, do a dance to the sun gods, and viola! your clock is set.  Of course, that's because they are trying to squeeze a lot of stuff in a limited amount of real estate, and they figure you're only going to set your clock twice a year for the most part.

In terms of CRM software, you definately should expect to be able to add, search for, and update contacts without resorting to the help system.  And maybe the program has a set of easy canned queries for you to use as well.  But if you need some super-advanced query, would you expect to be able to use the query builder without needing help?  Maybe if you're already familiar with buidling queries for an RDBMS.  But for a normal user, they're gonna have to RTFM.

Monday, July 12, 2004

"and they figure you're only going to set your clock twice a year for the most part"

If you're doing it twice a year then either it's really simple, or twice a year you'll be looking for the manual. If you're doing it every day, then after a week you won't care where the manual is, you'll just do it.

This applies to pretty much everything.

Incidentally, most people simply will not even bother trying to use an SQL (or similar) interface to anything, no matter how well written the manual is, how pretty the form looks, or anything else. They'll either write english requirements and give that to the expert, or they'll buy the system that has an understandable interface (even if it happens to make an SQL query under the covers).

They will not read the fine manual, no matter what you do. You can't make it simple and clear enough, even Mr Gates doesn't have enough money to pay them to read the manual and actually do it, and no amount of swearing and cursing will change this.

Either create an easy to use system, or acknowledge that most people simply will not ever use the advanced capabilities. (This may be acceptable, as long as your competition doesn't go for the first option, or messes up in some other area.)

making things easy is hard
Tuesday, July 13, 2004

*  Recent Topics

*  Fog Creek Home