Fog Creek Software
Discussion Board




"80/20" Rule

In an article[1] at the New York Times (that was mentioned at Slashdot) it was written that Yahoo! is trying to regain its stature lost to Google. Among other things, they're going to cut down on the graphical advertising and focus on returning search results more in tune with their web services. As you are probably aware, Google is very spartan; almost no graphics, their ads are text-based and their search results are consistently rated the most relevant. That, along with a number of beta technologies from their labs make them a strong presense indeed.

In fact, Joel once said[2]:

"A lot of software developers are seduced by the old '80/20' rule. It seems to make a lot of sense: 80% of the people use 20% of the features. So you convince yourself that you only need to implement 20% of the features, and you can still sell 80% as many copies. Unfortunately, it's never the same 20%."

Here's the thing: I believe Google figured out the "20%"! My God, can you think of anyone else who acheived this? This is the Holy Grail of programming; it's no wonder Google is #1.

[1] http://www.nytimes.com/2003/04/07/technology/07YAHO.html?ex=1050292800&en=821ae8a3ad2b7af3&ei=5062&partner=GOOGLE
[2] http://www.joelonsoftware.com/articles/fog0000000020.html
[3] This was originally posted at my site but I wanted to see what you guys think.

Chi Lambda
Sunday, April 13, 2003

Actually, I think Google figured out the 100%, people go to search engines to search for something, not to look at ads.

Andres
Sunday, April 13, 2003

that's a pretty thing to figure out... just do a survey of just about anybody who uses search engines.

i wonder how much advertising revenue they do get though. =)

Wei
Sunday, April 13, 2003

I once did some research for a client on "preferred placement".  Costs are either $5,000 or $10,000 per month for Premium Sponsorship and/or AdWords.

http://www.google.com/ads/overview.html

Jeff

Jeff MacDonald
Sunday, April 13, 2003

They offer "five times the industry standard for click through rates".

Five times almost zero isn't a great deal.

Stephen Jones
Sunday, April 13, 2003

I think Google has figured out something simple:  "people don't like being bullshitted".  It truly is as simple as that. 

When I search google, I get:
1. The most unbiased search engine available.
2. I get very useful results.
3. I don't get flashing banners, annoying popups, etc.
4. I can view power point, PDF, Word, etc. as HTML if I don't feel like incurring the app startup time to preview the link.

The other search engines would be (almost) as attractive as Google if they merely cut out the BS.

Software has a huge "Bullshit Attractor Factor". Layer upon layer of thick oozing excrement from a Bull with scurvy. The disturbed state of search pages (outside of Google) are the residue of an industry awash in the unholy cowpie.  The fact that they haven't figured out that no one wants to be within 10 miles of their barn is only remarkable.

Nat Ersoz
Sunday, April 13, 2003

Teoma is pretty good too, the 7 odd times I used it.

Prakash S
Sunday, April 13, 2003

Stephen,

Before you say that, you should read the NYTimes article referenced in the "No telepones" [sic] thread.  Apparently the conventional wisdom, that nobody ever clicks on those things, is wrong.

Kyralessa
Sunday, April 13, 2003

OT: anyone know how to change to Yahoo or other engine in the Safari Google search box? Couldn't be hard-coded into the app, could it?

pb
Monday, April 14, 2003

I don’t think that the problem is that yahoo has a few extra ads as compared to google. Both google and yahoo do a reasonable job of keeping the annoyances of ads down to a min. That is not the problem

The real issue is that google has a core group of developers who are very good at making search engines. Further, they are better at making assumptions.

Fact is, google came along and absolute kicked yahoos butt. It was SO SUPERIOR that yahoo did the most incredible thing that most companies would NOT do!!!

    They simply licensed the engine from goggle.

I was always stunned, and VERY IMPRESSED that yahoo reacted SO QUICKLY on this. They saw the writing on the wall before everyone else did, and made a very fast, and not doubt HUGE PERSONAL biting of pride and simply licensed the google engine.

The problem is here is they did not get the google team, and that what I would went after.

I bring two examples of what I said about making “assumptions”, and how google is so much better.


When yahoo came out, they had the worst possible default assumptions for key words.

If I was searching for an article such as how the VBA language is the “glue” that holds office programs together, one would normally type in:

VBA glue office

Simply three keys words. The problem is that yahoo assumed that you wanted all the articles with VBA + all the articles with the word glue in it + all the articles with office in it.

Now, I don’t know about you folks, but that has got to be the most stupid assumption possible!! I have NEVER MET ANYONE who would want the UNION of those three words. (but I sure the programmers said that doing the intersection of those 3 words used WAY MORE processing). Anyway, after some time, people learned to click on the advanced search, or learned to use the “+” plus sign between each word, and gives you the INTERSECTION of the 3 (which by the way, any human with a brain would really want). How could anyone possibly think that a user would want the union of all 3 keys words. It is beyond reason.

Hence, it was, and still is clear that yahoo  make a brilliant decision to purchase the google engine, they still continue to make bad assumptions.

Of course googol from day one used and assumed the correct default for those 3 keys words. It took yahoo way too long.

Another great example of how much better google works on general searches and yahoo fails. (yahoo was REALLY STUPID to have a google search engine link on their main page, since when a yahoo search fails, I then click on google. Again, google is kicking yahoo butt on the “assumptions” area. If yahoo was BETTER, then google, then it would be ok to have a link to the competitor, but right now, can you say “stupid!”. You don’t put a link to a better search engine then yours on your main page.

How dumb can these folks get!! (perhaps that was a requirement, or stipulation of the google license, since the google people perhaps know they have such a better product, and if they could only get a link from yahoo, it would be the end of the beginning).

For example the following fails, or is all but useless in yahoo:

VBA GLUE Oiffice

Yahoo returns no results for the above. Google does a auto spell check. In fact, it is THE BEST SPELL CHECKING engine I have ever used. 

A yahoo search of:
Joel on softwere

It does give 10 NONSENSE matches. Of course google will ask:

Did you mean
    Joel on software ?

In other words, most of the young generation that can’t spell very well is going to use google, because it naturally corrects what you type in. Yahoo does not do this.

This is but only two examples of the fact the google team makes FAR BETTER assumptions about how things should work.

I have NOT been able to find one person who can agree with yahoo’s original defaults for key  word searching. NOT ONE PERSON. Well, if you have company that does something that NOT ONE PERSON can agree with the dumb default…then that company has a problem.

By the way, yahoo is still my home page right now. They were really first in the search engine business, and they are very market savvy. When my search fails in yahoo, I then whack the “try google” button. (and it 99% of the time corrects my misspelling and gives me the results I want).

Once again, yahoo needs to purchase the updated engine that google has. Or, the yahoo team needs a wake up call…or they need to hire me!

Albert D. Kallal
Edmonton, Alberta Canada
Kallal@msn.com

Albert D. Kallal
Monday, April 14, 2003

Albert,
          Originally Yahoo used to do its searches manually. That is to say it had a human look at all the pages and then decide how to classify them and how relevant they were.

Kyralessa,
                The person referred to in that article sells Armani and Hugo Boss suits - that is to say she sells articles priced at over $1,000 in an industry where the retail markup for high street goods is routinely around 200%. Obviously she can afford a lot of click throughs that don't get to be orders at $0.21 - $1.50 a shot.

                Also she doesn't actually get that many extra orders for her $60,000 a year ad expenditure. Round about 1,100 extra orders so her ad expenditure comes in at $55 an order.

                There are probably enough people in her situation, or in a field where the marginal costs are close to zero, where Google ads will pay, and so the ocmpany needn't fear bankruptcy, but that doesn't mean that banner ads are going to be any way the healthy surce of income they were once supposed to be.

Stephen Jones
Monday, April 14, 2003

>>>Albert,
Originally Yahoo used to do its searches manually. That is to say it had a human look at all the pages and then decide how to classify them and how relevant they were.


That changes nothing about what I said. My point still holds if  “people” created the indexing, or a web crawler did.

They STILL USED THE UNION of the 3 key words. That was a very dumb assumption. I VERY CLEARLY remember the first few times using yahoo, and getting very mad at this fact. Even today, I often see people putting + signs between each keyword since they rapidly learned that searching was near useless. (these people now don’t realize that the + signs are not needed anymore).

It took yahoo several years to change this. In fact, it took yahoo way too long to realize the obvious.

Simply put, union of keys was a VERY stupid default. When a company makes mistakes that the general public can see, it hurts them.

I not writing off yahoo in anyway here…it just that they were slow to change some things. So slow did they get, that they eventually simply bought the google engine. As mentioned…they need to buy the google team….


Albert D. Kallal
Edmonton, Alberta Canada
Kallal@msn.com

Albert D. Kallal
Monday, April 14, 2003

Regarding Joel's original post:

Joel noted that everybody has a different 20% of most used functions, so you cannot really leave functionality out.  He also seemed to imply that the usability community was saying that the best way to make a programme usable is to remove functionality and therefore avoid 'bloatware'.  I think this is a misunderstanding of the HCI community's actual position.  Usability specialists don't want to remove functionality, rather we want the interface to focus on the most commonly used functions or tasks.  The core 20% should be most prominent and best supported.

Even though we all have favourite individual features, there is remarkable similarity in the core tasks most users undertake using a piece of software (i.e. when word processing most people tend to format documents, save and print them, spell and grammar check them, etc).  A good interface focuses on supporting these overall tasks well.  Less used features can be available but less prominent and hidden away.

Just a thought...
Sherlock

sherlock_yoda
Monday, April 14, 2003

I don't think Joel would dispute that the most used features should be easily accessed.  The claim he was disputing is not really made by interface experts.  It is made by software people, twisting the words of interface experts.  Every few months you hear about a new product, be it a word processor, email client, whatever, that claims to be "lite" and quotes the 80/20 rule as a reason why it lacks so many features.  Such products have been known to leave out features such as word count in a word processor (though how you figure that's not in the 20% is beyond me).

I don't doubt that the 80/20 rule has validity and should guide UI design.  But as Joel said, designing features with that as your premise can lead to disaster

Mike McNertney
Monday, April 14, 2003

Albert,
            I wasn't disagreeing with you, just adding some additional information.

Stephen Jones
Monday, April 14, 2003

Word count is used by journalists and students and almost no one else cares. I don't remember at ANY point since I left university caring about word count. However since journalists review stuff, you'd better have it as a feature.

Peter Ibbotson
Tuesday, April 15, 2003

Word count is also used by professional fiction writers, who get paid by word count.

Brent P. Newhall
Tuesday, April 15, 2003

Prakash. Thanks for the Teoma post. Tried it once. Not bad.

Will use it in conjuction with google, and alltheweb

tapiwa
Tuesday, April 15, 2003

Yahoo is still the #1 internet property, with MSN, Daum.net (korean), Naver.com (also korean), and Google at #5 trailing it.

http://www.alexa.com/site/ds/top_500

Interestingly, Passport is #6, Yahoo! Japan is number 7, Microsoft is #8, eBay #9, and another Korean site #10.

3 of the top 10 are Microsoft, and 3 of the top 10 are from Korea.

www.marktaw.com
Wednesday, April 16, 2003

It appears Yahoo is #1 among users dim enough to install the Alexa toolbar.

Adam
Wednesday, April 16, 2003

When Yahoo first started, the web was a pretty small place.  Hence the UNION behavior.  Search engines sold themselves on being able to find as many relevant pages as possible.


If google had been around at the start of the web...

AOLer:  "I searched on yahoo for "Los Angeles Lakers" and it found like twelve interesting pages.  Google only found http://lakers.com.  Google sucks."

But yeah, they shoulda changed that behavior right quick.  But it's hard to see you're making mistakes when you're still #1.

Richard Ponton
Friday, April 18, 2003

*  Recent Topics

*  Fog Creek Home