Instant Results and Spin

After reading the transcript of the SpoolCast episode Luke Wroblewski and Innovations in Web Input, I decided to write about two subjects they brought up. It's a double post! What does it mean?!*

*Yes. This joke is stale.

Isn't Google Instant Awesome?

I mean, really! You start typing and (as long as you're not using Opera) results just pop up (unless your connection's too slow)! It's like magic! It makes Amazon's search suggestions look quaint and your public library's catalog search look positively archaic! Why can't they do magic like that, too?

Well, Google has approximately one quadrillion times the resources of your public library, for one thing. Something like Google Instant takes a lot of resources. If a company tries to implement something like that without the required resources, the resulting performance hit (and probably downtime) is as bad as or worse than not having search at all. Yahoo! actually tried to implement something similar to this in 2005, but they ended up ditching it because they were concerned about server load.

This does put smaller organizations at a bit of a disadvantage, because if people get used to seeing results instantaneously, they start expecting it everywhere. Why should I have to type my entire search query, hit enter, and wait for another page to load? (I guess you could argue that Google is aiding Twitter's nefarious plot to shorten the attention span of the world's population.) Unfortunately, though, businesses have to allocate resources to the activities that will provide them with the most benefit, so Google Instant-esque capabilities are probably going to be fairly low on the list.

Their Numbers Don't Mean As Much As You Think They Do.

I used this analogy recently…there's 160 million iTunes users, and within 48 hours they announce that one million of them were on Ping. So the metaphor I used there was, well if 160 million people are driving down a road, and I put a pile of dog poo on that road, chances are one million will drive through it. — Luke Wroblewski

In a world where Apple can do no wrong, their music social network Ping is a fledgling feature that hasn't yet reached its full potential. In the real world, well, it's been compared to a pile of dog poo. "One million" is a useless measure without context. I'd love if one million people read my blog. Facebook would probably kill a new feature that only attracted one million users. For the iTunes Store, one million is less than 0.7% of their users.

On the other hand, one could argue, for the time period mentioned in this anecdote, maybe 160 million users isn't the metric we should use to judge Ping's success. Surely, not all 160 million people are trying to buy music at the same time. Maybe only half of them are typically online at any given time. That makes Ping's user base…just over 1%. The point is, a company can spin numbers in a multitude of ways to make themselves sound better.

(Wroblewski goes on to suggest that the reason for Ping's disappointing adoption rate is an uncharacteristic lack of attention to detail on Apple's part. Version 2 might be better.)

Usability In The Rainforest

On the web, usability refers to the ability of users to find and use stuff on a website. (How's that for a circular definition?) According to Mel Pedley on Accessites.com, there are five core components of usability:

1. Learnability: On a first visit, can users find what they're looking for?
   
2. Effectivity: Can users do stuff quickly?
   
3. Memorability: If it's been a while since a user was on the site, can s/he still use the site?
   
4. Reliability: How error-prone are users while using the site?
   
5. Enjoyability: Do users like using the site (or, at least, do they not hate it with the burning passion of a thousand suns)?
   

Let's explore these concepts with a site we're all familiar with: Amazon.com. Sarah has been living in a rural area with ridiculously slow dial-up for the past several years, and has just discovered that you can buy products online. She needs to order a new DVD burner so she can mail her mom some home movies. She types "buy stuff online" into Google.

She sees that Amazon is the first result, and decides to click on it. (Hey, bonus points for findability!) She is then dumped onto Amazon's main page.

Okay, not a problem. Amazon's apparently a big site, which makes sense since her search query was very general. She looks around for "DVD Burners", or "Computer Components", or anything that looks like it could be what she's after. There's "Computers & Office" in the sidebar, and voila! There's "Computer Components" right there.

Lots of computer components to choose from. She remembers now why she asked her nephew for help the last time she needed to buy computer parts. Where are the DVD burners?

She scrolls down and – oh, there they are! Her nephew told her once that optical drives are the same as CD/DVD drives, so she clicks there.

She's still looking in the upper-left corner when the next page loads, so she immediately sees the link for "DVD Burners" in the navigation and clicks on it, too. No use in seeing all that other stuff she's not after.

Now that Sarah has found the DVD burners, let's evaluate how Amazon has been doing so far.

• Sarah could find what she was after with relative ease, so they get a point for learnability. (On the other hand, if she had been looking for something more obscure, she probably would have had a harder time of it.)
• She didn't have to waste a lot of time to get to where she was going, but she had the benefit of already knowing that "optical drives" and "DVD burners" were synonymous. Therefore, Amazon's effectivity was mediocre.
• This is a first-time visit, so we're not analyzing memorability.
• Sarah could easily have missed the link she needed to click on a few different times, because it was not necessarily put in a place that made the most sense. Reliability may or may not be an issue.
• So far, Sarah doesn't have a problem with using the site, so its initial enjoyability is not an issue.

What could Amazon do better? This is a hard question. One's first inclination might be to say that they should make it easier to find what people are looking for. The problem with that in this case is that because Amazon carries so many products, if they tried to make it easy to find everything, that would make their site too complex and cluttered to find anything. Burying some things three or four levels deep may be the best way to organize the information. Sometimes you have to make compromises to create the best site possible.

Architecture, Latches, and Taxes

Information architecture is the confluence of context, content, and users. What does that mean? It means you design your website around your content, while keeping in mind usability, the perspective from which your audience is coming, and any relevant "stuff" (for lack of a better term) from your business.

Richard Saul Wurman first coined the term information architect in the mid-1970s. He explains his choice of words as the following:

Unfortunately, design, which used to be a perfectly good word, means to make something look better for most people....The designer is called in to make a magazine article look better, or an illustrator is asked to make a picture look arresting, or a photographer is asked to take an interesting view of an author or a subject. Nowhere are any of these designers used in the fundamental sense of creating meaning or understanding.

That's why I've chosen to call myself an Information Architect....I mean architect as in the creating of systemic, structural, and orderly principles to make something work--the thoughtful making of either artifact, or idea, or policy that informs because it is clear. I use the word information in its truest sense. Most of the word information contains the word inform, so I call things information only if they inform me, not if they are just collections of data, of stuff (qtd. in Wyllys, 2000).

So an information architect's job is pretty much all-encompassing: figure out what you're displaying, and present it in the best manner possible. Organizing the information is a big part of that. Wurman lays out the five LATCH methods of organizing:

• Location
• Alphabet
• Time
• Category
• Hierarchy
  

An example of information arranged by category and hierarchy is Jess Bachman's Death and Taxes poster, which divides the federal budget by department and uses a proportionally-sized icon to indicate the relative size of that department's chunk of the budget. (As a side note, it should come as no surprise to anyone that the Office of Governmental Ethics' icon is visible only with a microscope.)

Web designers should be concerned with information architecture. Website content is divided into multiple pages, of course. How do you organize them? This is where the context comes in. Is this a large e-commerce site? Organize by product type. Is this a site dedicated to the history of Elbonian widgets? Organize by year. What are your users going to be looking for? Figure out how to get them to what they need, as effortlessly as possible. You are an architect. Construct your virtual building.

Web Writers: "You Have A Short Attention Span"

My conclusion is this. Write short, grammatical sentences and have other people edit them.

I probably should have broken that into more sentences. Let's review what isn't acceptable in web writing:

• Semicolons
• Long paragraphs
• Long headers
  
• Complex sentences
• Sentences that aren't stuffed with keywords
• Formal writing style
• Exclamation points
• Excessive bold and italics
• "To be"
  

Some of this is common sense, carried over from print writing. Others were developed to work within the constraints of screen real estate. (Remember 640x480? Those were the days.) Killing semicolons, however, is advice I find odd. How else would you punctuate a list that has items requiring commas? "Mr. John Doe, Esq., Dr. Ibee Grumpy, MD, and Mrs. Jane Smith" is not as readable as "Mr. John Doe, Esq.; Dr. Ibee Grumpy, MD; and Mrs. Jane Smith".

Whoops, I exceeded the sixty-word-per-paragraph guideline. Take a deep breath.

How many words should there be in a paragraph? Let's take a look. Ask a Manager's most recent post will serve as an example. Excluding the introductory "A reader writes" paragraph, she averages 77 words per paragraph. Even a web writer's blog occasionally exceeds 60 words per paragraph. The paragraphs that don't are often one sentence long or sound choppy. I think paragraph length will follow from your writing style and your topic.

The next web writing tip is a short paragraph. Don't make grammatical mistakes. Proofread your text. Have someone else proofread your text.

Yes, make your content engaging. Yes, remember that websites aren't in order like a book. Yes, make your stuff easy to read and navigate.

The most important point to remember? Good web writing is good writing.

A Library for IT Folks

ITIL is the Information Technology Infrastructure Library, a collection of documents describing IT best practices. ITIL was first developed in the late 1980s by the UK's Central Computer and Telecommunications Agency The ITIL philosophy is centered around a Service Strategy, which supports the areas of Service Design, Service Transition, and Service Operation; these processes are continually being improved (image from SUBnet.192):

ITIL has had 3 versions: v1, encompassing 30 books on various aspects of the IT process; v2, with 8 books; and v3, simplified to 5 books. ITIL is in the process of transitioning from version 2 to version 3, though both versions are still widely used by IT companies.

One of the aspects included in v2 -- specifically, the Service Support category of v2 -- is the Service Desk. It is pretty much what it sounds like: the company's tech support function. There are three descriptors for a Service Desk, each with differing levels of actual tech support provided.

• A Call Center is essentially a relay station for taking customer problems and forwarding them to the appropriate support department. While Call Center employees can solve some basic problems, most of the time they just get information and forward customers to another department that is better equipped to solve their problem.
• A Help Desk solves problems as quickly as possible. They provide support, but don't handle maintenance-type activities.
• A Service Desk handles tech support, maintenance, change requests, etc. Pretty much, if it has to do with their system, they handle it.

The Service Desk activity is categorized under the Service Operation banner in ITIL v3.

There are many software solutions for ITIL, spanning programs that cover the entire breadth of ITIL activities and programs that focus on one aspect of it. One more focused product is ServiceDesk Plus, "a web-based help desk software that...[manages] all your communications from a single point". It takes requests via different methods (phone, email, etc.), applies the necessary rules and/or conditions, and routes them to a technician for fulfillment. (Workflow image from ManageEngine.)

There are also several companies that provide ITIL consulting services. One such company is Enterprise Consulting Services, "a boutique consulting firm specializing in IT Service Management consulting, implementation and outsourcing, using the ITIL framework". ECS focuses on translating ITIL best practices into concrete implementations based on a company's needs.

One such client with whom they have worked is Iberdrola USA (formerly Energy East), a multi-state energy corporation. ECS provided training for employees, allowing them to become certified in ITIL. They also redesigned the support and security systems and upgraded their computers (Windows XP, Office 2003, and McAfee security -- in related news, Iberdrola probably just had svchost.exe disappear for a while). These moves resulted in annual savings for Iberdrola of about $1.4 million. That's what applying optimized practices will do for a company.

Quote of the Day:
The Tech Support Glossary:
ID-Ten-T error: the user has just done something inane, like use their DVD drive tray as a cupholder. (Also known as ID10T.)

PEBKAC error: there is nothing wrong with the computer; the problem exists between keyboard and chair.

Layer 8: in the OSI Model, the user layer above the application. (See also ID-Ten-T.)

(from Wikipedia)

Buildings, Triangles, and Big Balls of Mud

Christopher Alexander had a problem. Actually, all architects had a problem. There they were, being hired to design buildings, and their clients understood the buildings more than they did. The problem, essentially, was translation. How could an architect take what the client described and turn it into what the client wanted?

Simple, Alexander said. Clients want basic variations of a set of buildings. We need a language that describes basic stuff (that's a technical term) in our problem domain, building buildings, that is simple enough to be understood and used by our clients: a pattern language. "The elements of this language are entities called [design] patterns. Each pattern describes a problem that occurs over and over again in our environment, and then describes the core of the solution to that problem, in such a way that you can use this solution a million times over, without ever doing it the same way twice."

Design patterns are not one-size-fits-all, cookie-cutter solutions. Instead, they are skeletons, templates, general descriptions of how one might solve a type of problem. For example, it has been determined that people like rooms where natural light comes in from two directions. That's a design pattern. Notice that the pattern does not say, "Put one 16"x30" window on the north wall, and one 24"x30" window on the west wall." It says, "The optimal room has natural light coming in from two directions."

Alexander came up with this concept in 1977. To this day, it has had a mixed reception in architecture. In 1987, however, Kent Beck and Ward Cunningham got the idea of applying the principles of design patterns to software. The "Gang of Four" (Gamma, Helm, Johnson, and Vlissides) popularized that usage in their 1994 book, Design Patterns: Elements of Reusable Object-Oriented Software (Wikipedia).

Good design patterns, in computer science, are classified in four categories: creational patterns, structural patterns, behavioral patterns, and concurrency patterns. Creational patterns define how an object is created. Structural patterns define relationships between entities. Behavioral patterns define methods and patterns of communication between entities. Concurrency patterns define how to deal with multiple threads in a program. Wikipedia has some good descriptions of software patterns.

An example of a good design pattern is the creational pattern abstract factory. An example of an abstract pattern is the AbstractCollection class in Java. This class provides the basic skeleton required to create a collection of objects. It has methods to add items to the collection, remove items from the collection, get the size of the collection, etc. Individual methods in the class can be overridden in subclasses to provide functionality specific to that implementation. Several classes implement AbstractCollection in a more specific way. ArrayList, by way of AbstractList, implements the methods so that they work with a resizable, sortable list of objects of a specified type. List implements them to work with a list of specifiable, but fixed, size.

To put it in simpler terms, an abstract factory pattern is like the definition of a triangle. The abstract definition of a triangle is that it is a shape with three sides. The abstract methods would be implemented in the classes RightTriangle, ObtuseTriangle, and AcuteTriangle, which would place restrictions on the angles of the triangle, as shown below (Wikipedia).


Bad design patterns are called antipatterns. While good patterns have dry names like "abstract factory", "composite", and "chain of responsibility", antipatterns have been given such monikers as "spaghetti code", "God class", and (my favorite) "big ball of mud". The big ball of mud starts out as a relatively simple piece of software, then through the addition of bits of throwaway code, quick fixes, and piecemeal growth, becomes an unwieldy, mutant program that would take a good deal more time and money to fix than it would have taken to properly maintain it in the first place. By leveraging all of my considerable artistic talent, I leave you with this insightful look into big balls of mud:

Google AdWords

Google ads are everywhere. They appear on searches and websites, usually identified as "Sponsored Links" (since that sounds so much classier than "Ads"). Google's ad programs are more useful than some other ad services, since they are driven by context - instead of getting a random "You are the 1,000,000th visitor to this site! Click here to claim your prize!" you will see something related to what you're reading. This, in theory at least, provides a better return on investment for the advertising company.

There are two sides to the Google advertising platform: AdWords and AdSense. AdWords is for companies or individuals wishing to purchase ads based on certain keywords, while AdSense is for companies or individuals willing to place ads on their website. I'm going to discuss AdWords and how it works from the customer's perspective. Here's a flowchart giving a basic overview of making the decision to use AdWords. It's much easier to read if you click to enlarge it.

Our company - we'll call it Fabulous Widgets - has limited marketing funds, so Marketing requests the budget from Accounting to verify how much they can spend. Once they know that, they can start researching keywords. The keywords for which they purchase ad placement will dictate the searches on which their ad will appear. Google's Keyword Tool provides information about, and suggestions for, keywords. In this case, the Marketing Department wants to purchase ads for searches that are as relevant as possible to their core product - widgets.

After typing "widget", "widgets", and "fabulous widgets" in the keyword box, the Marketing employee is presented with a list of possible search terms, along with popularity, search trends over time, and (optionally) the average cost per click, or CPC, for each term. The top matches include "widget", "dashboard widget", and "web widget". Well, Fabulous Widgets is a purveyor of only the finest dashboard widgets, so "web widget" is not really an applicable search term. Marketing decides, based on their budget, to purchase ads for "widget" and "dashboard widget".

Periodically, Marketing will need to review the performance of the ads they've purchased. How much more business is Fabulous Widgets getting from the new ads? Maybe AdWords is providing them with more profit than a different marketing campaign, and therefore needs more money allotted to it. Maybe the keyword choices should be tweaked. The Internet is a fluid place, and so, by necessity, must Fabulous Widgets' marketing efforts.

Quote of the Day:
I need subtitles walkin' in here.
--Booth, Bones, "The Man in the SUV"

Enterprise 2.0

Facebook, wikis, blogs, YouTube...what do these have in common? They are all examples of Web 2.0.

Web 2.0 is not like HTML5. It's not a new specification, it's not a specific way of formatting or configuring anything online, it is defined by the behavior of the users interacting with the sites. It is community-driven, dynamic, extendable. It's also everywhere.

So it follows logically that business would want to capitalize on that community, and so Enterprise 2.0 was...well, perhaps "born" isn't the right word; "coined", then. Enterprise 2.0 is merely taking the ideas inherent in the Web 2.0 model and applying them to business situations. (Forrester Research begs to differ slightly; according to them, Enterprise 2.0 does not include anything aimed at consumers. I personally disagree with that statement: how else are businesses going to reach consumers, if they're not using Facebook and Twitter-type technologies?)

Are you in the engineering department and you need to collaborate with your coworkers in design? Set up a wiki to provide a centralized place to put mock-ups, specifications, etc. Need quick bursts of promotion? Fire off Tweets about discounts or product launches.

Sounds great, right? So why aren't all businesses Tweeting and Facebooking and blogging and wiki-ing? Dion Hinchcliffe has compiled a list of 14 reasons why Enterprise 2.0 initiatives fail. In essence, most of the problems boil down to the fact that some managers treat Enterprise 2.0 as an afterthought, something that doesn't need to have many resources allocated to it. After all, one may think, how much effort does it take to Tweet "14-inch widget on sale for 20% off, limited time offer"?

But it's not just about Tweeting, or wikis, or any one thing. It's about collaboration across the organization, destroying barriers to communication, and paying attention to the customers. Only by gaining the acceptance of the entire organization can Enterprise 2.0 initiatives succeed.

Extra reading:

• "If Enterprise 2.0 is a Crock, Second Life Enterprise is the Stuff That Fills the Crock"
• "What is Enterprise 2.0?"

Quote of the Day:

Booth: A prodigy violinist disappears and a month later his skull ends up bouncing off a garbage truck?

Cam: Obviously, we are looking for someone who really, really hates classical music.

--Bones, "The Widow's Son in a Windshield"

All Kinds of 'bilities

Findability is the quality of being able to be located by potential users. You may have the greatest website about (insert favorite subject here), but if it's buried, unorganized and unlinked, in the depths of Angelfire, nobody will ever know about it. Findability != SEO; it is focused on humans, not search engines. If findability efforts benefit search engine rankings, that is good, but it is not its primary aim. Findable websites not only facilitate initial visits, but they encourage visitors to return and to recommend the website to friends.

Findability does not exist in a vacuum. It is a part of SEO, but it is not a bunch of keywords stuffed in a page. It is useful content, organized for people to find, read, and use. Once people find your website, they have to be able to use it. Usability describes how easy it is to, well, use the website. Usability is different from accessibility: usability is more about organization and ease of use, while accessibility is about making sure that the website is available to the maximum number of people.

Members of Aarron Walter's Findability class at the Art Institute of Atlanta came up with the following definitions for findability, accessibility, usability, and SEO:

Findability: The quality of being able to be navigated or located.

Accessibility: The general term used to describe the degree to which a product (e.g. device, service environment) is accessible by as many people as possible. Usability and accessibility are often confused. See Also: http://www.w3.org/WAI/

Usability: The term that describes the quality of a website that determines how well tasks can be accomplished and how easy the website can be used. This can be broken down into many qualities, such as learnability, memorability, and efficiency. See also: http://www.useit.com/alertbox/20030825.html

SEO:Search optimization is a hyper-competitive endeavor that requires intense focus and a thorough, up-to-date understanding of how the search engine algorithms and robots operate.

(source)

Resources:

• The 10 Worst Findability Crimes
• Findability, Orphan of the Web Design Industry
  

Quote of the Day:

http://www.flickr.com/photos/morville/ / CC BY-NC 2.0

Web Analytics and SEO

I've written a paper on the use of web analytics and search engine optimization. It's available on Google Docs. If you haven't been given access, and you wish to read it, comment on this post with your email address and I'll see about setting up view permissions. Don't worry, I will not display your email address or send it to those nice folks in Nigeria. :-)

Quote of the Day:

Castle: We make a pretty good team, you know. Like Starsky and Hutch, Tango and Cash, Turner and Hooch...

Beckett: You know, now that you mention it, you do remind me a little of Hooch.

--Castle