CircaSpecting: People, Tech, and Business

Lyza: Spreadmart-free Data Analysis

Last week, I met with a couple dozen industry experts and journalists to do show-n-tells on Lyza, in advance of our official launch on September 22.  Lyza, in case you missed my earlier blog hints, is our new desktop software for data integration and analysis. 

Anyway, after several years of research, design, and engineering work, you'll understand why I breathed a sigh of relief when....they liked it. 

Here is a sampling of what they've been writing about Lyza:

all-in-one data analysis tool for the desktop ...well thought out ...highly graphical ... very wysiwyg

an easier, more intuitive way of pulling data together from diverse sources, combining and manipulating it and creating results and reports for distribution

well suited to analytic needs...a graphical environment allowing analysts to build up analysis iteratively...access to most data sources...auditability

Read some of the articles via these links..

Barry Devlin, BI/DW author and former IBM Fellow
http://www.b-eye-network.co.uk/blogs/devlin/

Andy Hayler, CEO of The Information Difference
http://www.it-director.com/technology/data_mgmt/content.php?cid=10744

Shawn Rogers, Editor at The Business Intelligence Network
http://boulderbibraintrust.org/brain_trust_blog/2008/09/lyza-software-launches-at-the.php

Now, try it yourself, with a free trial download from www.lyzasoft.com.  And, spread the word!

Imagination is Elastic

The bigger the space occupied by your product -- the more specific and detailed and tangible and exhaustive (exhausting?!) your description of what it will do for your customers -- the less room is left for ... the customer.  For his or her imagination, for the unique mark, for the creative -- for identity.  Imagination is very elastic; it will contract to nothing if you leave no room for it.  (Kathy Sierra wrote a great post on this topic, here.)  No matter how wonderful our product’s features, we have to remember: it’s the imagination of the user that binds the features, holds the whole together, and makes the product come alive. 

Recently we've seen this demonstrated powerfully in our user labs, when people use features in ways we don't expect -- extending the definition of the product itself.  If we dictated to the lab participants exactly how to use something, we'd probably get a pretty boring regurgitation of our own views. 

So, our central challenge is to recruit the user’s imagination to the task of creating our product from the features we’ve engineered.

When products falter, the missing ingredient usually isn’t a feature; it’s IMAGINATION ... as in, the users'.  Make room for it.

Then with a pleasure which is both sensual and intellectual we shall watch the artist build his castle of cards and watch the castle of cards become a castle of beautiful steel and glass.  Nabokov

Now THAT is Cool: Skyrails

Though I work in technology, most technology makes me yawn.  The jaw-dropping, genius, cool-n-useful tech innovation is a rare thing.  But, I found something today that makes me babble on in a stream of Wow and Cool and OMG!

Skyrails

There are so many amazing, engrossing, and effective applications of this tech.  Nice work, folks.

Usability, Early Adopters, and Crossing the Chasm

We've spent a great deal of time and money this year running observation labs to evaluate user cognitive processes, toward the goal of understanding how to make Lyza (www.lyzasoft.com) software that is both powerful and easy to use.  While this research has been invaluable, it also contained a trap we did not foresee.

Initially, we thought that Early Majority users would express higher expectations for usability than did Early Adopters.  But, they did not.  Their expectations for good usability were almost identical.

At first, this puzzled us and made us wonder if we were perhaps ready for the hockey stick of the mass market.  (Given the optimistic psyche of entrepreneurs, this trap is particularly dangerous.)  Then, we noticed something else -- Early Adopters showed significantly higher tolerance for a gap between their expectations and the product's actual usability.  EA's were just better able to get past, work through, or work around perceived deficiencies than were Early Majority users.  Fortunately, the labs allowed us to go beyond simply asking about expectations and allowed us to observe the differences in behavior and tolerance. 

Many new tech projects see a stabilization of expressed expectations at some point, encouraging them to scale-up marketing of their products in hopes of tapping the exponential growth curve in the first graphic.  But, they jump in without the significant additional investments of money and staff required (see second graphic) to improve the product faster than expectation-performance gap tolerance is declining in their target market (see "The Cliff" in the first graphic).  I think these conclusions are simply another perspective on the now-classic Crossing The Chasm argument: don't count your money based upon Early Adopter response, because there is a lot of work left and a very tight window within which to do it.  As a professor of mine once said ...

Growth consumes large amounts of money and staff.  Have stockpiles of both before you start.

The consequence: we need to hustle to grow our Lyza engineering staff ASAP.  So, if you're an expert Swing developer, send me your resume.

Software and Dancing?

Sorry I've been so quiet the past few months.  We're launching a new software venture (www.lyzasoft.com, shhhh -- we're still in beta), so I've been balancing two jobs. 

Over the last couple of months at Lyza, we've been busy with some intensive research programs focused on software-user cognition, which I'll talk about in some posts next week.  Though we've been driven heavily by secondary research in this field for the last couple of years, our own recent primary research has allowed us to watch and get-inside-the-heads-of dozens of analysts working through their typical challenges of gathering data, analyzing it, reporting on it, etc. 

And, I can say with some certainty that William Butler Yeats had it right:

How can you tell the dancer
from the dance?

Analysts' work defies tidy boundaries like Business Intelligence, Data Integration, Desktop Productivity,   Data Warehousing, etc.  Technologists and strategists spend a huge amount of time and energy debating ideal dance moves ...

How will Search work within Business Intelligence? 
Should Data Integration products enable SOA? 
How can you deliver reporting without a Data Warehouse?
What's wrong with spreadsheets?

...without actually slowing down, watching, and absorbing the real people performing the dances.  When you see the dancer and the dance as inseparable, those arbitrary technical categories blur -- and you realize that the REAL cast of people involved in data integration, analysis, reporting is FAR, FAR larger and more diverse and more dynamic than you imagined. 

A Simpler BI Scorecard

Ben Schneiderman does an excellent job of classifying the tasks in the creative, analytical, innovative (aka “thinking”) process in his book Leonardo’s Laptop. 

Since, theoretically, the purpose of Business Intelligence is to feed people’s thinking, Ben’s framework seems like a perfectly simple scorecard for evaluating BI tools.

What follows are my assessments of the standard BI tools along Ben's criteria:

F    Searching for data
F    Visualizing data and process, to discover relationships
F    Consulting with peers
F    Thinking by free association to create new combinations of ideas
F    Exploring via what-ifs and simulations
F    Composing artifacts and performances step by step
F    Reviewing and replaying to support reflection
B    Disseminating results to gain recognition and to support the community
F    Allowing for all of the above steps to happen recursively and iteratively

It’s easy to see why standard BI tools are not the solution for the Thinking Caste. 

Technology That's Too Logical

Is there a software product you enjoy using?  One that you actually look forward to launching?  Odds are, it’s a product that does not require you to file a comprehensive and detailed work plan in advance, because that's a buzzkill.  You have fun with software that allows you to improvise, experiment, goof off, assume, adapt…to be opportunistic. 

It’s more than just convenience; it’s technology that amplifies you, not technology that interrogates you or spoon-feeds you like a kindergartner or locks you into a prescribed paint-by-numbers procedure.

Computer scientists and information technology professionals have the highest degree of introversion of any profession studied.  They prefer to work on problems in isolation, so the social issues of dealing with real users may be troubling. Ben Schneiderman, Leonardo’s Laptop

People do not always behave as full, logical, reasoning organisms, starting with high-level goals and working to achieve them.  Our goals are often ill-formed and vague.  Actions may be executed before they are fully developed.  Opportunistic actions are less precise and certain than specified goals and intentions, but they result in less mental effort, less inconvenience, and perhaps more interest. Donald Norman, The Design of Everyday Things

We need more software developers who create products for real-world whole people (aka emotional, inspired, insane, intuitive, impatient, illogical, goofy, brilliant, creative, analytical, etc) -- not for caricatured make-believe users like Mr. Spock. 

BI = Cheez Whiz (or, Why Spreadsheets Still Win)

Back before the term Business Intelligence was shoved upon us, analysts called it Decision Support.  Seemed like a reasonable name: someone needs to make a decision between alternatives, and so we want to evaluate the situation to figure out which one is best.  It was active, not passive.  It was a discovery process that proceeded (non-linearly and recursively) from "I am not sure" to "I know."  In other words, it was clear that the term Analyst means "One who proceeds from not knowing to knowing."

Then along comes BI, and suddenly the word Analyst is supposed to mean "One who has the answers."  (Newsflash: that’s called an oracle, as in the one at Delphi.)  Because these expensive, new, complicated tools don’t allow for much true analysis, the vendors work to shape our language.  Instead of analysis as an active process of discovery, they try to reduce it to a process of regurgitating pre-processed answers.  That bears about as much resemblance to Analysis as Cheez Wiz does to Tillamook.  To get a feel for just how wildly misguided the BI mission has become, check out Neil Raden's Feb-2007 article in Intelligent Enterprise.

Spreadsheets remain the tool of choice for analysts because analysis proceeds TO knowing FROM not knowing.  Uncertainty and discovery and trial-and-error and experimentation cannot be assumed away, despite the wishful thinking of the BI vendors. 

To find/build good technology for analysts, start with the right understanding of the job.

Chasing a Mirage: The 1% Feature

I see a lot of technology groups (software companies and internal IT project teams) focusing on features.  Communicating features, comparing products based upon features, slaving to build a remarkable new feature.  Here's the problem: people are more than just the sum of their tasks; they are lives.  In a human life, style points count. 

So, for a piece of technology to connect with a person, that technology needs to provide an experience that is more than just the sum of a bunch of cool features.

In fact, the ROI for a new feature declines exponentially the cooler the feature.

Ask yourself two questions about your product or project:

1. What IQ does your user/customer need to appreciate and make proper use of your technology?
2. What is the probability that your user/customer will take your technology and bend/remix/retask/extend it in ways you did not intend?

If you refocus what you are doing so that you reduce the first answer by 20 points and raise the second by 50%, your odds of delivering something remarkably valuable will improve dramatically as compared to focusing on delivering cool features.

Data Warehouses are Child's Play

Sometimes we REALLY overcomplicate things.

Technical and line professionals spend a HUGE amount of time debating what should be included in a data warehouse, so that users will be able to get answers to their questions.  It’s really not that hard.  Here is a (partially) tongue-in-cheek tip that will help you cut a few months out of the project plan next  time: play 20 Questions.  Seriously, go to www.20q.net and play it a couple dozen times using weird/random stuff -- and write down each question the game asks you.  Type?  Size?  Color?  Sex?  Etc. Etc. 

If that application can figure out what you’re thinking in just 20 questions, those descriptors must be pretty powerful.  So, you’d have to figure that if your data warehouse included similar descriptors, it would allow users to find the answer to just about any reporting question they are apt to ask about the enterprise.