Some of us like to name things in BI

If you haven’t already, ask around: Exactly what is “business intelligence”? Some say it’s all about business decision making, and others seem to think it’s all about tools.

We struggle with definitions, but usually not in public. Perhaps that’s why the recent uproar on the weblog of eminent visualization critic Stephen Few felt like a refreshing breeze.

It all began with Few’s damning review of a product whose promoters tripped and gave it the now-sexy “visualization” label. Oops.

Usually, Few’s readers sit back and enjoy the show. He’s one of the few Bi writers with the courage to call out a stinker. But this time, several people sat up in protest. Comments erupted into a weeks-long discussion-turned-scholarly-fistfight over definitions.

After a few swipes at his “mean-spirited” tone—which I don’t see—and other complaints, they found the deeper issue. Colin White, president of BI Research and a keynote speaker at this year’s TDWI World Conference in Las Vegas, arrived late to the discussion but soon led the charge.

One term they fought over was data visualization. To Few, it’s a business function. He wrote that it’s “the use of visual representations to explore, make sense of and communicate data…”

White disagreed. He prefers a more “pragmatic” definition to accommodate the term’s variety of uses. He wrote, “If data or information is presented to a user in a format that aids decision making, then that contitutes data visualization.”

Though White writes that experts must “use clear definitions and terminology,” he wrote in the next sentence, “However, it is important that we accept that other people may have different definitions, and we need to find common ground.” He went on, “We also have to accept that business users may employ technology and use some terms in a completely different way, and it is important to adjust our positions and explanations accordingly.”

Did he mean that terms mean what the person who uses them says they mean? White leaves that and other things unclear in his careful yet still foggy pronouncements. He doesn’t even state his definitions of business intelligence and data warehousing, even when he condascends to Few that his definition is “outmoded.”

Few politely called White’s definition of data visualization “not useful,” and I agree. No term can be useful that has lost its meaning. As Alice said to Humpty Dumpty in Alice in Wonderland, “The question is whether you can make words mean so many different things.”

Label inflation makes it tougher to find a toehold in the market, to write about techniques and tools, and even to have a conversation. When marketing collateral shouts “data visualization” to the general BI market, who will look up if it could mean bad Powerpoint slides? It hurts the whole industry if worthy products can’t find words that make would-be buyers listen.

Few’s review of Lyza looks to me like a case of mistaken identity. Perhaps the company should never have entered the visualization arena. Also, according to at least one BI expert I respect, it is actually a valuable tool. A bloody nose for nothing.

To Alice’s question about making words mean many things, Humpty Dumpty replied, “The question is which is master. That’s all.”

If everyone’s a master, we have label chaos. Instead, industry leaders, journalists and smart marketers should use words as they’re most widely understood. As a rule, the master should be business, the data train’s final stop.

Also see sascom editor Alison Bolen’s “What we call what we do: a lesson in evolving industry key words.”

9 Responses to Some of us like to name things in BI

  1. Ted,
    I read your blog with interest, but I am not quite sure how to respond to it! It might be worth a discussion on the phone before you publish it as an article because I think you are missing several key pieces of the story and discussions. A few quick points, however. Actually I was not late to the discussion. I was involved in the discussions from the beginning. Many of them were offline because Steve is a friend. Also, the objective of the article I wrote was to bring together a set of opinions from various experts and summarize them. I wanted to move the discussion away from the “review” of Lyza. My summarization in the article did express my viewpoint but it was also a summary of the numerous discussions I had with people as I gathered their input. The majority of the discussion (and disagreement) centered around the term data analysis, rather than data visualization. Steve specializes in business user visual data analysis which is a combined subset of data analysis and data visualization. He feels that any of the data analysis work done during data integration is not data analysis. The majority of the people in the discussion disagreed with this viewpoint. Data integration does involve data analysis (e.g., data profiling). It’s just that the results are used more by IT users than business users. I didn’t define BI or DW in the article because most people have a pretty good sense of what these terms mean and this was not the objective of the article. The point I was trying to make about Steve’s view of BI in a later blog was that most of his arguments are based on the traditional viewpoint that BI and data warehousing are inseparable. With the industry move toward operational BI and in-flight analysis this tight association is only true for certain types of BI. With the move toward a more dynamic BI approach, the BI model changes from an store data (in a data warehouse) and analyze it model, to an analyze data in flight and store the results model. This makes it even more difficult to separate data integration and data analysis and traditional definitions have to evolve to deal with this. As I said at the beginning I would be more than happy to discuss this in more detail and get your input on this topic. This is most probably a better approach than simply using blog material which tends to be instant and reactive in nature! Colin White.

  2. Ted,

    Thanks for weighing in on the discussion. After reading Colin’s response, I’d like to clarify a few points.

    Regarding the origins of the discussion, Colin directed our interaction into one about definitions (data analysis and data visualization), but he was not one of the original people to respond to my blog about LyzaSoft, which was what gave rise to the discussion. As such, you are right in saying that Colin came late to the discussion (that is, more than a month after it began), and Colin is right in saying that he was involved from the beginning, if you consider only that part of the discussion involving definitions.

    To correct one of Colin’s statements, I do not believe that “any of the data analysis work done during data integration is not data analysis.” Data analysis is data analysis whenever it’s done. What I have stated during this discussion is that data integration is not data analysis; data integration is something that we must often do to prepare data for analysis. If you engage in data analysis–that is, the process of making sense of data to understand the meanings that reside therein–while in the midst of data integration activity, you are interspersing data integration and data analysis activity, which we often do, and legitimately so. A software product that provides the means to integrate data but not to analyze it is not a data analysis tool; it is a data integration tool. Similarly, a software product that provides a primitive set of charts, as Lyza does, is not a data visualization tool.

    Regarding my view of business intelligence and data warehousing, I do not define them as inseparable. The concern that I’ve expressed over and over in my work is that, even though business intelligence by definition ought to be a human-centered industry that supports people in their efforts to make sense of data, leading to better decisions, with few exceptions it remains a technology-engineering-centric industry that exhibits little understanding of the needs and abilities of human beings.

    As business intelligence now strives to become more dynamically integrated into the flow of peoples’ work, it will fail unless it shifts from its predominant technology-engineering focus to one that takes the time to understand human beings: what they need from data, their strengths (such as visual perception), which can be tapped to meet these needs, and their limitations (such as limited memory), which must be augmented to support the analytical process. Unless designers with an understanding of human factors, usability, cognitive psychology, human-computer interfaces, etc., are involved in the business intelligence industry–not as an afterthought, but as leaders–BI will continue to provide faster and faster algorithms that end up doing little for human beings.

    My concern with Lyza was not that they attempted to support a seamless interaction between data acquisition, integration, transformation, analysis, and presentation, but that they failed to provide anything useful for data analysis and presentation. A good vision, but a failed execution.

    Take care,


  3. Ted,
    As I mentioned in my first reply to your blog my main mission in the material I published was to get together a group of people and try to reach consensus on definitions. Although the topic has become controversial that was not the intent of my article. Part of the reason it has become controversial is that it keeps being connected back to the original Lyzasoft blog and the good or bad job certain BI vendors are doing in the industry. I was trying to separate the two topics.

    On the topic of definitions I have sent you a draft article that is a follow up to the original one on the BeyeNETWORK. I am again gathering input before it is published. It is too big to put here. I look forward to your comments.

    On the separate subject of Lyzasoft I think this topic has been beaten to death and has run it course. I am really not sure how much more can be said about it. I would however like to make one final point because you did make a comment that you disagreed with my statement that Steve’s review was “mean spirited”. I stand by that statement for several reasons. The situation is summed up quite well by another blog about Steve’s review by Donald Farmer:

    “Ouch. This strikes me as a rather mean-spirited article. Perhaps the pun was just irresistible [this refers to the title of the article ‘Introducing Lazysoft: An engineer’s playground, but a data analyst’s nightmare.’] Even so, it is unmerited. Did you speak with Lyzasoft about their design and development processes? I ask because, from what I know of them, they take the process of design with deep seriousness. They do something that is all too uncommon in software development and rare indeed in a new vendor shipping version one: they test usability. In fact, I know from my contact with them that they test usability and all aspects of their design very thoroughly and they act on the results. This is far from lazy – the insult is deeply undeserved and unnecessary.”

    I agree completely with Donald Farmer. Steve did not do a product review. He simply reviewed the marketing literature on the web site. He did not get a briefing on the product, evaluate the product, discuss his issues with the company, or look at the usability testing they did. Like Donald I did all of these things and that is why I agree with Donald’s blog comments. The company during this period was willing to share this information with anyone who asked. I am the first to agree with Steve that the company did a poor job of their marketing material. I also with agree with you that Steve does a good job of keeping vendors honest. I don’t always agree with his style but that is okay.

    In this particular case Steve fell down on the job regardless of whether his comments about the product are correct or not. When I asked him about this his reply was he simply doesn’t have the time to evaluate products at that level. My view if you don’t have time to do a good job then you shouldn’t review the product. If you have serious issues with a product (particularly a small start up) then I believe you should discuss them with the company before writing a scathing review. I also believe even if you write a negative review it can be done in a professional and non-confrontational manner. This is why I described the “review” as mean spirited. Of course this is only my viewpoint because these are the rules that I follow. My objective is not to create controversy in reviews but to be helpful to folks that read my articles.
    I am not sure where you go with this. I believe this topic has been beaten to death. Good luck! Colin.

  4. I’m not going to add to the discussion here because I think the positions are clearly enough drawn. But tangentially, I do want to support Steve’s definition of BI because I have held the same position for many years, in print and in person:

    “business intelligence by definition ought to be a human-centered industry that supports people in their efforts to make sense of data, leading to better decisions, with few exceptions it remains a technology-engineering-centric industry that exhibits little understanding of the needs and abilities of human beings.”

    The dismal uptake of BI in organizations is a direct result of this problem. I’m not going to criticize anyone, but if you examine the work of my colleagues, you will find it overwhelmingly dominated by IT architectural diagrams, data and process flows and virtually nothing about how people make decisions with information and how they use information in their work.


  5. Ted,

    I feel compelled to respond to Colin’s remarks about the “mean-spirited” nature of my blog comments about Lyza. Both Colin and Donald Farmer of Microsoft, who Colin quotes above, are friends of LyzaSoft’s CEO, Scott Davis. I don’t believe that Colin’s description of my response, when he challenged my comments about LyzaSoft, accurately expresses what I said. Here’s what I wrote to Colin, word for word:

    “As I have now said in the blog discussion, my critique of Lyza is completely consistent with any critique that I have written of a product that I find ineffective. The tone of these critiques is usually provocative, which I believe is professional, appropriate, and the kind of response that is warranted when a vendor makes erroneous claims. Take a minute to review previous blog posts and you will see that this is the case. Perhaps you haven’t read these other posts, or if you have, perhaps this one seemed different because you know and like Scott Davis and have empathy for him, which is understandable.

    I have not engaged in tabloid journalism (a claim that Colin made), and find the accusation upsetting. I have engaged in the kind of journalism that you will readily find in fine publications like Newsweek, the Economist, etc. If I were writing about Sarah Palin in this fashion, I suspect that you would be cheering. I have done a fair and accurate evaluation of the data analysis and presentation capabilities of Lyza, based on everything that LyzaSoft has written and shown. I have clearly stated the limits of my review and have restricted my judgments to the things that I’ve seen. Taking the time to review the product thoroughly would not change the facts that I have stated.”

    Take care,


  6. I agree with Neil, extracting engineering from his post, it’s like saying engineering is about tools and equipment, when they are barely dealt with in an engineering degree.

    However, as I mentioned in a post (completely in a parallel world of my own, blissfully unaware of this discussion) what is important is not that I agree with Neil or which analysts agree with each others definition, it is the understanding of the consumers and users – the masses if you like about the tools at their disposal in the practice of business intelligence.

    Arguing about semantics is truly futile. If we take ‘visualization’ then we can say that text is a visualization of an idea, or speech. We can also go up the stack and say a check or a cross is the next level, symbols instead of text.

    The next stage might be ASCII art, then line charts, etc, etc. Is there a granular hierarchy of ‘visualization-ness’ or is it simply a boolean.

    Colins thoughts on terminology are right on the money, but the next step is to try and actually do something. Reading the threads it appears that Colin and Stephen are friends. If friends are fighting, then it seems like a referee is needed.

  7. Ted, I thought to jump in here as the original dissenter to Stephen’s blog. I enjoyed yours too and the many comments.

    We have moved on from the specifics of Lyzasoft. Nevertheless, in passing, I will clarify my original objection just one more time. My view was simple. I knew that Lyza had conducted considerable customer-focused HCI testing. I know from personal experience that Lyza’s approach to HCI is exceptional for a small vendor. I believe it is laudable that they did so – whatever Stephen, or I, think of the results. (I was not defending the results!) I know it takes money, time and effort to conduct that research. Therefore, I protested the casual pun that Lyza were lazy. I especially dissented because the criticism referenced only weak marketing materials rather than a product review, or a discussion with the vendor about their methodology. My motivation was not friendship, but rather fellow feeling for a new vendor: I have been in that position myself. Of Stephen’s style in general, I should say that I subscribe to his blog, and when my RSS reader flags an update, it is time to grab a latte and to sit down for a few minutes reading: minutes that are always well spent.

    Now to the meat of this thread. Ted suggests, “Industry leaders, journalists and smart marketers should use words as they’re most widely understood.” It is hard to disagree, but of course, the devil is in the detail, and in this case, the detail is blurry. Colin is quite right that words evolve, regardless of our attempts to restrain them. The most familiar example is “gay.” As Dylan Moran says, “The meaning of the word gay has changed. It used to mean all colourful and happy and homosexual but now it’s a word children use to describe something that’s a little bit bleh. ‘You’re eating Weetabix, oh that’s so gay’ ”

    As words move from a restricted sphere of influence to general usage, we must expect these changes, and can barely resist them without being Humpty-Dumpty’s ourselves.

    This is happening as we speak with Business Intelligence. As Stephen says, “business intelligence now strives to become more dynamically integrated into the flow of peoples’ work.” If this is successful, and it appears to be, then more people will be familiar with our terms and the terminology will evolve, blurring our favoured definitions.

    This has already happened with “data mining” which has not only moved on from any technical sense it may have once held, but has also become pejorative in the process. The interesting question is whether we should regret or resist this. I personally think not. There is always the alternative of coining new usages to revalue the original meaning: let us say “predictive analytics”. As for the subject of this blog, shortly before Christmas, I was walking through a decision-tree model with a customer. I described the tree visualization. They said, “That’s not visualization, it’s just a diagram!” For them, “visualization” implied something much richer. “Like on CSI,” was the touchstone.

    For Business Intelligence there is a complicating factor. We coin many Business Intelligence terms by re-using (often by narrowing) terms that already carry broad common meanings. These vulgar meanings still resonate with customers. Every so often, I meet an executive who thinks a “data warehouse” is primarily an archival store, based on his understanding of a warehouse. The problem is not unique to information technology. My otolaryngologist once told me that I appeared to have suffered an “insult to my oesophagus” whereby I visualized something quite different from what he had in mind. Yet, when we use terms which still have vulgar resonance, we can hardly insist that our preferred usage supersedes the vulgar implications – the laryngologist may reasonably expect others in his profession to follow his meaning, but he had to accept that my somewhat amused confusion was justifiable.

    Within our narrow domain, I do believe it is useful for us to formalize terms in some contexts. For example, there is another controversy continuing here with Andy Bitterer of Gartner regarding the Data Integration Magic Quadrant. When it comes to inclusion criteria for a formal industry review, I heartily support a prescriptive terminology. However, I do not think that Gartner, or anyone else, can insist that only their definition of “Data integration” is acceptable more broadly.

    If we do move, as Neil rightly exhorts, to be a more human-centred industry, less “dominated by IT architectural diagrams, data and process flows” then we are no longer masters of our domain, or of our terminology. Breaking out Business Intelligence from its narrow, technical, sphere to be an expansive, humanist, activity is, to my mind, the greatest priority, and I am happy to accept that when we do so, our terminology slips from our control.

  8. Hello Ted,
    Coming a bit late to this party — here now thanks to your twitter follow — but I do want to say —
    I largely agree with Stephen Few and independently arrived at similar conclusions in my initial look at Lyza. LyzaSoft oversold their tool. See my blog article, Lyzasoft’s Non-Analytical Approach to Analytics at .
    I participated in subsequent discussions of definitions and now see that Lyza has moved away from leading with the “data analysis” and “data visualization” terms, which says that the company now also sees that its core strengths and capabilities are in other areas and that, for whatever reason, there are better identifications than continuing to promote Lyza with the DA & DV labels out front.

The data industry thrives on conversation. Please submit a comment.

Other recent posts

Bohemian Grove a la BI

The Bohemian Grove of the BI industry convenes for the fifteenth time in just three weeks. Naturally, you ask the obvious question: Are you serious? The Grove? A summit? The answer begins with a fond recollection of the Grove. If you’ve never attended the Bohemian Grove yourself — I haven’t, though I live in the… Continue Reading

Favorite Star Trek, a data story

This story shows how elemental data stories really are. Humans come ready to tell and hear them, requiring no plug-ins at all. This young person can do a good job of it. There was a question, followed by data, then questions and answers, and and finally a conclusion. It’s all there. It’s elementary. Sure, this… Continue Reading

Bad stories stop good data at the water cooler

We agree by now that data’s a good compass. One neglected question is tougher: Which map? Everyone’s known the kind of “grouchy guy” TDWI instructor Kellee M. Franklin, Ph.D tells about. This guy knew better than most of his co-workers about how their Washington, D.C. defense agency worked. And he was frustrated. Over the years,… Continue Reading