News Flash: BI Discovers the Obvious

I was surprised to find the obvious reported as news today when I read Ted Cuzzillo’s latest article for TDWI’s newsletter BI This Week titled “The New Breed of BI Analyst.” I suppose I shouldn’t have been surprised, because vendors and other organizations that support the business intelligence industry sometimes claim to have discovered something new that has existed and has been obvious all along. The huge population of data analysts who are, according to Scott Davis of Lyzasoft (as quoted in the article), “determined, resourceful, and distrustful of data managers who presume to think for them” is probably the oldest and largest group of data analysts that’s been around since the advent of information technology (IT). Anyone who has ever worked in decision support (data warehousing, business intelligence, business performance management, etc.) has been aware of these analysts all along. Some have opposed them as “annoying users who make our lives difficult by working around our controls” (my paraphrase of a sentiment often expressed by folks in IT) and others have worked hard to support them by providing access to data and useful tools, but they all know that this is not a “new breed of BI analyst” that was recently discovered as the article suggests through “research by LyzaSoft” and work by “Microsoft’s Gemini program.”

When a vendor or a consultant has nothing new to offer but is desperate to win clients, it is tempting to claim a new discovery by shrouding the obvious and familiar in new terminology. In this case, however, no attempt has been made to disguise it; it’s just old news. Come on Lyzasoft and Microsoft, don’t insult the intelligence of the business intelligence community by gluing a carrot on the head of a goat and calling it a unicorn. That only works at carnivals for children and drunks.

The fact that two of the BI industry’s longtime leaders-Wayne Eckerson and David Wells-were cited in the article should not be taken as a confirmation of its claims. I respect their work and know firsthand that they don’t view the type of analyst that’s described in this article or the importance of their analytical needs as a new discovery. I also respect the work of the article’s author, Ted Cuzzillo, and can’t imagine why he considered this news. Has the BI industry really become this desperate for something worthwhile to write about? Are we allowing vendors to tell us what’s news rather than finding it on our own? Is any BI organization supporting the work of journalists like Cuzzillo in a way that makes it possible for them to really investigate what’s going on or are they supporting them just enough to pick the low-hanging fruit?

Take care,

15 Comments on “News Flash: BI Discovers the Obvious”


By Donald Farmer (Microsoft). April 12th, 2009 at 6:38 pm

Stephen, I suspect the key clause in Ted’s piece is “a kind of analyst that many know of but few have measured.” Taking the measure of these users is, for me the interesting change in approach to this class of analyst.

I do not disagree with you that experienced practitioners have been aware of these analysts all along. In that regard, I do agree that there’s no new discovery here.

However, I do find that less experienced practitioners, such as companies starting thier first BI projects, are often surprised to discover, some months after deploying their “one version of the truth” that these analysts are out there in the wild in their organization. At first, the approach tends to be, as you put it, opposition to these “annoying users.” Rarely, IT departments try to service them directly with data and applications: and I think that in many cases the IT departments do so only after trying in vain to reign in these characters.

In this regard, Ted is doing a service by drawing attention to this problem, especially on a site such as TDWI which, by its educational nature, attracts many who are newcomers to the field.

I think it is refreshing that vendors such as Microsoft, Lyza, Tableau and others are now seriously researching the needs of these analysts, and attempting to bridge the gap between them and those who find them “annoying.”

For those of us who have seen it all, the existence of these analysts may not be “newsworthy,” but I think it a useful service to TDWI readers to reflect on a trend in BI software to deliver new applications specifically geared to these users.

Donald Farmer

By Stephen Few. April 14th, 2009 at 10:41 am

Donald,

I’m thrilled when BI software companies actually become acquainted with the people whose needs they supposedly serve. I object, however, when particular vendors claim credit for discovering something that we’ve known all along. I also object when vendors demonstrate ignorance of their customers by developing and selling products that fail to address their needs, especially when those vendors claim to take their needs seriously. Of the three vendors that were mentioned in Ted’s article, only Tableau has a product that demonstrates knowledge of what data analysts do and a commitment to helping them do it.

I wholeheartedly agree that TDWI readers should reflect on the needs of people who struggle to make sense of data and then present what they find to others. As you know, my work is dedicated to raising such awareness and addressing the needs, including the work that I do with TDWI. I object when articles by TDWI or any organization suggest to their readers that the vendors who have failed to address the needs of data analysts (including LyzaSoft and Microsoft) are leading the charge in supporting them.

Steve

By Seth Grimes. April 14th, 2009 at 12:10 pm

The earliest that this distinction was usefully made, so far as I know, was by statistician John Tukey in defining exploratory data analysis as a contrast to confirmatory data analysis. A quick read of the Wikipedia entry on Tukey (http://en.wikipedia.org/wiki/John_Tukey) implies that this distinction dates to the ’70s although it surely has very deep roots.

Microsoft Gemini will focus on Excel as the user interface, won’t it? (It isn’t slated for availability until the first half of 2010 so a lot could change.) Excel has always, for better or worse, facilitated data exploration, has it not? The worse part is that Excel discourages collaboration and the exploratory trail — the transformational steps undertaken — is often not easily traced. Lyza supposedly overcomes the second of these problems although it appears to do nothing for collaboration and to have far fewer number crunching functions than Excel.

With all the Web 2.0 to-do, it had become obvious, I thought, that modern tools are supposed to support collaboration and traceability, and is it not also obvious that new tools should generally be more functionally rich than the ones they intend to replace? So perhaps BI has (re)discovered SOME what’s obvious and willfully ignored other, inconvenient (to some) bits of obviousness.

By Donald Farmer (MSFT). April 14th, 2009 at 1:39 pm

I suppose we’ll need to see whether Gemini meets the needs of analysts in 2010 or not. It certainly will be focussed around Excel, and it certainly will include features strongly geared to collaboration and understanding the “exploratory trail” - although in the latter case we’re providing oversight and control for IT rather than the detailed analytic lineage that Lyza supports.

The development gate is rarely so strait that new applications inherit all the good and discard all the bad of previous generations. Some features, like the QWERTY keyboard layout, persist by sheer persistence, even when provably inferior. I’m sure it will always be so.

Whether we do a good job of supporting analysts, especially those who are underserved today, is something I am happy for the analysts themselves to judge. I have no doubt that we, as vendors, are doing the right thing in reaching out to them. Also, from all I hear, data analysts find it refreshing that vendors are paying them this long overdue attention.

This may not be news to Stephen, but it *is* news to many struggling to solve business problems within enterprises today. We sometimes talk about “people who can’t spell BI” - meaning thereby no slur on their intelligence, only that they don’t even know such a praxis exists. For them, it’s all Excel. They are the Gemini audience.

By Stephen Few. April 14th, 2009 at 2:32 pm

Has thoughtful discussion among professional colleagues now been reduced to twitters? I don’t twitter, so I wasn’t tuned in when Donald Farmer of Microsoft and Claudia Imhoff of Intelligent Solutions exchanged comments about this blog post today. A reader kindly forwarded their remarks as follows:

Claudia mused re: Donald’s wondering if his comments about your comments on Ted’s Cuzzillo’s piece about “new” BI analyst types were a mistake:

Claudia: No — S Few’s rant was uncalled for — AGAIN. Your response was spot on. Not new but still a problem - yes!

to which Donald replied:

Donald: Thanks. I don’t know why Stephen’s on his high horse so much these days. Such a good blog when he’s on form.

Does Claudia actually disagree with my position? Although she calls my blog post a “rant” that was “uncalled for,” she didn’t explain why she feels this way or to what she objects. In the blog I stated that this supposed “new breed of BI analyst” that LyzaSoft claims to have discovered is not new. Claudia apparently agrees that they are not new. She implies that I don’t agree that the existence of this body of analysts who have never been well served by the BI industry is “still a problem.” I absolutely believe that this is still a problem. Most of my work addresses the needs of these analysts and bemoans the fact that they have always been underserved by BI. So what is the substance or even the gist of Claudia’s objection? I actually don’t know.

I object strongly to claims that aren’t true, such as the false claim that I exposed in this blog. When vendors make such claims, they hurt the BI industry and confuse the people who rely on it. Every thought leader in the industry should find false and misleading claims offensive. If I get my facts wrong in anything that I say or write, I expect to be corrected, not with vague allusions to the “high horse that I ride” but with substance.

Steve

By David Raab. April 14th, 2009 at 6:15 pm

AS it happens, I was briefed today by Scott Davis of Lyzasoft. My recollection is that Scott described his research as defining different types of analysts, but didn’t particularly claim these analysts hadn’t previously existed or that Lyzasoft had discovered them. Nor, now that I read it, does Ted Cuzzillo’s article say anybody claimed that these types of analysts are a new phenomenon.

As to Lyzasoft itself, I have some reservations about the product but am surprised by Seth’s comments. The system is designed specifically to deliver traceability, and seems to do so reasonably well (at least compared with alternative approaches, which are close to non-existent). Its calculation functions are Excel-style and seemed quite extensive: that’s not my favorite approach, but if Excel is the standard, then I think they meet it. As to collaboration, they are releasing a collaboration solution tomorrow that certainly addresses the issue, whether or not it is the ultimate solution. I do agree that the data visualization and interaction capabilities are very limited compared to something like Tableau, so at least that criticism makes sense to me. I’ll get around to writing a more detailed assessment in a day or two.

By Stephen Few. April 14th, 2009 at 10:23 pm

David,

I drew my conclusion that Lyzasoft considered these analysts a relatively new group from the title of the article: “The New Breed of BI Analyst.” Either this is how the information was presented to Ted Cuzzillo who wrote the article or he mistakenly read this into the claims that Scott Davis of Lyzasoft made about their research and its findings. If the latter is true, then I was wrong to fault Lyzasoft for fostering this misconception. Perhaps Ted can clear this up.

The fact remains, however, that nothing that Ted reported about Lyzasoft’s research findings is new information. To be fair, however, little of what’s reported as BI news is actually new and useful information. It’s often old worn out information recycled in a new wrapper.

The needs of data analysts have changed little in 30 years. One thing that has changed is that, due to data warehousing and business intelligence, more data is now available to them with less effort. Because few vendors in these industries have ever understood and addressed the needs of data analysts, we have vendors such as Lyzasoft that are now discovering what has actually been known for a very long time. If this really is new information to Scott Davis, I’m glad that he’s now aware of it and hope that he does something real and effective to address it. So far, his claims to provide “a powerful desktop analytics solution that enables analysts to synthesize, explore and visualize data, then to publish compelling presentations” [from a Lyzasoft press release) are hollow. Because so few vendors are actually providing useful tools for data analysts, I consider it harmful when a vendor that lacks a solution feeds off the hopes of people who desperately need one.

Steve

By Stephen Few. April 15th, 2009 at 12:47 pm

Donald,

Data analysts will indeed be thrilled that vendors are reaching out to them, but only if vendors are doing so in ways that actually serve their analytical needs. The proof is in the product. Although Project Gemini is moving in the right direction, even if it were released today rather than in 2010, it would still be years behind the simple and useful analytical tools that vendors such as Tableau and Spotfire already provide.

Although functionality that supports collaboration and a history of the analytical process are useful, these are not the primary features that products such as Excel and Lyza lack. It is a mistake to focus on features that are currently gathering buzz when you haven’t provided the basic analytical interactions and visual displays that support the bulk of analysis. It is the habit of most BI vendors to focus on features that are getting media attention because of their marketing appeal, resulting in a slapped-together implementation of those features, rather than doing the hard but much more important work of addressing the fundamentals.

Steve

By Ted Cuzzillo. April 16th, 2009 at 9:48 am

Steve, you wrote in your reply to David Raab, “Either this is how the information was presented to Ted Cuzzillo who wrote the article or he mistakenly read this into the claims that Scott Davis of Lyzasoft made about their research and its findings.”

Actually, neither. Here’s the whole lede: “Two makers of data-analysis tools have taken a hard look at a kind of analyst that many know of but few have measured. Conclusions of the two generally agree.” That is, the news is the research.

Nowhere in the text is there any mention of a “new breed” nor does it imply anything like it. In fact, Wayne Eckerson and Elissa Fink indicate an awareness of such analysts that began before the LyzaSoft and Microsoft research.

Generally, it’s best to base discussions on what a story actually says, not the headline.

Ted

By Stephen Few. April 16th, 2009 at 12:21 pm

Ted,

So the title of the article was wrong? I did read the entire article, but I started with the title, which provided the framework for interpreting everything in the article. The title of an article should reflect its content. Was this a case of an editor coming along and slapping a title on your article that misrepresented its content?

Setting aside the issue of whether these analyst are a “new breed”, which the title claimed, Lyzasoft’s research has revealed what anyone who deals with data analysts already knows. There’s little value in confirming the obvious, except to those who don’t already know the obvious. So the real news reported in this story would have been better headlined as “Lyzasoft finally discovers what any vendor in the business intelligence industry should have already known for years”, but I suppose this is a bit too long.

Steve

By Ted Cuzzillo. April 16th, 2009 at 3:19 pm

News flash: Something about LyzaSoft just bugs the hell out of Stephen Few.

By Stephen Few. April 17th, 2009 at 8:35 am

Ted,

Again you’re reporting something that isn’t news. I’ve described my problem with Lyzasoft very clearly in the past. I have a problem with any BI vendor that makes misleading claims about its product’s abilities. Despite the vendor’s claims, Lyza is not a data analysis and visualization product. It is to visual analysis what a broken-down tricycle is to the Tour de France. And yet, Lyzasoft has received a great deal of attention in the BI press. This is a result of the company’s marketing efforts and relationships to industry leaders, not its merits–certainly not in the realm of data analysis and visualization.

By Natasha Lloyd. April 30th, 2009 at 9:16 am

[Full disclosure: I work for SAP BusinessObjects.]

Stephen,

From a user experience perspective, the info published in the article is very useful. It summarizes findings from the user research performed by two makers of data analysis tools. Besides the title (which I agree is misleading), I don’t see them claiming that anything they’ve found is “new”. It’s more a matter of sharing their findings with others in the same industry.

“There’s little value in confirming the obvious, except to those who don’t already know the obvious.”

What is obvious to you is not obvious to everyone. No one starts out an expert in data analysis. It takes many years to accumulate the amount of expertise that you have on the subject. Those who take the time to become experts are not the same people who will write software. They are the analysts - the target audience for the software.

It’s clear that people producing data analysis software should understand analysts and the process of data analysis. And one of the ways this is done is by performing user research. This article publishes the results of two such studies. Yes, the findings are not surprising (they rarely are in user research), but they do organize and summarize the information for people who are trying to learn about their users.

It seems like your main gripe with the BI industry is that they don’t understand the needs of their users. I think research like that described by Ted is one of the ways BI vendors are trying to change that. I’m surprised to see that you are so against it.

- Natasha -

By Stephen Few. April 30th, 2009 at 9:54 am

Natasha,

Don’t you think it’s sad that “leaders” in the business intelligence industry are claiming to have discovered through research what has been obvious to those who work with data all along? I don’t expect everyone to already know this stuff, but I do expect companies that claim a deep knowledge of business intelligence–LyzaSoft, Microsoft, SAP Business Objects, and others–to know it.

The opportunities for real research to discover what we don’t already know about data analysis and ways to support it are unlimited. Why isn’t this kind of research being done by the vendors? (By the way, exceptional work of this type is being done at Microsoft Research, but for some odd reason, Microsoft’s products rarely reflect this fact.) We who work to make sense of data and use it in meaningful ways expect those who make the tools that we use to understand what we do and what we need. We expect this, but we rarely get it.

When you refer to the “studies” that LyzaSoft and folks in the Project Gemini group at Microsoft conducted, you’re perhaps imagining something that wasn’t actually done. Did they actually do anything that people involved in software usability, human factors, cognitive science, or information visualization would actually recognize as research studies? If they did, I would love to see the research papers that were produced. The anecdotes about research findings that were reported to Ted reveal no evidence of actual research. Vendors tend to use the term “research” rather loosely.
Have you seen results produced by these so-called studies that “organize and summarize the information for people who are trying to learn about their users”? I would applaud such research. What I will not applaud, however, are claims without substance to research that furthers our understanding of data analysis. This is not research; this is marketing.

Steve

By Dan Flanery. May 8th, 2009 at 10:20 am

By the way, exceptional work of this type is being done at Microsoft Research, but for some odd reason, Microsoft’s products rarely reflect this fact.

This happens at many large companies - great R&D work gets lost in the shuffle, or buried under the weight (and marketing hype) surrounding other (usually pre-existing) products. Remember, it was Xerox who invented the modern personal computer as we now know it, starting in the mid-1970’s. They were never able to exploit their creation, for a variety of reasons, most of them internal to Xerox. It took another decade for Apple to successfully popularize those Xerox innovations, like the Graphical User Interface, the mouse, and the concept of desktop publishing.

Microsoft has recently started what looks to be a long and painful decline, as the PC market implodes (at least in dollar terms) and computing shifts to other devices (like cell phones) where Microsoft’s established Windows monopoly does them little good. In the enterprise space they have to contend with competitors like Oracle, and partners like IBM and HP who would just as soon not have to split their customer’s money with Redmond. As MS continues to shed jobs over the next decade, look for many of the innovations they’ve developed, but have been unwilling or unable to exploit, enter the market via other entities.