Thursday, September 20, 2007

They Myth of BI Standardization

BI standardization is a topic that vendors have been pushing on their customers for years. There are no shortage of press releases, case studies, and customer references that talk about X company "standardizing" on "so and so's" BI platform. And from a vendor perspective, it's a good course to take. Everyone wants to become the standard. It's the corner square, the top of the hill, and and defensible position whereby you get to dictate the other technology that your product interacts with in the client IT environment. The nirvana catch-phrase of "we can't use them, they're not the standard" is the dream of every vendor who attains such lofty status, and the big 7-figure deals that BI and EPM companies tout in their quarterly calls are often predicated on the issue of "standardization."

The only problem is that it's not true.

At least not in the way that the vendors would have you believe it to be true. And we all know it. First, just look at the customer logs of all the big vendors. If company X was the standard, with 85% of the Fortune 500, wouldn't it stand that company Y could only say they had 15% of the same list at most? Why then is it that every vendor counts 7 out of the Top 10 pharmaceutical companies, the top 30 retailers, the largest of the global 1000--whatever your measure--amongst their customer base? It's because there is no standard. It's all about departmentalization.

Truth be told (I've always wanted to type that phrase), there are actually very few mainstream technology "standards" in companies today. Microsoft Office is one--or more broadly--Windows. But even there the free-apps crowd is starting to encroach in a few organizations. Still, let's give Redmond that one. ERP as a standard gets muddled when you've standardized, say on SAP, only to acquire a company running Lawson that's so customized that your SAP system can't do the things this other system does, so you keep that system and interface it to SAP. Is SAP still the standard? Maybe. But I think you get the point. Bigger companies have multiple ERP systems in play. Sure, SAP may be the "standard," but they're paying maintenance to three other vendors as well. The pure-play BI vendors for years have feasted on "heterogeneous" environments, data sources, etc. as the arbiter of all data. "Sure, you can have as many data sources as you want," they'll say, "but you need one BI standard if you REALLY want to get the benefit from business intelligence or performance management. So use us."

But do we actually need a BI standard?

The oft unspoken truth is that the bulk of the deals done by BI vendors are at the individual, or even departmental level, and they'll likely stay that way. Even the "global accounts" teams in these companies are usually in with only part of the account, or at most a few business units or geographies. GE for example, owns every product under the sun in their organization, and yet multiple vendors tout "standardization" by GE on their products left and right.

And we know why they all do it--it's for credibility. If you can say that "GE" has standardized on your products (vs. the competition), and people look at GE as a company they'd like to emulate, then that may be worth something in a sales cycle. There's just one problem--it's not true, and more importantly, it shouldn't be.

Here's why. Just like the iconic Apple 1984 ads that have evolved over time to encourage us all to "think different," the needs and uses and sources for information that people use to make decisions and solve business problems are not and cannot hope to be addressed by one tool or set of applications. There are too many use cases, too many data sources, too many new ways to use and share and analyze the myriad of information that bombards a typical worker on a daily basis. Do we really think that a typical project manager is going to use just structured report data to address all the issues they have in front of them? Are we to expect that everyone will access the same universe and metadata to query the database and get the answers to the questions in front of them? It's just not going to happen.

And why is that? Well, usability for one. Say I have to put together an analysis of the potential revenue associated with a new product launch. Now accessing my "standardized" BI tool for historical reports and forecasts is one place I'd go for information. But what about information that's not in that system? Past launch plans. Ad hoc analysis on a spreadsheet. Third party research data. Am I going to use the same BI tool to get this information? Probably not. First, the tool doesn't support getting me that kind of data. And second, even if it could, it would have to be a highly customized pre-set query that would let me get just what I need (not to mention helli-smart to know where to get it within the vast wasteland of both my hard drive and the company network).

So what do I to do get all the information I need? I use not just the reports from my BI system, but also things like spreadsheets, business process tools, files on the share drive, IDC or AC Nielsen data--whatever I need to accumulate enough information to put the plan together and send it around for review.

And all these tools also comprise my business intelligence environment.

As a manager, I WANT my people using all the available data and tools at their disposal to help make the right decision. The thing is, is that all the available data is not just in the tool we've "standardized" on, it's in a lot of places, and I'm likely to use not just a report I build from the BI system, but a lot of different BI tools in order to come to my conclusions. That's hardly standardization, and yet, that's the reality of business today. The goal of "business intelligence" is to help people make better decisions with the information at hand. But there's no one "standard" for how that should be done. There are lots of them. And they're going to be different based on how I work individually, how my team works, our industry, our company size, our technology model--all of which could be wholly different than the person and the team on the floor above me.

So when you see the next press release come out, the next huge client that's "standardized" on one product or another, keep in mind that they've done no such thing. They may be "officially" declaring that they'll use only that BI tool vs. any others, but there will always be a multitude of tools and applications at their disposal to ensure that their people solve the problem in the fastest way possible. That's when you see business intelligence achieving its promise. And as our close friend Martha Stewart* would no doubt say, "that's a good thing."

*This statement is a lie. We don't know Ms. Stewart and hope she doesn't sue us for stealing her catchphrase.

3 comments:

Timo Elliott said...

Guy,

Let's face it, you know you're (mostly) attacking a straw man.

Having IT standards is not exactly uncommon in today's organizations, and it would be strange if BI were any different.

The only useful definition of BI standardization is "pragmatically implementing BI standards to reduce overlapping tools, lower costs, and maximize the benefits of BI" -- we may never reach nirvana, but it's hard to argue that stepping back and accepting information chaos is a better answer...

Finally, I agree that if business intelligence is to mean anything in the future, it has to step up and cover a broader information set than it has done up until now.

Link to my post on "Is BI Standardization a Myth?"

Regards,
Timo

Darren said...

Timo, as I commented on your blog, I think you've done a fantastic job of helping large companies standardize and I think it's been a very useful mandate for the IT departments' of enterprise companies seeking independence from the monolithic suites of Oracle and SAP. But what do you make of on-demand applications like salesforce.com that have primarily taken hold in companies where the lines of business have refused to wait for the bureaucracy of a standards committee to bless their use? What do you make of "shadow BI" that takes root under the radar as IT struggles to migrate between versions or manage the multple metadata layers of even one vendors' on-premise BI suite (for example)? Also, for what size company is BI standardization even a relevent conversation in the first place?

In the meantime, I tend to agree with Neil Raden, who points out in this whitepaper - http://www.hiredbrains.com/sfdc.pdf that "Standardization benefits accrue mostly to IT and to vendors, not to the wider audience of BI whose empowerment should be the enterprise IT mission."

Guy said...

Timo--actually it wasn't a strawman at all--the thought came from a Cognos press release (http://cognos.com/news/releases/2007/0820-2.html) the highlighted a standardization win.

My point is primarily with the narrow definition of "BI" as it's been traditionally defined. In the latest entry, I point to BOBJ's own moves with Inxight into new areas that the structured BI tools are just not able to handle.

BI is out of the genie's bottle--it's in the Universe, but it's on-line, on your hard drive, in a word doc, a spreadsheet, in XML--and there's no one place to go these days that people are going to use.

Great conversation