JAN 29, 2010 5:00am ET

Related Links

USAA Uses IBM's Watson to Help Members Adjust to Civilian Life
July 28, 2014
Behind The Rise of the Chief Data Officer
July 24, 2014
IBM Introduces Watson to Consumers in Service for USAA Clients
July 23, 2014

Web Seminars

How Intelligent Digital Self-Service with Customer Analytics Can Lower Costs and Raise Revenue
Available On Demand
Improve Omni-channel Shopping Experience with Product Information Management
August 21, 2014

Pros and Cons of Using a Vendor Provided Analytical Data Model in Your BI Implementation

Print
Reprints
Email

The following question comes from many of our clients: what are some of the advantages and risks of implementing a vendor provided analytical logical data model at the start of any Business Intelligence, Data Warehousing or other Information Management initiatives? Some quick thoughts on pros and cons:

Get access to this article and thousands more...

All Information Management articles are archived after 7 days. REGISTER NOW for unlimited access to all recently archived articles, as well as thousands of searchable stories. Registered Members also gain access to:

  • Full access to information-management.com including all searchable archived content
  • Exclusive E-Newsletters delivering the latest headlines to your inbox
  • Access to White Papers, Web Seminars, and Blog Discussions
  • Discounts to upcoming conferences & events
  • Uninterrupted access to all sponsored content, and MORE!

Already Registered?

Advertisement

Comments (20)
Right on with the pros.. If you look at industry models for an EDW you have a much better chance of finding a better fit (once again a good starting point only) as opposed to getting "analytical" models which are typical subject oriented schemas for datamarts. This is certainly not as plug for Sybase but their IWS datawarehouse solution does have industry logical/physical models. The EDW model does have transaction/archival/aggregation layers modeled right into it.

Most folks combine how information is sourced/integrated and how information is used into a single EDW model. Add on top of that the challenge to keep everything "start schema"ish. The uses/use-cases of information are many and continue to evolve in any business. If we look at the Bill's corporate information factory, the EDW is an integration point for the enterprise and the datamarts represent various applications or uses of that information. Even when an organization does have the talent to develop an enterprise model, getting business to define things accurately and consistently might be a challenge. This is a problem no packaged or home grown model can solve. Bringing consultants in to facilitate might make the exercise a little more efficient at best. At the end of the day the organization has to appreciate the "value" of information or the "cost" of not doing so and business/IT leadership need to be articulate this well and press for neccesary change.

Posted by Kary K | Friday, January 29 2010 at 5:55PM ET
Boris

As usual, you are commenting on something very relevant and current. We get this question a lot and I hoping I can point my prospects to your Blog post.

I agree with all the Pros.

Regarding the cons:

I agree with all the Cons, but have questions/comments regarding the two below.

1. Competitive Advantage: I believe the competitive advantage derives from proprietary business processes. In this case, a packaged analytic model will not or at least "should not" support the proprietary business process. If it did, there was only an illusion of competitive advantage. There should be not fit. Typically vendors will pick business processes that are common to build a packaged analytic model, so that they can monetize he engg. investments and stay away from highly custom processes as it relates to packaging.

2. "Agile BI": Given that the model is ready, I would assume it would save time and deliver a faster deployment. Confused about this one, hopefully you can elaborate, what you meant.

Posted by Ajay D | Tuesday, February 02 2010 at 12:57PM ET
In addition, business processes need to be aligned with metrics in order to be effective. In many cases, the business processes you have will need to change to align with the metrics in order to be able to utilize an analytical application, rather than having metrics defined specifically to align with your business processes if you have a custom system. Not changing business processes may significantly reduce the relevance of the metrics, while changing them introduces organizational risks. Either way, you have a potentially significant obstacle to the success of the project.
Posted by Ralph M | Tuesday, February 02 2010 at 1:19PM ET
Boris,

I agree with your post. Unfortunately some analytical vendors focus so much on their analytical model that they lose touch with the customer business needs as well as use and integration issues. The vendor cannot adequately address these issues and the result is a troublesome and isolated anayltical tool that starts the countdown to its own obsolescence before it is even up and running.

Posted by Adam G | Tuesday, February 02 2010 at 2:04PM ET
I agree with most of your analysis - I feel stronger about the con's than the pro's. I believe utilizing an industry model does make sense for mid-size companies or for industries where the a large part of core information domains are either highly regulated or based on industry standards - Telecom comes to mind. In Industries where information models differ highly between competitors or in most Fortune 500 companies acquiring an industry model as a set standard will most likely create more headache than benefit. Here are a few dealbreakers I have seen before:

.) Large companies are having a hard time agreeing on data concepts to begin with. The vendor model is yet another player with a different opinion on what a 'Customer' or a 'Product' is .) As you mentioned above large players have competitive advantage in their data model. I know of a credit card issuer that collects in excess of 1400 data elements for every swipe of a credit card .) The need for customization will outpace the benefit and create unwieldy models

I think the main benefit for large organizations could be to get a head start, see concepts they may not have thought of and provide general modeling guidelines that the organization may want to adopt.

More on gaining competitive advantage from information at http://www.theinformationadvantage.com

Posted by Oliver H | Tuesday, February 02 2010 at 3:51PM ET
It was great for me to read this article and comments. Many companies seem to be going down the path of a "data-mart-in-a-box" solutions (i.e., SAP BW cubes, and Business Objects RapidMarts) that result in less than 100% of the business requirements. In some cases, this results in a penalty to the implementation team when the solution doesn't fully meet business expectations. Prepackaed/designed solutions may be a good starting point, but the practice that has proved to work for me is the top down performance management approach. Nothing works better than delivering a solution that aligns with company goals that are measurable.
Posted by Steven P | Tuesday, February 02 2010 at 3:53PM ET
Dear Boris,

I think no standard model/approach for modeling and designing the DW systems adopted by vendors. Each vendor have their own methods and tools. However, building DW from scratch and by your own is not smart, but helps from experiences vendors are crucial important in order our DW systems developed successfully. So, my opinion the users need well understand what is BI and DW all about, and discuss with vendor what approach is suit for your requirements ... and get commitment from the management to implement it. Indeed, analysis requirements from the perspectives of goal and decision maker are priority to ensure the value of information is well pressing.

Posted by Azman T | Tuesday, February 02 2010 at 8:43PM ET
In general, vendor-provided frameworks in very well-defined problem domains can work well. Frameworks for anything with any lack of rigor and standardization fare much less well. BI frameworks are much more likely to fall on the less-well rather than on the more-well defined end of the spectrum. Enterprise ERP and CRM experience should provide a healthy dose of skepticism for anyone considering adopting a pre-constructed model. Any start-up advantages offered by a framework are likely to be squandered in the struggles to make adapt it to the organization's real information models. Unwinding and rewinding an established model to fit can be taxing to experienced professionals, and overwhelming for those less experienced. It's usually a much better idea to have some experienced professionals develop a high quality model in the first place, implementing the essential model without creating cruft that will obfuscate things going forward.
Posted by CHRIS G | Tuesday, February 02 2010 at 8:59PM ET
What about the cost of "Building your own Model" Vs. "customizing a Vendor Model". I have seen customers who have adopted Vendor Model and doing well, but yes they are really worried about future changes they have to incorporate. Some customer tried using Vendor model and ended up customizing 80% of it, but worth doing as they had something to get started. I am not sure about the "Cost" factor in long run, any comment on the cost effectiveness. In my experience my preference will be to deliver the first set of KPI using a vendor model(if its freely available) and then build my own as companhy realize the benefit of DW/BI.
Posted by anup k | Wednesday, February 03 2010 at 5:29AM ET
Hi Boris

well the pro's cvan do you some good at the start but in the end it is a hell of job to adapt such models to your real business world. And in the long run you have the same problems when you need to change your business information model due to changing business conditions compared to building a model from scratch. In my experience it is much better to adopt modeling techniques like eg DataVault because it will give you means to reduce the length of a development/change cycle. And it is still better to use Kalido (and i'm not paid by them; it is just based on experience)as this tool use a generic datamodel "under the motorhood" and (re)generates you information model. Just a pity that there are no more suppliers offering similar functionality. Flexibility,the capacity to adapt and minimal leadtimes are crucial in building and maintaining a model.

Posted by ad s | Wednesday, February 03 2010 at 5:38AM ET
Interesting comments. Do you wish to meet 100% of business needs, or will 80% do? Time spent getting it perfect may be the difference between satisfaction and no job - the company loses out through lack of ability to react (solution not ready). Also, regarding a "con" issue, differentiation comes from action taken upon analysis, rarely upon the advantage of having perfect information - since as soon as you have your 100% fit, the business realises it needs more data!
Posted by Martin W | Wednesday, February 03 2010 at 6:59AM ET
The challenge as always in creating information systems that can deliver tangible value and benefits to an organisation is how closely the business and IT functions work together to provide both business and departmental strategies and objectives, but also a true reflection of the data's 'fit for purpose' state. Having delivered solutions for some of the largest grocery retailers in this space both using in-house designed models and vendor purchased models the question that I always get asked is how do I choose what is right for me? We have found 3 key questions help to point you in the right direction: -

1. How many business entities or separate P&L's are to be users and contributors (data and intellectual) to the solution? - Parent organisations delivering a solution for lots of smaller 'independently' run businesses will find that their data won't integrate easily into vendor models in key subject areas unless the business is prepared to be quite brutal in making decisions 2. How well are the business requirements defined to enable design - If the delivery team is saying there not sure how to model based on the requirements purchasing models really help kick-start the project 3. Where is the business value of the vendor model - often vendor offerings will provide both the warehouse and associated marts, if a majority of the value is in the marts or analytical tables it is critical the warehouse model is not customized too much as that customization will need to be replicated throughout the solution

Posted by Chris C | Thursday, February 04 2010 at 6:45AM ET
I agree with all the cons. The other alternative of building your own data warehouse environment or BI environment can be expensive and time consuming.

Rather than building expensive warehouse or even reporting data repository - leaverage the existing investment by data unification. Cheaper, easier solution to use the existing data and make it available to the application teams. Check out www.linkwex.com for more details on data unification.

Posted by Amit H | Thursday, February 25 2010 at 6:56PM ET
Hi, Thanks for sharing this post with us. It's really an amazing post. Keep posting the good work in future too. Small Business Loans
Posted by Peggy B | Friday, April 01 2011 at 5:42AM ET
The blog was absolutely fantastic! I like to keep on visiting this site once again thanks for sharing the info. Social Network
Posted by julie l | Tuesday, April 05 2011 at 10:05PM ET
I admire the valuable information you offer in your articles. I will bookmark your site and have my friends check up here often. I am quite sure they will learn lots of new stuff here than anybody else! Newport Beach Houses| Komodo Island New 7 Wonders| Beach Houses Newport
Posted by doel k | Saturday, April 09 2011 at 4:26AM ET
I admire the valuable information you offer in your articles. I will bookmark your site and have my friends check up here often. I am quite sure they will learn lots of new stuff here than anybody else! Newport Beach Houses| Komodo Island New 7 Wonders| Beach Houses Newport
Posted by doel k | Saturday, April 09 2011 at 4:27AM ET
Add Your Comments:
You must be registered to post a comment.
Not Registered?
You must be registered to post a comment. Click here to register.
Already registered? Log in here
Please note you must now log in with your email address and password.
Twitter
Facebook
LinkedIn
Login  |  My Account  |  White Papers  |  Web Seminars  |  Events |  Newsletters |  eBooks
FOLLOW US
Please note you must now log in with your email address and password.