Tue 14 Mar 2006
There are far too many conflicting and confusing definitions of Data Mart and Data Warehouse floating around. The long running debate between Ralph Kimball and Bill Inmon, the two Titans of Data Warehousing, only adds to the confusion.
In this post, we’ll try to get some sanity around the concepts, without getting drawn (hopefully) into the crossfire.
A Data Mart is a specific, subject oriented, repository of data designed to answer specific questions for a specific set of users. So an organization could have multiple data marts serving the needs of marketing, sales, operations, collections, etc. A data mart usually is organized as one dimensional model as a star-schema (OLAP cube) made of a fact table and multiple dimension tables.
In contrast, a Data Warehouse (DW) is a single organizational repository of enterprise wide data across many or all subject areas. The Data Warehouse is the authoritative repository of all the fact and dimension data (that is also available in the data marts) at an atomic level.
Unfortunately, this is where consensus begins to break down into chaos. There are two broad schools of thought lead by Kimball and Inmon that disagree on the details.
Kimball School:
Ralph Kimball began with the Data Mart as a dimensional model for departmental data and viewed the Data Warehouse as the enterprise wide collection of Data Marts. This is the bottom-up approach. You may begin with the Sales Data Mart, after sometime you put in place the Ops Data Mart, and so on an so forth. If you want you could have even more specific Data Marts serving specific questions like customer Churn. If you take care of consistency of metadata (making sure each departmental Data Mart calls an Apple an Apple) and connectivity, you have a Data Warehouse. So the Data Warehouse is really a virtual collection of Data Marts collected together on a Data Warehouse Bus, and in that sense the data flows from multiple Marts into the Warehouse.
Inmon School:
Inmon’s approach is the exact opposite and avoids the problem of metadata consistency by looking at the Enterprise Data Warehouse as a single repository that feeds subject oriented Data Marts. You still have your Sales, Marketing, Ops and Churn Data Marts containing atomic or aggregated information, but they are based on the Data Warehouse and are really subsets of the data contained therein. This is the top-down approach.
Kimball’s approach is easier to implement as you are dealing with smaller subject areas to begin with, but the end result often has meta data inconsistencies and can be a nightmare to integrate. Inmon’s approach, on the other hand does not defer the integration and consistency issues, but takes far longer to implement (which makes it easier for the project to fail). Also, in my experience, organizations that are just starting to do analytics usually do not have the patience or commitment required for Inmon’s approach.
Any BI initiative is extremely iterative in nature. Unless you are confident that you would still have the CEO’s buy-in and a budget one year down the line, it might be better to begin with a Data Mart (to start delivering, and to manage expectations) keeping the meta data consistency requirements in mind, and then scale towards the Data Warehouse.
If you are interested in knowing more about the Great Debate, you can read an article by Katherine Drewek called “Data Warehousing: Similarities and Differences of Inmon and Kimball”.
P.S.: This is what I think, and I’ll be glad to hear of your experiences and views, specially if you have been a part of an Inmon style implementation.
March 30th, 2006 at 6:40 am
I appreciate the post that you’ve made. Bill is a good friend of mine, as is Claudia Imhoff, yet I work as a Data Warehousing consultant for the industry, and as a leading speaker on the subject.
For the past 12 years I’ve been building a data modeling technique called the Data Vault Data Model, you can see more at: http://www.DanLinstedt.com
It’s a hybrid data modeling approach squarely between Bill and Dr. Kimball - In other words, I believe there is a balance, and a place for both to exist and co-exist. However, I will say that I feel neither Star Schema nor 3NF by themselves fit the bill to “be” or “Act” as a wholistic data warehouse across the enterprise.
I’d love to hear your comments on my approach after you’ve read through the material. The Data Vault is not new, has established customers, and is being taught as a viable data warehousing technique at TDWI.
Respectfully,
Dan Linstedt
My blog, which is Bill’s blog is at: http://www.B-eye-Network.com
March 30th, 2006 at 9:53 pm
Dan,
Thank you for your comment and for pointing me to your Data Vault concept. I appreciate your taking the effort to do this. I have been a reader of your blog for a few months now, and it is a privilege to have a conversation with you.
I am exploring the Data Vault white papers and find it to be quite an interesting concept. It would still take some time for the ideas to really sink into my brain and for me to appreciate the implications.
I agree with you that there must be a way for Bill and Kimball’s approaches to co-exist, specially because I have found myself agreeing to both. Personally I find this (possibly unnecessary) disagreement to be one of the biggest impediments to widespread adoption of Data Warehousing and BI in organizations. The lack of coherence at the core itself leads to numerous confusions, misinformation and misinterpretation at all levels which in turn contributes to the fact that most BI projects fail.
I noticed that Data Vault is trademark (must be necessary to prevent misuse and degeneration of the term) and is patent-pending (like DW2.0 concepts that you took a permission from Bill to discuss). I (and this blog’s readers) would be grateful if you could clarify on the following:
a. I guess I would need permission to post about Data Vault concept and to link some of the material on your sites.
b. Would it be illegal if I (or another consultant) were to use the Data Vault concept for a in-house/client implementation? Would we need to pay a royalty? Would just a permission do? Or can we go ahead and use the concepts?
c. Do you think I may have transgressed on any IPR while discussing Bill’s and Ralph’s approaches in this post? (The lack of articles on this extremely important subject makes me wonder…)
Sorry if my questions sound stupid and if you have already explained some of these on your websites (My lame excuse: I’ve just browsed them for a couple of hours).
Thanks again, and have a very nice day ahead.
sincere regards,
Nishith
April 4th, 2006 at 11:53 am
Update:
In a mail conversation Dan clarified that the Data Vault concepts are not being patented anymore and are in the public domain so that anyone “can read, teach, speak, use the Data Vault modeling concepts.” He would appreciate if people use the discussion board on his site for questions so that he can ensure that the vision stays straight. The entire series is available as a download from http://www.MyersHolum.com
Do note that “The Data Vault” is trademarked and not the phrase “Data Vault” as such. This is understandable in order to avoid dilution of the terms and concepts.
Dan would be teaching Data Vault modeling architecture at TDWI, San Diego in August 2006, and that may be a good opportunity to learn directly from him.
September 26th, 2006 at 11:04 am
Hi Nishith,
You have done a good job.You explained the complex definitions in a simple manner.Keep it up and keep writing.Thanks a lot.
November 23rd, 2006 at 11:16 pm
In continuation of similar discussion. I have seen people talking about Kimball much when it comes to data warehouse development project. But in really people prefer top down approach. Till date I have hardly seen any organization serious about their DWH initiative going ahead with bottom up / federated data mart approach.
Any comments?
January 20th, 2007 at 12:37 am
I think a balanced approach is required. Inmon’s approach is necessary to determine what the end goals are, however these usually mean measuring complex business processes that require multiple transformation steps, that are complex to model, and costly and lengthy to implement.
So instead of trying to bite off that project first, it’s more practical to design a roadmap that takes baby steps like the Kimball approach, implementing data marts that analyse a subset of the complex business proceess, but still deliver business benefit. These ‘base’ data marts should eventually form the building blocks of the complete business process and this allow the complex analysis. The additional advantage though is that by having the consituent base data marts it the data transformation are apparent and allow drill down to understand the source of the data and the transaction that make up the measures.
This approach delivers tangible results so the business sees benefit, building confidence in the data warehouse initative to allow it to deliver on the longer term goals of the more complex business analysis.
February 5th, 2007 at 6:37 pm
Please excuse my butting in to your discussion. I am a Researcher working on an Executive Search assignment for a UK Utilities provider, who are looking to recruit a Senior Data Manager. I would be interested in your thoughts as to suitable companies where I might be able to find such a person. They must have proven influencing skills, commercial accumen, along with knowledge of data modeling structures. I would be very interested to hear from anyone who might be suitable for the role themselves, or able to point me in the direction of suitable candidates. My email address is mail@warrenpartners.co.uk
February 15th, 2007 at 6:30 pm
I have just started doing research into Data warehouses/Data marts. There is a lot of information out there and it can be confusing. Thanks for clearing up some of that confusion. But I do have a question, and not sure if I asking it the right way, but on a transaction level, which is more advantageous, a Data Warehouse than a Data Mart?
February 16th, 2007 at 9:34 am
Hi Jim,
Ideally you should keep detailed transaction level data whether you are using a Data Warehouse (trying to capture enterprise-wide data as part of a single project) or a Data Mart (subject oriented and hence limited in scope).
Assuming that you would want to begin with a Data Mart initially, there could be performance issues in case you have huge volumes of transactional data in the data mart that you are analyzing/reporting on. In such a case it might make more sense to pre-summarize the transactions in another table to aid common reporting and analysis while keeping detailed data in the main table.
In either case, it is not recommended to throw away detailed transaction data to keep only the summaries. You never know what you may end up needing tomorrow.
I hope I have answered your question.
February 20th, 2007 at 8:32 pm
Nishith,
Thanks, you have.
March 28th, 2007 at 7:52 pm
Hi Nishitu,
I work for a big North American Bank currently re-architecturing its existing Data warehouse that consists of data marts using Kimball approach. We are now moving towards an Enterprise Data Warehouse. I find it very odd that after spending so much resources and efforts for the past eight years we are now moving towards Inmon approach.
The Data warehouse management team at the Bank justifies that the cost associated with maintaining Data Marts is very high. Another strong reason is given that there is no Drill-Across capability across the existing data marts.
I can understand that the cost for maintaining individual data marts may be high. However, the management has not been able to provide any numbers that maintaining EDW will be less costly.
I think our main reason for failing to drill-across our data marts is because we used we did not use the Conformed bus approach properly to integrate our data marts.
I am not an advocate of any particular approach. However, I am really bothered about the decisions that are made to justify this significant change. Your comments will be appreciated.
Regards,
Rehan Shaikh
March 31st, 2007 at 11:01 am
Hi Rehan,
You are right in having identified the main problem to be the lack of a conformed bus to integrate the data marts. Even though this is critical to the success of Kimball approach, it is sometimes skipped because of poor design (probably the implementors did’t understand the concept), execution challenges (difficulty in getting different departments to sit together for stitching the data marts together), or sheer laziness (if the client’s not asked for it, why bother).
I too am bothered by assertions that these problems would automatically disappear in the Inmon approach. It would still involve the same execution challenges and there is no guarantee that the ‘connecting’ information would be successfully identified and utilized in Data Warehouse design.
In short, I agree that building the Data Warehouse again from scratch is not going to solve any problems unless the root causes are fixed. I wouldn’t be surprised if some top shot consultants (working directly or indirectly with vendors) are involved in convincing your management.
Do have a look at the post The Sad State of Analytics Consulting.
Hope this helps.
April 1st, 2007 at 8:30 pm
It is possible that such a huge change would be pushed for by folks with a particular (political/economic) agenda, whether or not the data spanning multiple data marts was properly stitched together using the bus approach.
Federations, in the political sense, are usually in some sort of tension with a centralized authority. It’s about power, control, and, often, money, e.g., top salaries to executives/consultants who “unify the data into a centralized data base”. You know, the mid-level folks have already done the work w/ data marts, and a central figure/group takes the credit.
Often technical folks are blind-sided by politics, because they believe that the “technology speaks for itself”. The key, I believe is not to be jaded or cynical about this reality, but, like any system, understand that politics matter, and we must learn about and be aware of them as we move forward w/ our data mart/warehouse initiatives.
-sch
April 4th, 2007 at 6:39 pm
Completely agree with you Scott. You’ve hit the nail on the head.
April 6th, 2007 at 6:51 pm
quetion
compare and contrast database and dataware house
April 6th, 2007 at 6:53 pm
quetion
compare and contrast database and data warehouse
compare and contrast data mart and data warehouse
May 29th, 2007 at 7:26 pm
Would someone be kind enough to provide some answers to the following question?
Thank you in advance.
1) Data Marts… (do we create these virtually, physically etc.? when? guidelines)..
2) Where do we need to source data from for data marts? (thru warehouse only or ??)..
3) What about real time integration of data? (thru ODS? Warehouse? Mart? Virtually? Physically?)
4) How do web service or SOA come into the picture with respect to a data architecture?
5) What is a business intelligence vs. report arch? We need both…. Tools? etc.
June 8th, 2007 at 4:51 pm
Hi Cameron,
Some of your questions are quite broad and I’m not sure if I’ll be able to answer them well without an idea of your context and environment.
A persistent problem with analytics/data-mining/BI space has been the extreme proliferation of jargon and obfuscating terminology that means widely different things to different people and hence has lost meaning.
A good beginning point would be to first read the post Database vs. Data Warehouse that explains the difference between operational (OLTP) system databases and data marts/warehouses.
Once the data is extracted from source OLTP system is made available for analysis, a whole set of ’solution and process’ capabilities can be built to use this data for meeting the analysis, reporting, modeling, and decision support needs of the business. This is the (often unrealized) promise of “Business Intelligence”, and Reporting Architecture forms a small (though important) part of it.
The data mart concept is a conceptual “business analysis” design framework that then leads to the physical tables etc. that make up the ‘physical’ data mart through a process called “Dimensional Modeling”. For more details you can read about Dimensional Modeling, and then see it applied at the post Sales Data Mart - Dimensional Model for Retail.
Do note that Data Marts are subject specific, and you would need to create separate marts for different subject areas. As for whether the data from source OLTP systems can be directly loaded into Data Marts, or whether you first want to take it to a Data Warehouse (that then feeds Data Marts) is a difference between Kimball and Inmon approaches as outlined in the current post.
As for real time integration, SOA and web service, the answer depends a lot upon what you really want to do. Guess it’ll help if you can provide some more information.
Hope this helps.
Nishith
June 20th, 2007 at 12:14 pm
Hi Nishith
my question is in regards that what is better in data warehouse and data mart and why we keep them differ on which bases, and what are there advantages respectively, i m working in house and waiting for your reply
June 21st, 2007 at 5:03 pm
Hi Rao,
As explained above, a data mart is subject specific whereas a data warehouse is much larger in scope. Its more a matter of choice and taste (between the Kimball vs Inmon approaches explained above) whether you consider the warehouse to be a collection of marts, or whether you feed the marts from a warehouse.
In my view, for an organization just starting off with analytics, it is better to follow the Kimball approach (of incrementally creating marts with conformed dimensions to form the warehouse). But again, a lot depends upon the circumstances.
Maybe we can help if your question is more specific.
Cheers!
You’ll need
June 27th, 2007 at 8:58 pm
At last I got the difference between Data warehouse and Data mart
August 9th, 2007 at 1:36 pm
it’s nice and very easy for the beginners to understand the difference between data warehouse and data marts.
August 12th, 2007 at 2:53 pm
Shoukatullah, i’m completely with you
August 31st, 2007 at 4:55 am
Hi Nishith,
For a long time I am doubting if I am an 100% Kimballian or an 100% Inmonian. When I design a Data Warehouse solution it’s a little bit depending on the volume of data, if the central Data Warehouse would be a relational model or a dimensional model (which is in fact a relational model as well). I mostly store my base facts of the same subject area in one table (not looking at partioning). Than I will design a datamart which in my opinion is a tuned star schema for a specific group of users (department, fraud annalists etc) and with a limitation on history & granularity, but still all information for the datamart(s) is resourced from one and only one central Data Warehouse (’Single point of the true’). I will use conformed dimensions, so different datamarts shared these dimension which enables users to compare data between different subject areas. Furhtermore I often have multiple base fact tables in one Data Warehouse (database) schema sharing only the conformed dimensions with the other base fact tables, if appropiate.
Bill Inmon’s paradigm:
Data warehouse is one part of the overall business intelligence system. An enterprise has one data warehouse, and data marts source their information from the data warehouse. In the data warehouse, information is stored in 3rd
Ralph Kimball’s paradigm:
Data warehouse is the conglomerate of all data marts within the enterprise. Information is always stored in the dimensional model.
There is no right or wrong between these two ideas, as they represent different data warehousing philosophies. In reality, the data warehouse in most enterprises are closer to Ralph Kimball’s idea. This is because most data warehouses started out as a departmental effort, and hence they originated as a data mart. Only when more data marts are built later do they evolve into a data warehouse.
I consider myself as an combination of both Inmon: central Data Warehouse, with an optimised datamart for a specific group. Kimball, because I often model the Data Warehouse as more different schema’s in one database schema.
My question to you is, what does it makes me and is the wrong to combine both “religions”. I read a couple of articles and nobody could give me a satisfied answer. I hope you are able to discuss this further with me. Now Altis is focussing more on the Kimball approach I was furthermore wonder is I need to change me approach and way of advising for the future.
Kind regards,
R
August 31st, 2007 at 4:14 pm
Hi all,
I am new to Business Intelligence. I want to learn Data warehouse, Data Mining, B.I. BTW, I got the basic idea of D.W. and D.M.
As I told, I don’t know much about these. Can anybody provide me links or materials to clarify my concepts? I want very introductory materials. If anyone has, please forward me at chintu_the_best2003@yahoo.com
Thanks in advance,
Chintan.
September 13th, 2007 at 5:29 pm
Thanks to the authors and to the contributors for this dialog on what can be a very confusing topic. There is one particular question I would like to pose, and it relates to synchronization of data mart analysis results back into the centralized data warehouse. We wish to extract from a centralized DWH a series of information subsets that we publish to different data marts for use by analysts with different areas of interest. These analysts perform work on the content in those data marts, and this work yields information updates we wish to migrate back into the DWH. We also wish to ensure that any other mart deriving any of its content based on these updated records receives the updated information near real-time. Are there preferred techniques to accomplish this “feedback” flow to the DWH, and to accomplish this synchronization with other data marts? Thanks in advance for your reply.
September 22nd, 2007 at 9:49 pm
All of the questions, and the reasons, and the issues that organizations are faced with today is _why_ I created the Data Vault in the first place.
It’s real name is: “Common foundational warehouse architecture”
What’s happening in the industry is that businesses (like the bank above) are running in to the limitations on the Federated Star Schema, these modeling approaches are “breaking” under volume, and limit flexibility of I.T. - Keeping a lid on agility, and build out of solutions.
So, really - the “2nd generation” or “3rd generation” data warehouse efforts are looking at the Data Vault modeling architecture, and the whole approach.
The Data Vault (as I said before) is freely available, and is a Top Down architecture, bottom up implementation. I’ve recently added more information about the Business Impacts and reasons why a Data Vault would be helpful.
Raw level EDW’s are necessary for compliance, and auditability. It is hard to “adapt” a federated star design, and it gets harder and harder to KEEP conformed dimensions as “conformed” as the enterprise efforts grow and horizontal integration requirements get larger.
The Data Vault also manages this by focusing on the definition of business keys (the Hub Entity) - I really encourage people to investigate this approach, not just from my perspective but from the customers that I have interacting on the forums.
Real-time data is hard to “load” to Federated models for many different reasons, this problem is also solved by the Data Vault modeling architecture.
I’ll be posting additional information on the architecture on B-Eye network over the next couple months. Also, feel free to contact me directly with questions.
Cheers,
Dan Linstedt
DanL@DanLinstedt.com
September 25th, 2007 at 9:00 pm
Sorry folks for not being able to respond over the last few weeks due to a lot of travel with limited connectivity at times.
First of all, my sincere thanks to all the participants in this discussion. All of us gain from this interaction, and I hope this healthy debate continues.
R,
I myself have long believed that both the sides in this argument have merit, and a judicious combination of both paradigms is the ‘correct’ approach. If it is ‘wrong’ to combine the two ‘religions’, then I guess both of us are indulging in a lot of blasphemy here! ;o)
Having said that, ‘combining’ the two approaches leaves a lot to individual skill, interpretation, judgment, and integrity. One cannot assume that every hired consultant would spend the time and effort required to find the best solution (see earlier post on The Sad State of Analytics Consulting).
So I guess, the need is for professionals to try and standardize the key approaches in ‘marrying’ the two ‘religions’. It would be interesting, for example, to look at the ‘Data Vault’ approach proposed by Dan above.
Dan,
You are right in pointing out that it is tough to keep the conformed dimensions conformed as the business relaities (as well as the Data Warehousing teams/vendors) change over time. I’ll keenly follow your postings on the Data Vault approach. (I’ve wanted to review and post on the same for sometime now, but just havent been able to do it do far).
JimMc,
If I have not misunderstood your question, you want to move some of the analysis results back into the data warehouse for use by other analysts and operational systems. Depending upon what kind of analysis data is being sent back, and what kinds of operational processes are to use them, you could consider this as an operational data store and design accordingly. Ideally, I’d want to keep these results in tables separated from the main DW tables (you don’t want to be updating the data in main DWH tables, both from performance as well as audit/proofing point of view). You’d also need to store ‘Freshness’ of the analysis results (as analysis_date, say). If the analysis and operational processes are extremely complex and voluminous, you could probably adapt some ‘transactional clearing house’ ideas (treating the analysis results as just another data source) from Dratz’s paper on BI as an abstraction architecture.
October 1st, 2007 at 8:31 pm
Thank you very much for the replies. I will think further on this to be certain that I understand your suggestions, and I’ll also read the paper on abstraction architectures.
I’d like to get “smarter” on techniques that are applied in situations such as mine to coordinate changes to a record that may occur in different data marts to a given source record, especially as these updates are incorporated back into the DWH. I can think of many challenges in terms of record synchronization, problems related to the temporal nature of the data seen by analysts in other marts, etc. Managing this will be a challenge, and I’d like to learn more about unique ways others have tackled this in the past.
Thanks again to each of you for your ideas and feedback.
JimMc
December 10th, 2007 at 5:39 pm
The article on Dw Vs DM is good but,it will be very nice if you followed the differentation in a point wise manner i.e differentate their usage ,character,advantage etc.
regards
yadu
December 21st, 2007 at 4:24 pm
My experience with larger organisations (Banks, etc.) is that they adopt the Inmon approach - and then find it all too difficult, so divide the DW into domains with separate data owners. These domain areas have strong political power to go their own way, so after a while, you need a lot of smarts in the data extract into cross domain data marts to match the data.
I’ve seen millions of dollars spent on what becomes a little more than an off-line copy of operational data, rather than a proper data warehouse.
I don’t know know if the Kimball method will work any better, but it at least is designed to be based on separate data ownership for different domains. Recognising this, the governance issue can be addressed directly. Maybe people will play better if the rules are already set.
February 20th, 2008 at 3:01 am
I agree, but I just need to know which of the current DBMS (DataBase Manager Systems) is able to aply these technologies.
Please, I´m working on it and I need to know wich DBMS is the best choice to develop Datawarehousing or DataMarts.
I´ll be thankful for your answer.
Cheers, from Mexico.
February 20th, 2008 at 3:05 am
Please, answer to:
p_d_g86@yahoo.com.mx
February 20th, 2008 at 8:17 am
Hi Pedro,
Data Marts and Data Warehouses can be built on any RDBMS such as Oracle, DB2, MS-SQL, MySQL, etc. The difference is only in the way you design your database schema and tables - otherwise these are just regular databases. So you do not need to go looking for new software.
You could try reading my post Database vs. Data Warehouse to understand this better.
I’ve particularly found MySQL to be an excellent solution for large data volumes (read Data Warehousing with MySQL), but I’d suggest you stick to whatever RDBMS you and your team are competent on.
March 27th, 2008 at 1:54 am
Pedro,
If your biggest concern is which DBMS to choose, you have a lot of other knowledge you need to acquire.
There are many tools to accompany a DBMS in the Business Intelligence (BI) arena. There are hundreds of vendors and some great products out there. Generally DW/BI projects are hundreds of thousands of dollars (> 100K USD) and many months. Even using a Kimball approach, it will take some time compared to the Inmon approach that could take over a year to complete all the analysis and requirements.
If you are merely looking for a repository, MySQL would be just fine, but I would guess that you want to do more than store the data.
It seems that no one is 100% Inmon or Kimball, but in either case, you must still keep in mind that there is always a big picture to construct.
Good luck in your endeavors. You are headed in the right direction by finding great blogs like this one to help in the education.
– David
April 23rd, 2008 at 8:52 am
I’ve just about come full circle, and have gotten the endorsement from Bill Inmon last summer.
Bill Inmon says:
“The Data Vault is the optimal choice for modeling the EDW in the DW 2.0 framework.”, June 2007
We’ve gotten huge responses to our classes on the Data Vault, which we teach in Denver, Colorado at: http://www.GeneseeAcademy.com, and also in the Netherlands.
I’m nearing publication of the business chapters of my book, which will be announced soon on B-Eye-Network.com
I’d love to hear from anyone that is interested, and would be happy to go through a 30 minute webex to discuss the Data Vault approach. I’ve got some large customers using it, including: Office of Naval Intelligence, US Army Medical Division, FDA, FAA, JPMorgan Chase, Cendant Timeshare Resource Group, and a host of banks around the world.
I’d love to hear from you all,
DanL@DanLinstedt.com
September 1st, 2008 at 12:30 pm
hi nishith,
i read your content on difference between datamart and datawarehouse,i just wanted to know which would be the best option to choose for an organisation and why?kindly reply to me