When it comes to Business Intelligence (BI) and Data warehousing (DW) there are always contradictory attributes arising within the industry. One among them is vendor selection. Whether to consider a consultant or go for an outsourcing vendor and if outsourcing, what can be outsourced to onshore, off-shore or near shore?
Most companies have used Data warehouse & Business Intelligence consultants at some point of time – but what about outsourcing? What exactly the companies’ interested in? What works or what does not works for BI and DW projects?
When it comes to Business Intelligence (BI) and Data warehousing (DW) there are always contradictory attributes arising within the industry. One among them is vendor selection. Whether to consider a consultant or go for an outsourcing vendor and if outsourcing, what can be outsourced to onshore, off-shore or near shore?
Most companies have used Data warehouse & Business Intelligence consultants at some point of time – but what about outsourcing? What exactly the companies’ interested in? What works or what does not works for BI and DW projects?
Such questions are certainly expected from various decision makers with respect to the vendor selection till they get a proper solution.
How to select a vendor
Well, the unique selling point of any outsourcing vendor is their customer testimony. The published customer testimonial will show you only an ideal picture of the vendor’s perfection and not the failed projects. Practically failed projects of a vendor are tough to be identified.
Ironically the selection criterion for a suitable vendor becomes very critical. So how do you pick the one that is right for you?
You need to gather answers to some fundamental questions involving availability, security, performance, and customer service. From these information, the best way is to follow high-level, yet restrictive, criteria and only then compare them on a feature-by-feature basis.
- Meeting your Business Requirements: Be sure to carefully consider your business requirements and go with a vendor that meets them. You may risk going with a smaller vendor, but you are more likely to get the BI deployment you want.
- Availability: It’s important to understand if the service is deployed on an underlying infrastructure that is backed up with a meaningful service level agreement. The level of availability you need depends on the criticality of the service to your business. Nevertheless you’re looking to move your business intelligence systems to the cloud, it’s likely very important for the cloud service to be highly available. However, for moving secondary application you might not require high availability of the cloud solution. Security: The biggest objection about cloud computing is security, taking corporate information outside the four walls. It should be notable that vendor’s may use the same server to host even your competitors data. Most vendors go about cloud security through the conventional means of various levels of encryption, firewalls, etc. So make sure where is the vendor server that hosts your data center and what is their business continuity during natural disasters? This will give you a satisfactory figure of how your vendor will be able to manage your data in a secured way.
- Performance: As the cloud services gets matured; companies are expecting performance metrics from the vendors to support the growing business critical systems that make their business run. They expect a better performance than when it was on their own data center. So test your vendors asking how does the cloud service provider define performance? What metrics do they use? How does their definition of performance relate to end user satisfaction?
- Customer Service: As a company you need to analyze the vendor yourself by
- How quickly does a vendor return your calls?
- How open are they in discussing what kind of technology they use or who their providers are and how often they have downtime?
- What servers they use and the software they run and their disaster recovery plans?
What can be outsourced?
As generic rule, you can outsource anything but what do you want to achieve with outsourcing?
Is it the lower costs? Better service to users? Or both.
When you look at the triple factor Quality, Time and Cost,
- Quality and Time ~ Cost Increases
- Cost and Quality ~ Time Increases
- Time and Cost ~ Quality increases
So quality, time and cost are invariably proportional to each other.
If you ask is it ‘the right choice’ for a company to offshore or even outsource it’s BI projects:
It depends on how much DW work you have, versus how much BI work. There are organizations which need to pull together over 6-8 sources of data, in large volumes (say over 10 Terabytes for 12 months data). I would say this is a DW environment, where a lot of integration is required. On the other hand, many organizations may need to pull BI from only 1-3 sources, or have very small data volumes (under 1 Terabyte). For the first type, having a DW is probably an essential pre-requisite for BI. For the second, it is not so much. The 2nd type of organization may even implement a BI tool directly on top of their operational source systems. My point is that it is easier to off-shore / out-source the DW work, but not so much the BI work. It’s important to keep the BI work on-shore as much as possible. I’d recommend a 70% onshore ratio for BI work. If you are building a fresh DW, it’s probably best to give responsibility to the same party for the BI, so that there is a single vendor to hold accountable. But even if you are going with an off-shore DW + BI implementation, insist that the BI team is predominantly onshore. But if you are basically doing a “BI only” implementation, it’ll be better to find a local vendor. BI only iterations tend to be smaller (can be about 1 to 3 months). But where a new DW needs to be set up or new sources added to an existing DW, then it would be a longer project.
Off-shoring:
Off shoring is other dimension in outsourcing, where based on understanding of BI program, it can be thought of what can be delivered from offshore without degrading the quality of deliveries, over all. In any case, your offshore supplier should play the role of a partner in successful delivery of BI program. It is important to recognize that Business Analysis, Requirements Gathering, design documentation and Test Plans are the key to delivering a successful BI or DW Project, and these tasks must be performed on the client site by people that not only understand the Technology to be implemented, but also have Subject Matter Expertise. Off-shoring works, But ideally part of the off-shore team (30%) should travel and remain on-site for the duration of the project. If it is going to be a long project then plan to rotate the on-site visiting team. This increases exposure of the off-shore team to the local business customers, and IT staff. It also ensures someone goes off-shore who has personal relationships and understanding of the customers IT environment.
Considerations during Outsourcing
- Involve your outsourcing supplier at the beginning of program and start involving them very closely with your BI program, that would set the stage where at later part of program you can think of portion of program or bundled deliveries to be outsourced.
- For those services where requirements are not clear, or change often, and performance of services are difficult to measure, like a nice complex DWH and BI development project, you may go for outsourcing.
- Outsourcing can be done with DWH/BI program, but one should not start it without proper due diligence and proper thought process in mind
- Outsourcing of any application environment is something to be considered very carefully and that is double true for BI/data warehousing as the volatility of the environment is much higher than for any operational OLTP system
- Information Security can be a road-block. You should be clear on 3 things (a) How sensitive Information Security is for your organization? (b) What concerns does your Information Security officer or organization have about the project? (c) How will your vendor address these concerns, and satisfy Information Security? If you have a dedicated Information Security officer, ask them what their concerns would be. If your source data contains information that needs to be securely managed (credit card numbers, health-care details, addresses) and is regulated, then you may want to setup a remote-desktop capturing solution to help the remote team access local systems, without being able to copy any data. Depending on your budget, you may also invest in a test-data management tool, or a data masking tool, to provide the off-shore team test data that is anonymized, but still useful. Bottom line, it can be managed, with relatively little overhead, but you need to ensure your Information Security team buys-into the controls in place.
What works for what?
Roughly an estimated BI Project life cycle would include
- Project Initiation and Analysis:
- Planning and Designing:
- Development/Testing:
- Go Live
- Support
Nevertheless it is time and again proved that outsourcing is successful provided if you take care certain business perspective in order to manage your offshore team. This question whether to could be now has no meaning.
Below is a description which could be a better plan for Outsourcing,
- Project Initiation and Analysis: Could be Outsourced but needs longer onsite visits for offshore business heads/experts to understand the environment to Analyze the existing business models of the industry
- Planning and Designing: Could be outsourced with short term onsite visits for offshore experts, this area is the technical and functional experts who needs to work together to provide an overall design and upcoming plan for the BI environment.
- Development/Testing/Support: This could be completely outsourced. The crucial part here is communication channel. The better the communication, the more service you receive from offshore. To build up the environment as per the agreed terms one need to have proper communication from time to time with offshore team, so that they could deliver the required development. Most of the time this area is misunderstood from the company’s perspective as they always feels that they have provided enough information. No one thinks it is a necessary to understand how much transition is being done within offshore team members, as they are the people who cater further services. However the company provided information at the beginning that does not really give full picture of what company needs. The co-ordination happens only when client gives a better time to their offshore team. There is always a possibility that few things are forgotten once after the analysis and few things are placed in the document after the analysis, and few things are not focused by company business consultants even before finalizing documents, even under 6-sigma there is a possibility for an error. So it significant that IT industry runs with communication, where there is less communication there is always a danger for disaster. So if this is handled with due care you are always successful in off-shored outsourced projects.
- Go Live: This once again needs Onsite Visit by offshore team for giving the feel of end deliverable.
Hope I could put the points, as I tried to use very few technical terms in order to present it in simple manner for better understanding about how offshore and outsource models would work.
So I conclude by saying outsourcing provide you follow certain steps as above or you could gauge according to the company needs and understanding. Depending on your organization’s maturity to off-shoring, the type of project activities can be done; it is possible to look at optimized off-shoring. You could also look at a staggered approach where learning of one release are incorporated into the next and progressive increase the level of off-shoring
To know more on ZSL’s BI services and solutions, Contact Shaughn Knight.
