Big Data – Choosing a consulting vendor for your pilot projects
- Jan 02, 2013
- By admin
- In Outsourcing
- Share on
2013 could be the year when even mid-market firms have Peta bytes of data and their relational database systems start to fail them on real-time decision making. If you are the CIO of one such organization and is planning to do a pilot on Big Data, this article is for you.
Your big-data landscape shall have a way for collection of data, a place/appliance to feed the data to (Hadoop?) and a method to piece the data together to arrive at patterns (Map Reduce?), followed by typical POS, CRM, BI tools that work on the (now available) user profile, predicted behavior to deliver contextual response. (A custom digital merchandise served to your mobile phone when you are watching a movie, is a good example.). For the purpose of this article, lets assume that you have identified a DB vendor and the infrastructure vendor who would help make your enterprise ready for Big Data analysis and decision making.
Some tips to choose a big data consulting / SI firm:
0. You’d quickly come to a realization that there will not be one vendor who plays across hardware, database, design, and modelling/analytics. [tweetherder]Be prepared to engage multiple vendors or a single anchor vendor with a set of niche vendors working with them.[/tweetherder]
1. See if you could work with a product vendor who can do professional services and your pilot could add value to their product direction. You get qualified expertise this way. A better method yet, would be to[tweetherder] have the SI do a pilot for you without you having to invest in the hardware, storage, DB[/tweetherder].
2. If a product vendor comes forward to do a pilot for you, you would get a few employees trained with them. They could then be in a better position to pick niche operators for your next series of pilots and would thereby make it easy for the business to see incremental benefits in adopting this new way of organizing/mining data.
3.[tweetherder] Review the security policies of the vendor, as you are giving access to your company’s intelligence and often sensitive customer data (even if its not a production project)[/tweetherder]
4. See if your vendor has the ability to move the data interpretation/modelling part to your premises, because for these to happen, they need to be close to business. Distance kills here.
5. Take one problem defined by your business which has a lot of legacy data, where RoI can be demonstrated within 10 weeks into the pilot. Make your vendor sign up for metrics – weeks, expected insights and what that translates in terms of business value.
ContractIQ helps mid-market firms find big data analytics vendors that have successfully executed several pilots and strategic initiatives for its customers. Leave us a note if you are planning for your big data pilot this year!