1 – Completely agree with the alignment with business goals, which emphasizes the importance of collaboration between IT, analysts and the business community.
2 – Agility requires the ability to manage complex and numerous analytical models that rely on lots of variables and massive data volumes. Not only in the creation of the initial model, but the ability to modify models as business factors change.
3 – Good point about right time, because it certainly helps to load balance processing and analytics work based on the response needs based on moving everything to real-time.
4 – Data quality is certainly key as is the ability to derive a trusted single view of critical master data, but it’s also interesting to consider analytics as part of the cleansing or discovery process, vs. just thinking about data quality as an enabler for analytics.
5 – The prototyping recommendation is in-line with the reality of the analytics lifecycle – that is creating a sandbox environment that allows for the development of multiple models using an iterative approach.
6 – It’s important to realize that one size does not fit all – depending on the situation, it may make sense to analyze the total dataset, which is now possible with advancements in in-db, in-memory and grid technologies. But I agree that sampling is a legitimate approach in some scenarios, then it’s all about getting the right sample. We are actually starting to use analytics up-front to create an intelligent, relevance filter to determine the most important information to include in the analytics.
7 – Analytics and information is all about driving decisions, so delivery of those analytics to the point of contact is key. That means embedding analytics in operational systems – which increases the need for feedback and model monitoring and tuning.

I’ve blogged about many of these topics at http://blogs.sas.com/content/datamanagement/

Thanks,

Mark Troester
CIO/IT Thought Leader & Strategist
SAS
Twitter @mtroester