As you rightly point out, agile is the right direction. We have bi-weekly reviews with our clients to make sure things stay on track, and to verify that we are building the right product. It is a collaboration. Costs do tend to rise a bit, but the client will rarely get upset with a 10-20% budget miss – if the product works really well. Even taking a month or two longer than planned is rarely questioned.

The key to a happy relationship is building a good piece of software, without completely blowing the schedule or budget. I think most clients are really good about it, as long as you earn their trust, and confidence.

Just finished reading your book. I am considering a technology practice consulting role with a national firm. Having always been on the client side of consulting engagements, and leading technology projects, your experiences, insights and perspectives have been hugely helpful.

The risk vs reward proposition as well as believing I am at a place in my career to make a decision to start a consulting career steered me to find your book.

Thanks again

I suppose such a thing is possible, in theory. However, I fail to see a great reason for doing so. The analysis is usually going to be in some alt structure in most deep analytics scenarios (think Hadoop or other map-reduce platforms).

Why do we need everything on the same platform anyway? Why even bother trying? We can analyze data in context of other data across multiple platforms reasonably easily. The only people I see wanting to build platforms are people that either sell the platforms (i.e. think EDW vendors) or IT departments who have already invested heavily in such systems and for some reason need to justify said system’s existence.

ok. but wouldn’t it be possible to leave space and create a rubric for either manually adding or automating additions to edge data later. To analyze it in context with other data.. which without specific examples I can’t be sure is necessary you would have to pull it from it’s silo anyway, right?

When I say edge, I mean things like logs (structured, unstructured), sensor data coming in from devices around the network, systems in shadow IT, and anything else that tends to escape the control of IT directly. The “dark matter” of corporate data, if you will – which almost always escapes direct control from corp IT.

Thanks for replying. It’s been a while since I managed a corporate data structure and even then they were only segments of the enterprise. Not sure i know what you mean by the efge… and when you say platform you mean like Oracle?

I actually don’t think it’s ever possible to be on the same platform. Too much data being generated at the edge in most organizations, much of which is the most useful/interesting. I am doing this now with a major IaaS provider – where the amount of data being generated, the organizational complexity, and the need for both owned and outside data in order to do proper analytics implores us to take a multi-platform strategy.

In other words, data generation has reached escape velocity. It is being generated far faster than any org can warehouse it. You have to take a strategy of dealing with data as it is, not as you wish it to be.

However…I do think one data standard is a key idea. If information flow stimulates growth..getting rid of data silos is important. Is it not possible to migrate the data of an existing organization onto one platform?
I am in the opening stages of an experiment to see if all data…. can be stored in the same warehouse with different permissoning structures.