Give your company a landline number, not your private mobile.
Go back 15 years and big data as a concept was only just beginning. We were still shivering in an AI winter because people hadn't yet figured out how to funnel large piles of data through fast graphics processors to train machine-learning software. Applications that gulped down data by the terabyte were thin on the ground. …
This Commentard hereby corrects El Reg.
"A data strategy should define a single source of truth in your organization, thus avoiding any inconsistencies and duplication."
Should read as follows:
"A data strategy should define a single source for the opinion of your organization, thus avoiding any inconsistencies and duplication."
When you prove that humans can overcome their own biases by coding AI advanced enough to find and correct errors propagated by those very same biases I will retract this comment.
Until that point in time I will stand by my firm (and very depressing) belief that J. Goebbels was right about the inconvenience of truth. Boardrooms are ruled by market forces and market forces believe in the concept of Infinite Growth. The concept of Infinite Grow violates the Laws of Thermodynamics and cannot exist. Basing the future of a corporation on on what cannot possibly exist is amoral in that the worst outcome is that the corporation and those that depend on it will suffer. Basing the future of humanity on on what cannot possibly exist is immoral.
If I need to explain why that matters then none of it matters at all.
But the very content of the article kind of contradicts it's starting premise. After having read the article, it is clear that you do not set a data strategy "by Friday". It seems to me that that is something that will take weeks to elaborate properly, with many a meeting along the way.
I don't think Goldman Sachs got their petabyte-sized data lake described and specced in one week.
But, apart from that, good read.
Although the content of the article seems good, I seem to remember the Boy Scouts had the motto "Be Prepared". This seems to be padding on the same concept.
Oh and if my holiday included a one-off event my answer would be a flat 'No'. It would also be a 'No' if they weren't prepared to give a written offer to fully reimburse my costs, and arrange (at their expense) a similar holiday within a reasonable time frame.
But then that's prolly why cheapskates firm only employs kids they can lean on.
"This is something McKinsey reckons can help reduce your IT costs and investments by 30 to 40 per cent"
Be afraid. Be very afraid. When the consultant (who probably has never seen an array in his/her life, let alone had to actually manage a data farm on a day to day basis) tells you that you're going to save and save big by doing the latest-and-greatest gee-whiz dedupe and migrate operations, it's time to run like hell.
I am sick and tired of companies or individuals swaggering around quoting they have XXXX P/T/G in their data lake/swamp/warehouse. Big friggin' deal. How much of that data do you use ? Regularly, Honestly.
Companies live in the now. How much did we make today, this week/month/quarter/year. They may compare back to the previous time period , but that is it. If you have 5 years of transaction history. Straight away your lake contains 60% dead data. Sometimes, depending on the industry, you run a few models to look at seasonality of your business. I'll give you that, 3 years. So still 40 % dead data
Most companies claim massive lakes or oceans of data. The truth is they only need a kiddies pool.
The concept of a data strategy is something every CxO will claim they need to implement or have implemented. Right up to the point they start to understand the inherent costs in doing it. How many times have you heard this from the senior managers. We don't have the money/resources to do that now. We'll just go with the cheaper option and then we will get the money in next years budget to do it . or "The project timelines don't allow for that. Looks like something we can do after going live". In many instances the data strategy does not deliver hard bottom line returns, the in-direct benefits are massive and not just on the reporting side.
Yep working through the data governance model for this company. Apparently it's something they absolutely need. They are having problems with the concept of not cow-boying everything, but taking time to design things properly. baby steps, baby step
Completely agree. This whole "Gotta keep everything forever" mentality is so stupid. Companies think they have to keep everything now so they can have some analytics/machine learning wizard wave their magic wand over all that data and extract something useful. In reality, it's not going to happen.
I firmly believe 99% of the historical data companies keep is worthless, and it only serves to line the pockets of the storage and cloud vendors they use to house it.
I'm out of the storage field now (thank God) but if I were in charge of data strategy, I'd dump EVERYTHING that wasn't mission critical or required by regulation or law. Everything else would get dumped, and I'd free up all that operational and/or capital expense.
Data should only exist if it's used to generate qualitative (not quantitative) info used to make actual business decisions. Everything other than that isn't data, it's trash, and it should be deleted.
Biting the hand that feeds IT © 1998–2019