Blogs

Post a Comment

Business Transformational Analytics Performance, and Cat Food

March 7, 2012

Last week I was in Bucharest and Ljubljana, but before I headed to the airport on Monday I recorded a webinar – doubtless I’ll be blowing my own brass section about how and when you can watch it, when I know that. Usually for webinars, as for other presentations, I have my own slide deck, which is almost always a slight variation of a previous one, but on this occasion I was using a pre-set deck because I was a late substitute for the planned speaker. The deck, as usual, included a bunch of Netezza customer stories, including one I wasn’t familiar so I had some reading to do over the weekend to get acquainted with it. It was an interesting application: an investment bank wanted to implement an algorithm to identify pairs of equities in the market whose price curve have some kind of correlation. Luckily for them there’s a very nice bit of mathematics called the Spearman’s rank correlation coefficient function that assesses the population covariance and the standard deviation between equity combinations. But for Spearman to do its magic it first needs to calculate volume weighted average prices and return from previous close values for each equity to be compared, and of course to do the job properly every combination of pairs of equities must be compared and you’ve got to go through all the trades of all the instruments to derive the calculated values. A pretty demanding bit of analytics and on their previous data warehouse it ran for 2 hour without completingi. When they ran it on Netezza, it completed in 2 minutes.

A great story to add to my repertoireii - at least 60x faster, which wasn’t the biggest performance multiplier in the customer stories I had in the deck, but it’s a good example of why great performance isn’t just ‘oh wow my report ran really quickly, that’s nice’. It’s another example of what I call ‘business transformational performance’iii. So I took a copy to add next time I’m talking to banking prospects and customers. Mind you, I’m going to have to read up some more on the maths; I suspect Spearman wasn’t born when I took my maths degree.

As you might possibly have gathered, I spend a lot of my time in airports, so I am occasionally exposed to Oracle’s big poster ads, and I saw one as I arrived at Heathrow. For years Oracle used to do the ‘9 of the top 10 xxx firms use Oracle’ and I always used to wonder if that was inspired by the old UK TV ad for cat food  ‘8 out of 10 cats prefer Whiskas’ with a clip of cute moggies tucking in enthusiastically to the Whiskas. Today they had a different ad. It was an Exedata ad and the big number leaping out at me was a 13. I think it was company x runs 13 times faster with Exadata (concentrating on my driving so I didn’t read it all). 13 times faster? I suspect 9 out of 10 Netezza customers wouldn’ t get out of bed for 13 times fasteriv. Nice, and maybe even worth the money, but hardly business transformationalv.

Off to Lisbon in two weeks for more big data and happy Netezza customers.

For more information:


i Ok, you wrung it out of me; it was an Oracle warehouse.

ii There’s a bunch of Netezza customer stories here

iii As soon as I typed that, my marketing bollox radar started to twitch, but actually I can’t, off the top of my head, translate that into succinct, plain English. I suppose it comes out something like ‘so fast you can achieve stuff that just wasn’t possible for the business before’.

iv That’s a gag; please don’t sue me.

v This is not the bank story I talked abut above, but it’s a very cool story about MediaMath’s online ad platform