An Innocent Abroad

A Light Touch for Data Warehouse Appliance Customer Support

dclegg@netezza.com — Wed, 04 Aug 2010 21:15:18 GMT

I’ve been at Netezza now for just over three months and I still feel a bit like the innocent abroad but at the same time, I’m clearly a veteran, as other people join our growing team after me. We have our own EMEA Telco Solutions Manager (Chris Smith) and EMEA Alliances Manager (Kate Tickner). I’m just jealous they seem to be on top of their missions, much more quickly than me. It’s great that we have more dedicated resources for working with partners and customers, but i was struck last week by some other the stuff we do, working with customers, that is really much less visible.

Over every weekend I get cc’d on a bunch of weekly reports from folk in the org, which I scan, occasionally raise a question about and file. And amongst the reports are those from our technical account managers. These reports list customer accounts, each with a status - green, amber and red (not many of these – no really), the activities in the account this week, issues outstanding and what’s happening next. All sensible stuff and my input is sometimes to ask for a briefing if an account is at amber for two weeks or more, which generally they aren’t, and at red even more infrequently (but not never – get thee behind me corporate).

Then i rewalized that what I’m seeing is every single customer having their account reviewed every single week – proactively. I even read a report the other week in which a customer had requested less frequent contact; a sort of “everything is fine, don’t call me, i’ll call you”. And I was stunned. This is a customer complaining in that their supplier’s customer service is too responsive! Now that’s not a situation you meet often in my experience. Yet Netezza regard that level of customer service as standard process.

This was the next report for that customer:

Customer: XXX

Status: Green

Activity this week: No activity this week. Model customer J

Issues on hold/carried forward: None

Next steps: Light touch.

I just loved that ’Light Touch’.

A Little Late Musing on the Greenplum Acquisition

dclegg@netezza.com — Tue, 13 Jul 2010 14:39:58 GMT

I wasn’t planning to blog about EMC’s acquisition of Greenplum since Phil Francisco has commented here and many others, more well qualified than me, have had their say (eg here), but it did occur to me that one point this illustrates is how data warehousing got interesting again after ten years as a bit player in the drama of information technology. Suddenly, led i have to say by Netezza back in 2003, a whole generation of disruptive innovators have entered what was a stagnant market of established players and are redefining the segment. Richard Hackathorn said as much at EnZee Universe.

And it's what i had that in mind last week talking with someone from one of Greenplum’s software-only competitors. His take was that the choice was data warehouse machine or software only on a self-assembly hardware configuration (sounds like a grid from IKEA – i’d like to see the allen key in that kit). And Greenplum had decided warehouse machine was the way to go. Of course i wouldn’t be a good Netezza corporate citizen if i didn’t observe that there’s two classes of data warehouse machine: the true appliance (Netezza, out-of-the-box) and the customized data warehouse machine (either vendor-assembled hardware configuration or workload-specific tuned database on commodity hardware, or both).

That doesn’t alter what i took away from the conversation, which was that Greenplum tried the software-only route and then plumped for the machine option. Of course they may well have been made an offer they couldn’t sensibly refuse (in the original sense not the horse’s-head-in-the-bed sense). If so, i guess it’s a case of EMC not being content to be the optional storage component in a configured data warehouse machine and indulging in a bit of supply chain integration. These are interesting times for vendors, customers and, as ever in such situations, analysts and consultants.

Mobile BI and the Vuvuzela Queen

dclegg@netezza.com — Wed, 07 Jul 2010 09:55:27 GMT

I took a little time out yesterday to think about the implications of a possible huge upsurge in mobile BI apps. You don’t have to share Michael Saylor’s unwavering beliefthat iPad and iPhone will be the dominant deliveryplatform in order to acknowledge that there is a real opportunity here. The use cases that I have heard talked about seem to fall into two distinct classes. The first type would be role-based dashboards, for example top ten best movers/worst movers/best aggregate margin etc for a retail store manager. These might be database-intensive queries, but they would generally be cached because they would be slow moving. The second type of query would be ad-hoc requests for specific data. For example ‘what’s the inventory for product xxx?’ These could be run very frequently, with different parameters, but would not be database intensive. And both kinds of queries would be pre-baked into the app that the user downloads to their mobile device. All of this seems eminently doable for MSTR using their new mobile app development toolset and their intelligent server run time.

And the same might well be true for any other comparable app development platform. The critical success factor here is not the deployment platform, though some folks may get burned there depending on how the fight for dominance plays out. The critical success factor is fast response time. A BI specialist running operational reports or complex predictive analytics across the whole dataset may be happy with a multi-second or even minutes response time. But a mobile user has to have the answer quickly or it’s not worth having. From a Netezza perspective this is all music to my ears. Any kind of app that needs fast access to a mass of data is going to need the fastest database they can put at the back of it. We have already partnered with Microstrategy and Quantisense to deliver the Retail Analytics Appliance. It will be interesting to see how the market for mobile BI apps develops. I think it’s clear that retail is a fertile segment, but not the only one.

As you do, if you’re a partner-vendor exhibitor at a conference, you bring along what i’ve always called a twomm (to rhyme with from) - total waste of marketing money, otherwise known as the stand giveaway. Of course they are not a waste of money they represent a unique opportunity to build lasting positive brand identification, blah, blah, blah. But it being world cup semi-final week and with it being my choice, we have vuvuzelas as our giveaway. And give ‘em away we have. I’m not even sure i’ll have one to take back to my Dutch footballing friend who i was texting with as I watched their game (1-0 pleased : 1-1 worried: 2-1 confident:3-1 triumphant: 3-2 clenching) against Uruguay at the conference beach party (thanks Microstrategy – good party). Next to me throughout the second half was a tall Dutch (from appearance, mien and allegiance even if she had said nothing) woman doing a great job with a Netezza vuvuzela. Thank you madam for promoting our brand, and congratulations on a (just about) deserved victory. I’m hoping for more branding opportunities tonight. Not sure who i want to win: hugely talented Spain or unexpectedly fluent and uninhibited Germany. Or should i say who would i prefer to lose: envy-inducing talented Spain or England-crushing Germany.

Anyone get that somewhat torturous music reference in the title?

Your mother is a hamster

dclegg@netezza.com — Tue, 06 Jul 2010 08:45:29 GMT

Up at ten to five to get an early flight to Nice to attend the MIcrostrategy World conference in Cannes. Very interesting keynote presentation by MSTR CEO Michael Saylor at the launch of their iPad/iPhone mobile BI apps platform. Saylor was passionate about iPad; he regards it as a game changer; the biggest thing since the .COM boom, even bigger he suggested. I tweeted it and got a 2 to 1 vote in favour of iPad vs ‘it’s a fad’ (no, literally 2 to 1). I think I can regard that as a statistically valid sample. So bet the company Michael, you have the technorati with you all the way. And we’ll just call the rest the troglodati. The more I play with an iPad the more I feel myself agreeing with him. But later I came across an intriguing argument for why IPad BI apps will succeed. Think of all those managers who’ve bought or want to buy iPads; the perfect justification is ‘so I can be connected to my KPIs and access my critical performance reports wherever I am’. That’s going to get the expense approved, no brainer; well better than ’but it’s so cool, and everyone else has one’.

Then I came across another example at dinner. There was a boys-and-their-toys moment during the hiatus before dessert and everyone is slapping his phone on the table and arguing the case. Most of which was legitimate geekery but it reached its climax when someone observed that the key differentiator for his iPhone4 over his colleague’s Android Samsung with touchscreen (Omnia 2?) was ‘yeah, but it’s Apple’. And his justification was ‘you have similar job offers from two companies and one gives you a Peugeot company car and the other offers you an Audi. Which one are you going to pick?’ The quality, function, form and whatever can all be overridden by brand. And Apple is certainly the brand of the moment.

But the highlight of the dinner was discovering a Frenchman whose favourite film is ‘Holy Grail’. You haven’t laughed quite enough at the French taunting scene, until you’ve heard a genuine Frenchman do it. It’s still making me smile as I draft this later.

In fact i was drafting that much too later. For a bloke who started by complaining about how early he had to get up, I made a poor job of catching up on my beauty sleep. I’ve been on the booth in the partner exhibition hall from early this morning. I made it through my morning shift without falling asleep on the shoulder of a prospect while I listened to his earnest questions about query times for cartesian product joins. Just as well, I might have snored and dribbled on his suit and that’s just plain unprofessional (you can see I have high standards).

Trading privacy for advantage: the way to better service or an invasion of privacy?

dclegg@netezza.com — Fri, 25 Jun 2010 09:26:15 GMT

When he wasn’t being distracted by the footie (today I use a Brit term – you can work out what I’m talking about) and amongst a lot of other themes, Curt Monash made a few observations on privacy in his EnZee Universe presentation. I find this topic scary but fascinating. As I wrote previously, there’s some fantastic opportunities to use customer data to deliver better, more customized service. My feeling is that we’ll willingly make the trade of our data, and by implication our privacy, for the advantage of being better served, provided we are better served. Although we might, each of us, respond differently to our grocery store knowing more about what we want to buy this morning than we do ourselves. We saw what happened at Facebook when they overstepped the mark on re-purposing our data, but I feel there is a subtle difference between on the one hand re-purposing data gathered in an intentional relationship to improve that relationship and on the other selling it to third parties to create new uninvited relationships. But from the reactions I saw of my own non-tech friends in that instance at least part of the resentment was the shock at how much privacy they had inadvertently traded for the value they get from the site.

There’s a lesson for us all to learn about the need for pro-active transparency about data, privacy and re-use, certainly with my generation. Maybe generation-Y just don’t care.

I’m drafting this in breaks between sessions of the Netezza Advisory Council that has been running since the conference ended. It’s a forum for interested and innovative customers with Netezza engineering, product management and product marketing management (not sure how I snuck in). The objective is to help Netezza prioritize development by taking input from customer pain points and use cases and from seeking customer feedback to research projects that are maturing in the NZLabs. In the course of the discussion we looked at some very exciting new technologies, but more of them another day. It’s time for me to head for the airport. Knackered and keen to get home.

Hi to anyone i met at the conference. Especially the guys who wanted to get revenge on me for sticking a mic under their noses when i was doing my roving-reporter impersonation by sticking a phone video camera in may face late at night in a pub. Cheers guys. I hope you got some lovely footage. I doubt it.

Intelligent Economy out of Diapers

dclegg@netezza.com — Wed, 23 Jun 2010 17:49:09 GMT

In Jim Baum’s keynote at the EnZee Universe conference Intelligent Economy was central and it featured elsewhere in the conference. Not always by that name, but with the same essential characteristics – a mass of data and imaginative use of that data beyond the core application it was gathered for. I’ve written before about the analogy with human intelligence - that babies have all the data but they haven’t developed the capability to deal with it intelligently. With us humans (and I’m assuming we have that in common, no offence to my more exotic readers if it’s a false assumption) we gradually learn to do that sort-of automatically until we reach the rational point where we can choose to learn more skills, acquire more data. Organizations entering the nascent Intelligent Economy are already rational (ok, stay with me, i know there are exceptions) and can choose. That was Jim’s final slide on Monday, “think big”. The organizations that choose to exploit the data they have more fully, will have an advantage. For example, retailers who have big loyalty card programs started them primarily to buy loyalty (there’s a clue in the title) but increasingly they, unlike their competitors with less well-developed loyalty schemes, can now do market basket analysis over time for an identified individual customer – genuine micro-segments of one! And that's Jim’s real message, echoed elsewhere, is what do you do next? What new capability do you choose to acquire? What further data do you choose to collect?

Sticking with my retail example, is the next step in-cart technology for the customer to insert their loyalty card into so it can display their usual shopping list? Or custom suggestions? Or shortest route through the store for the list? Or automatic scanning and tick-off of each item as it drops into the cart? Stephen Baker’s book is packed with examples of mind-boggling, but possible ideas like this.

One of the threads that Stephen Baker and others have emphasized is the importance of the big-brained analysts who will be needed to build the models that enable the complex analytics that drive some of these scenarios. But it occurred to me that the limited supply of such mathematicians would be a barrier to progress. A bit like the early assessment of telephone adoption - that if everyone in the US had a phone then one third of the population would be working as switchboard operators (Is that an urban myth? I couldn’t track it down. Let me know if you can source it). That didn’t happen and i suspect the parallel scenario won’t happen; apart from anything else, quantitative analytics is a bit more intellectually demanding than operating a manual switchboard. In the BBBT forum on Tuesday Richard Hackathorn suggested one way to achieve the necessary scalability will be the development and productization of industry-specialized models. These might then be operated, combined, incorporated into business processes. I’m looking forward to the technology advances, Richard’s suggestion and/or others, that will create the breakthrough for mass adoption of deep analytics. But meanwhile the existing technologies and opportunities provide ample scope for the next wave.

And yes i know i'm a Brit using an American English word in my title, but i suspect i have more readers for whom Diaper is the right word.

Pitching in an Elevator

dclegg@netezza.com — Tue, 22 Jun 2010 18:47:16 GMT

Curry was up to scratch last night, definitely the best I’ve eaten in the US, but I spent most of my time at the event with a microphone in my hand and a cameraman with a big, professional-looking shoulder-cam right beside me while I interviewed customers and partners about the conference. If they weren’t just being polite, I’d say the consensus is it’s the best EnZee Universe yet. It’s certainly the biggest.

As well as the interviewing I did last evening, we’ve had a number of people wandering around the conference with flip video cameras in their hands. I’m one of them so I have been trying to do it justice by recording clips in the talks I have been at and sound-bites from Netezza technologists. Today I have been following up the exercise i kicked off yesterday. I toured the sponsor stands in the partner pavilion asking them if they would record a 30 second elevator pitch to camera – to go on the community web-site and to show at the conference. I offered a prize for the best one and I said I’d be back today to record them. So this morning i have been discovering just how wide a range of reactions a little hand-held video camera can have on folks.

Basically my modus operandi was to walk onto the stand just say “I’d like to shoot a 30 second elevator pitch to camera for our conference website, do you want to do it now, or shall i come back later?” Only one person said ”Ok, let’s go”, plenty of people said “Can you give me a few minutes to think what I want to say?” several said “We need to get so-and-so, she/he is the best person to do it”. All of which struck me as pretty reasonable reactions. And i predicted the “let’s go” guy just by his appearance – a frustrated actor in a marketing job i suspect (and no, it wasn’t my alter ego, though maybe it could have been).

One guy on a small booth did turn me down. He said his CEO had left the conference early and he was a techie who didn’t feel confident to do the pitch. Well fair enough, but what really surprised me was the reaction on one of the bigger stands. They delayed while they conferred with the other staff they had at the conference, but in the end they said thanks but no thanks. It just shows that people who can be great one-on-one don’t all have the thespian gene when it comes to one of those scary little lenses.

Off to capture some more mischievous footage and to some more serious meetings. I need more time, which i don’t have right now, to write up some of the surprising things i learned in the keynotes on Monday and a panel on analytics this morning. That should be coming this evening or tomorrowI

Beer and Curry

dclegg@netezza.com — Mon, 21 Jun 2010 18:09:27 GMT

Well this is a baptism by fire for me. When i read my schedule at the weekend i assumed that the 7am /10pm start/finish was just the conference activities from registration desk opening to the evening’s entertainment concluding. I hadn’t realised i would actually be on the go for the whole 15 hours. Six hours in and so far i’m lasting the pace.

I don’t know how many people at the conference know about the Netezza UK tradition of Beer and Curry evenings. And i don’t know for sure that’s what inspired the curry banquet tonight but i suspect it did, so for those who don’t know the background let me elucidate...

It’s a tradition that started some years ago, and basically ‘beer and curry' is just an informal and, we hope, convivial get together of Netezza staff, customers and prospective customers for beer and curry. It’s as likely to be wine nowadays and it will be a pretty classy curry (the last one was here) but we remain true to the spirit of informality. When i first met the idea i was amazed that Netezza just lets prospective customers talk to existing customers freely and without intercession from staff. I went to a beer and curry evening before i joined Netezza and it just confirmed i’d made the right decision to escape the DeathStar for a company so sure of its technology that it doesn’t need an army of marketing and sales staff mediating any interaction between customers and prospective customers.

I guess tonight is the same intention writ large and i’m looking forward to it. I’ve eaten curry in a lot of places, including Heston Blumenthal’s favourite curry house, but i’m expecting this to be a bit special.

I'm trying to think of analogies from consumer sales to this idea of customers and prospective customers mixing freely. In retail the stores are all next to each other and they are competing on every sale to build loyalty. Maybe a better example is the insurance industry. I bet not many insurance companies would host an unmediated forum of their customers to direct you to when you requested a quote for new business. Did i just invent Which for the blogosphere? Or is it already out there and i didn't notice?

Since their Wings have got Rusted.

dclegg@netezza.com — Mon, 21 Jun 2010 14:58:18 GMT

At EnZee Universe this week in Boston, Netezza’s user conference. My first one, of course, so a great opportunity to learn directly from the customers what they are doing and what they want. And i have a challenging role to play. I get to be a flip holder, one of a small group of Netezza staffers who will be wandering the conference with flip video cameras in hand. I’ll be shooting clips from presentations, and any attendees who will talk to me. We’re going to put the clips up on the conference website (attendees only access). And i’ll be in the partner pavilion looking to shoot the slickest, most diverting elevator pitches to upload.

And since i have a very long history of data modelling i’m going to keep an eye open for any presentations on the topic. In Netezza we’re pretty agnostic about logical data modelling; we only get concerned about the physical database design, but i have heard stories about physical warehouse designs that have been optimized for their original target database and continued to evolve over the lifetime of the implementation, By the time we see them they are a long way from the original TNF model. I wrote a little while ago about this and why it’s no barrier to a successful Netezza implementation, but what i haven’t tracked down yet is the best practise for building a Netezza warehouse from scratch. Any pointers/resources/suggestions?

I’m going to try to post every day from the conference, so i hope that flip will help me gather material. If you want to help out with your own Netezza anecdote, i’ll be the slap-head Brit in the red shoes. And a glass of champagne to the first person who correctly tells me who made my shoes. And another for who tells me why the title.

Tea at the Dorchester

dclegg@netezza.com — Wed, 16 Jun 2010 06:58:30 GMT

I’m at a telecoms conference this week, and in previous years i’m told the talk here had all been of networks and switches and bandwidths and copper. This year it’s all about content and customers and services and customization. Very interesting how the industry is moving from building infrastructure & acquiring market share to adding value for customers and delivering content.

The new watchword 'is the customer data is king'. I talked to an infrastructure provider who is planning to collect data on the content they deliver on behalf of customers and then sell access to that data back to advertisers. We talked about an analytics-in-the-cloud type service for the advertisers. And another innovative mobile provider was very interested in the churn management application one of our partner has, although what she really wants is predictive analytics that tell her which subscribers to call, because they might become dissatisfied. The example was subscribers who live near the edge of a cell and get higher than normal dropped calls. Does that need spatial data? Not sure, but even if it does it’s not massively complex. I suggested that combinations of call history to dynamically identify ‘friends and family’ of very recent defectors might be interesting and we talked about a few other possibilities. I don’t think she realized it was feasible to deliver demanding analyes of huge data volumes quickly enough to support this kind of agility in the business. More than one person I spoke to today appreciated that if they didn’t push the boundaries with how they get value from their data, their competitors will.

Of course there’s a big advantage to Netezza (with our newly enriched in-database analytics) for customers to identify more and more complex ways to analyse more and more data. But that doesn’t mean I would steer the conversations in that direction. No sirree, not me. (where’s that Pinocchio emoticon when you need it).

I had to get to the conference at 7am to set up the stand. About 10:30 the previous night i got a call from the sales executive who was supposed to be there to help me to tell me she’d be delayed. I’m still a newbie so this is my first time building and manning the exhibitor’s stand. Good job there were instructions in the case. That way I could ignore them and struggle – well I am a man. We can’t be following instructions until all other possibilities have been exhausted.

And my predictive analytics tell me my sales executive colleague – who incidentally worked her (insert your own inappropriate simile here) all day – will have another urgent appointment come tear-down time tomorrow. We’ll see.

PS. Top conference catering – courtesy of the Dorchester. We got the same pastries at afternoon break as the dowagers taking tea next door. Nice.

Big data and a big dinner

dclegg@netezza.com — Mon, 31 May 2010 12:24:40 GMT

The big event for me last week was introductory training in TwinFin i-Class. What is does is explained in the link, what was most interesting for me was the discussion at dinner afterwards. I’m still learning the technology and the products, but the business case for what we’re doing just seemed so obvious to me right from the start. It starts with, but goes way beyond what Netezza insiders call the ‘speed and feed’ story. Speed and feed just means feed the Netezza box your data and it will deliver greater speed than your existing warehouse infrastructure. The key question then is what do you do with all that speed? What we’re seeing is a world where more and more data gets collected from more and more sources and we’re seeing our customers thinking about some very smart things to do with that data. One example is retailers tying point-of-sale data to customer loyalty cards or branded payment cards, so they can analyse individual spending patterns - scary or cool, depending on how well they turn that knowledge into really helping me with individualized advertising.

For sure, i’ve had enough of so-called eCommerce leaders spamming me with mail. Apparently there’s one global bookseller that thinks i’m interested in fantasy, gardening, learning Italian and technology. A little better analysis of the data they have about me would tell them that i only buy Italian, gardening & fantasy books in November, December or around dates that represent family birthdays. They could even work those dates out if they traced my purchases to the wishlists i pick from. Now it doesn’t matter, i can unsubscribe or i can just delete the mails unread. But in that case we’re both losers and it makes me cross to see the knowledge represented by that data go to waste. And why should it? It’s not expensive or rocket engineering to unify wishlists, purchase records, products and categories and run some simple analytics, though it might need some horsepower and a lot of historical data available.

I hear a lot of discussion about the intelligent economy, and i’m a big believer. But sometimes i think the development of the information economy it’s a bit like the development of the human being. As babies we have all the sense data of an adult but at first we just use it for basic purposes – your hand jumps back from something sharp and your brain registers distress. Later we get smarter and remember that data so we resist the urge to poke the cat in the eye. Later still we learn to coordinate so we can decide to stroke the cat instead. So we use the same data for increasingly sophisticated processes as we retain the data and analyse how we want to use it. And unlike humans, organizations can decide how fast they want to develop. How well they want to exploit those data resources.

There’s always more data you might collect and keep, but i’m thinking that right now the challenge is more about analytics. Our customers , telcos, retailers and others are pushing on hard with smarter ways to use their data. That’s why we’re focussing on moving analytical capability into the database with the i-Class, trying keep ahead of the curve as customers really start to ramp up their deep analytics. I also saw a presentation from one of our partners last week (TEOCO) who provide specific high-value applications for telcos. Exactly the same trend - doing smarter things with more data.

Stuff to do in an Emergency

dclegg@netezza.com — Thu, 20 May 2010 14:30:28 GMT

I’m drafting this as we fly ‘Compact First’ from Johannesburg to Cape Town. Actually it’s a one class plane, but if you want to get the status-conscious sales executive on a one-class plane you tell him it’s ‘Compact First’. But one class or not Kulula Air are pretty classy. Their emergency procedures leaflet in the seat pocket is entitled ‘Stuff to do in an Emergency’ and features a picture of a member of cabin crew with her hands thrown up in horror and a look of panic on her face. Inside, along with the standard sensible stuff about oxygen masks dropping and leaving your high heels behind, there are little sidebar comments like ‘we tried laughing gas instead, but they insisted on oxygen’ and ‘yes ladies I’m afraid you’re going to have to leave the Jimmy Choos behind.’ Also on Kulula Air flights the smoking lounge is on the port wing - they say if you can light it you can smoke it. It was an early start so i was slightly regretting the long discussion about South African wine the previous evening.

I have a long history of conventional RDBMS database design. You use indexes, partitions and other more esoteric features - whatever the database can do for you. And you look at table design ‘data tricks’ - denormalization, data duplication and so on. So as a Netezza neophyte I was keen to learn a bit about how the database features available differ and to learn some design tricks of the trade. I had to sign a non-disclosure before the tech account manager I was grilling would answer. Not because it’s secret, but because he doesn’t want a half-educated marketing director running round thinking he’s a ninja-level design consultant. Turns out that in Netezza it’s pretty straightforward to get going - you just select your distribution key and start loading data. I’m educated enough now to know there’s more to it than that, but not educated enough to explain it. And anyway 90% of it is not important for 90% of databases. The whole Netezza architecture of hardware, firmware & software is holistically designed so you really have to be down to the very short straws before you need to do anything significant for any particular database design.

I thought that was great and filed it away as interesting stuff I should learn more about as I go. But it was brought back into my mind when I was talking to the DW architect of a potential new customer about how we should set up our proof of concept project with them (more about Netezza Test Drive proof of concept - POC - projects here). And he was very concerned that the way his database is designed (the data tricks I mentioned above) wouldn’t add any value in a Netezza database and might even be sub-optimal (duplicated data always carries a load; you only do that if there’s a concomitant benefit). I said he was right that some of the hyper-tuning he had done to get the last ounce of performance out of his existing warehouse (before they threw in the towel and started talking to us) would no longer help. But the debate was around whether he should devote time in the proof-of-concept project to re-design. The Netezza view is don’t. Just load your data and compare the results. We don’t say this to shorten the timescale so we get to a sale quicker, though our sales folk might feel that is a good enough reason. The real reason is more scientific. In any experimental process you change one factor at a time, so when you see a difference you know what caused it. We like to just put the Netezza appliance in to replace an existing warehouse and throw the same queries at it. That way we and the customer know the performance improvement is just down to the appliance itself.

I don’t say there isn’t value in subsequently re-addressing the database design, just don’t do it as part of the POC, otherwise you’re making it impossible to calculate ROI on the appliance and you’re also delaying time-to-payback. At Netezza we’re pretty confident you will get sufficient improvement from a TwinFin that you’ll have all the time you need to address any historical design tweaks. But when I was discussing this with my friendly neighbourhood deepTechs, they said it almost never happens. Customers just never get around to removing these historical idiosyncrasies of a database design because it’s not worth it.

Maybe we should think of them as what first the students of gothic architecture and then the evolutionary biologists call spandrels. They don’t add value, but they are doing no discernable harm.

Next stop Cape Town.

An Innocent Abroad

dclegg@netezza.com — Wed, 19 May 2010 15:46:35 GMT

Monday I flew into South Africa for the first time, just a month too early to watch the footy. That was bad planning, but the reason I’m here was to launch the partnership with BITanium – our new partner in the region. The launch was this morning at a very grand hotel on the outskirts of Johannesburg, I was a bit nervous – this was my Netezza debut in front of customers, but i shouldn’t have bothered. After Calvin Low of Edcon presented there wasn’t much more that needed saying.

Edcon are our first customer in South Africa - a major retailing force in the area - who had taken our ‘Test Drive’ challenge and went straight ahead with their first project. And the numbers he shared with the audience were great. They did a really thorough job on the Test Drive proof of concept (POC); they set us up against the latest version of their existing warehouse and threw a whole spectrum of their existing queries at it. The worst performing query ran for 48 hours on the incumbent kit, then fell over. On the Netezza box it ran to completion in 20 minutes. Another query that had taken 5hrs+ ran in less than 3 minute on Netezza.

It was funny watching the body language of some of the audience who had done the ‘raise eyebrow, lean back in chair, fold arms, cross legs’ response when Chris O’Connell of BITanium had put up the old ’10 to 100 times faster’ slide. When Calvin had finished with them they were doing the ’lean forward , elbows on table, chin on fist, stare intently’ posture.

My personal triumph of the day was answering a serious techie question on solid state media to the satisfaction of the Netezza SE & Tech Account Manager who were in the audience (presumably both doing the ‘bite fingernails, clench buttocks’ response when I launched into my answer).

Now it’s late after dinner at a Greek restaurant, discussing how conventional warehouse designs get so convoluted over time as they are tuned and tuned to keep up with new demands and increasing data volumes. The debate was about whether you should clean up the schema before move it on Netezza or whether you should just lift & shift, let the Netezza box solve your performance problem and clean up at your leisure. In fact we know from experience that clean up often doesn’t happen, it becomes less of an issue when the knotty performance problems that led to the design complexity just go away.

By the way it was Greek because of the vegetarian-friendly menu for one lacto and one pescatarian (me) in the party. You get some funny looks from the average South African confronted with a vegetarian Meze for his starter