Sunday, October 11, 2009

Aweekstweets Oct 4-11 2009

Aweekstweets Oct 4-11 2009

  • Social media in customer svc: "witness factor": clients testify to your screwups & successes: performance anxiety keeps you at peak. #btf0931 minutes ago from TweetDeck
    • JK2---To some degree, social media just intensify the hellish online culture of flaming, blaming, and shaming. But, if you have fire-resistant skin and are not agoraphobic in the slightest, social media are the platform and proscenium for your world stage. The audience witnesses everything you do. You may or may not witness them in return. You need validation, or just a place to vent? You up to it?
  • Why doesn't TweetDeck let create column of tweeters I've never accepted, have unfollowed, or just received? A tweet-spam folder, so to speakabout 6 hours ago from TweetDeck
    • JK2---I just want to witness the passing parade I’ve chosen to let pass me by. Or in whose passing parade I’m currently marching.
  • Problem re "crowdsourcing" is that "crowd" often far more noise than signal. Trick is tuning your receiver to hidden harmonies: throngsong.about 7 hours ago from TweetDeck
    • JK2---The harmonies are usually in the overheard overtones: the distant chitter-chatter that plays back in your head when you’re not paying attention.
    • JK2---Let the world witness your internal processes, including your backlogs and bottlenecks, your deficiencies and inefficiencies. Yes, your competitors should be able to peer deeply into your innards too. What do you have to hide?
    • JK2---Real friends stay with us. Faux friends merely witness our outer manifestations, if they’re really paying attention, which they’re probably not.
    • JK2---Lean socializing means a clean, uncluttered experience. A comfy, non-distracting, pressure-free personal experience of community. As comfy as a stroll in the neighborhood, among friendly, non-judgmental people who welcome you into their presence—and you them. And who make your day a little nicer—and you them. Lean means a single “neighborhood.” To the extent that social media means having to hop back and forth among Twitter, Facebook, LinkedIn, MySpace, blogs, etc., it’s far from a single neighborhood, and none of them feels completely like home. It feels like a sprawling megalopolis in the making. So, no, it’s not lean. It’s bloating out of control.
  • Clay, #btf09, says audit BPM tech portfolio re specific biz requirements. Get bloat out of your BPM tooling. Reuse tools across BPM projectsabout 8 hours ago from TweetDeck
    • JK2---Lean business process management (BPM) goes against the grain of traditional enterprise adoption of this technology: the tactical, bottom-up, project-specific implementation of app-embedded, process-specific, siloed workflow platforms and tools. You can only implement lean BPM if your companies has a BPM competency center and a senior-level sponsor (e.g., CIO) who empowers it to consolidate existing workflow environments into a corporate standard platform—and to dictate use of that platform for all future projects.
  • Audience Q at #btf09: ROI metrics to justify lean BPM project? Clay's A: Don't hide behind ROI numbers w/o strong narrative re biz impact.about 8 hours ago from TweetDeck
    • JK2---The BPM justification/payback narrative is key. A single powerful ROI anecdote, with or without actual numbers to back it up, can make all the difference.
  • At #BTF09, question re BPM ctr of excellence relevance to lean BPM. Clay says BPM CoE must be aligned w/ biz strategy, connect to biz value.about 8 hours ago from TweetDeck
    • JK2---This is, of course, motherhood and apple pie advice for any technology. Where BPM is concerned, it’s even more critical: the business value of a business process platform must be unimpeachable. The business value of, say, a new data storage investment is also important, but it’s more indirect and bottom line. A business process platform should be a top-line investment from the start: it should have clear payback in helping organizations be more productive and agile in responding to challenges.
  • Clay, at #btf09, says business analysts must have passion for process, but work closely with tech analysts, to deliver on lean BPM projectsabout 8 hours ago from TweetDeck
    • JK2---Truth be told, passion for results is what counts. Workflow experts become passionate about process at their peril. That’s like being bullish on bureaucracy or rah-rah about red tape. Business analysts must work closely with tech analysts to trim process to its leanest, least obnoxious, most agile essence.
    • JK2---Consolidate your many process monitoring dashboards down to a single correlated view, updated in real time, available to all internal and external stakeholders, triggering automated and manual exception handling actions.
  • Keys to lean processes: Process alignment to strategy. Process governance and skill building. Process waste reduction. C. Richardson. #btf09about 8 hours ago from TweetDeck
    • JK2---This is the essence of every scientific management, business process reengineering, and quality assurance philosophy ever put forth by anybody anywhere.
  • C. Richardson, #btf09, on lean BPM: Watery metaphors: old BPM "waterfall method"; new BPM "cloud." Which processes should I put in cloud?about 8 hours ago from TweetDeck
    • JK2---This is the age-old “what’s my core competency vs. what should I outsource” debate, latest chapter.
  • Listening to Clay Richardson on lean BPM at #btf09. Michael Hammer's BPR was euphemism for firing people. Please Hammer, don't hurt 'em.about 8 hours ago from TweetDeck
    • JK2---This is the age-old “shareholders uber alles vs. employees come first” debate, 80s flavor.
  • Building personal brand as an analyst. It's all about personal integrity, viewpoint, approach, & passion. Being a whole person. A personage.about 18 hours ago from TweetDeck
    • JK2---This is the age-old “demarc between life, career, and job” question, IT analyst rumination variety.
    • JK2---Everyone wants to cash in on being the start and/or end of someone’s online voyage of exploration, or being the map that connects those points.
    • JK2---Everyone wants to have at least one absolutely indispensable product that customers cannot live without—and that is their last best hope for sustaining a business model in the face of erosive competition.
  • "Fit to purpose frameworks" for lean software. An anti-monolithic deploy. In analytics, user-config modular hetero BI/DW solution. #BTF099:23 AM Oct 8th from TweetDeck
    • JK2---I’ll spell that out. A lean analytics deployment is one that users customize from modular platform, application, and tool components—business intelligence (BI), data warehousing (DW), and whatnot—sourced from heterogeneous vendors. It’s “roll-your-own” but entirely standards-based, rather than a monolithic stack from a single vendor who locks you in.
    • JK2---Lean is anti-lock-in. Lean is ability to throw it all out, or repurpose and reassemble any set of components as needs change.
    • JK2---Lean is clean. Consolidate to fewer system-of-record applications, eliminating the need for intermediary solutions that do downstream profiling, matching, merging, cleansing, correction, and governance of data.
  • Lean analytics? Trim cost thru BI/DW consolidation, fit to purpose thru deep domain content, agile to new biz needs thru user empowerment8:58 AM Oct 8th from TweetDeck
    • JK2---In the coming year, I plan to go deeper in my Forrester research in all these areas.
    • JK2---Just my basic working definition of “lean,” extensible to any application, information, or process management topic.
  • My replacement laptop will be here tomorrow in time for #BTF09 #FORR. Tons of new e-mail in my inbox. Kiosking hard on my feet. says4:53 PM Oct 6th from web
    • JK2---New laptop works great so far. I’ve caught up on e-mail. I worked out my legs over the weekend. I’m on my feet now, typing this at the food/coffee bay in the back of a moving airplane. Flight attendants are up front serving people. You work where you can.
    • JK2---Essentially, everybody reusing everybody’s mashed-up analytic apps.
    • JK2---Workload isolation—transactional vs. analytics--makes great sense for so many reasons. If they’re both using the same underlying platform, we’re talking separately deployed servers, clusters, and tiers.
    • JK2---In 10 years time, we won’t be sad. We’ll forget. We’ll be so surrounded, suffused, and saturated by online everything—including the devices to access, view, consume, and interact with it all—that hardcopy will just be another option. We’ll be comfy and habituated to the new world. Much as nobody seems to be mourning the loss of ubiquitous hardcopy family photos anymore.
    • JK2---And I’m sure they’ll have interesting announcements at IOD in 2 weeks.
    • JK2---“Next wk” is here. I’m on way to San Francisco for OOW. Sun acquisition not complete yet, so I suspect there won’t be much discussion by Oracle of Sun futures.
  • RT @bevelson "Pentaho acquires piece of Lucidera for new OLAP UI" JK--BI SaaS pure-play space not yet viable. But great IP breeding ground.9:43 AM Oct 5th from TweetDeck
    • JK2---DW SaaS also underdeveloped, emerging, immature space in late 2009. But will be huge in 2010-11.
    • JK2---Also a 2010+ focus area for my research at Forrester.
  • Few phrases make me chuckle like "think tank." Older term for "research, analysis, and consulting firm." Implies hermetic isolation chamber.9:07 AM Oct 5th from TweetDeck
    • JK2---Forrester is one such “think tank.” No place could be less hermetic. A little isolation now and then would be great—so I can ward off distractions and get work done. I doubt that analysts in other firms have the luxury of shutting themselves off. That would sort of defeat our core function: deep market research into the latest/greatest and future of everything IT.
    • JK2---For analytics, core function of intelligent storage layer is faster queries.
    • JK2---In high tech, the most important phenomena are those entities either above or below the threshold of everyday perception. Above that threshold? Wide-area, wireless, virtual, software-driven clouds. Below that threshold? Invisible processes executing in memory inside boxes, inside molecules, inside atoms, inside quarks, astride the tiniest strings in superstringy base reality. Inside imagination.
  • Endured yet another bee sting today mowing the lawn. Last mowing pre-frost. Last sting pre-extermination.7:18 PM Oct 4th from TweetDeck
    • JK2---I have forgiven the little guys. Besides, I can’t tell which of them is the culprit. They all look alike.
Posted by at

Sunday, October 04, 2009

Aweekstweets Sept 27 to Oct 4 2009

Aweekstweets Sept 27 to Oct 4 2009

  • "80% of Twitterers tweet mostly about themselves" (http://bit.ly/1esLMg): JK--Not social networking. It's I'M-SO-SPECIAL networking.19 minutes ago from TweetDeck
    • JK2---Let’s not imagine we’re becoming altruistic with all this “social networking” going on. We’re primarily tooting our own horns, letting off steam, indulging in our obsessions, measuring ourselves against others, finding shoulders to cry on, grabbing whatever we can put our virtual hands on, and otherwise just being more completely into ourselves. I don’t find a lot of actual collaboration, charity, and mutual aid being transacted among all these so-called “friends.”
  • Thank you "sesli sohpet" for your comment "eğlencenin sohpetin arkadaşlığın tek adresi" on my latest blog. My reply: "rwejl ajfqkt rjigfjs."about 4 hours ago from TweetDeck
    • JK2---Ah yes, remember blogs? This was a bot-generated gibberish response to my most recent Forrester blogpost. I turned off the comments feature in James Kobielus’ Blog ages ago. I was tired of bot spam. I was also tired of moronic, superficial, or irrelevant comments by people who obviously hadn’t taken the time to actual read or understand my post.
  • RT @santaferraro "SOA+BPO can produce self-optimizing models for... customer offers, supply, etc" JK--Yes. Auto model scoring/eval/promotionabout 4 hours ago from TweetDeck
    • JK2---One of the key solution differentiators in the predictive analytics and data mining market is the ability to a) generate multiple alternative models (with different algorithms, approaches) from a common historical data set, b) promote the “champion” model with the most predictive power to inline production status, c) regularly re-score the champion and all “challenger” models against fresh data, and d) automatically promote a challenger to production when it shows greater predictive power than the former champion. From what I can see, that’s the emphasis of “self-optimizing models.”
  • RT @scottkarp "journalists can find 'clues' on Twitter & provide thoughtful analysis & summary." JK--More than clues: can find entire "lead"about 4 hours ago from TweetDeck
    • JK2---Hey tech industry professionals, believe it or not, you can stop reading your e-mail and just focus on reading your tweets each day. If you get tweet feeds from the right mix of publications, vendors, analysts, consultants, users, and others, you can stay abreast of the day’s news and commentary through your Twitter client. If you want further depth and context, you simply follow the tinyURLs back to longer articles or blogposts. There is far richer aggregate content every single day on Twitter than we give it credit for. In fact, I’d rather monitor Twitter than the blogosphere, which is not an efficient news-dissemination medium.
  • Service-Oriented Analytics and business process optimization? Yep: reuse models in BPM/LOB apps, & optimize decisions thru real-time scoringabout 5 hours ago from TweetDeck
    • JK2---Predictive models will become just as central to transactional applications in the future as relational databases are today. Predicting a customer’s likely response to alternative offers is key to customer relationship management. Predicting future resource requirements is central to enterprise resource planning. Predicting likely logistical bottlenecks is essential for supply chain optimization. And so on and so forth.
  • Service-Oriented Analytics and technology populism? Yep: enable end users to mashup, publish, share, and reuse analytics collaboratively!about 6 hours ago from TweetDeck
    • JK2---Companies can only meet the future head-on if they empower everybody with predictive modeling expertise. No, you don’t have to become a statistician or soothsayer. You simply need to be able to create business planning, time series, and what-if scenarios that leverage the best models and visualizations that others publish in spreadsheets, reports, and dashboards. And you need to be able to share out your best models for others to use and extend for their own needs.
  • Service-Oriented Analytics and lean management? Yep: don't build any analytics that you can't reuse to the max!about 6 hours ago from TweetDeck
    • JK2---Nothing’s worse than buried treasure—millions of spreadsheet models trapped on desktops, unknown and inaccessible to others who might benefit from them to address new business problems. Every business model—expressed as a Excel spreadsheets, logistic regression model, neural network model, whatever—should be searchable, shareable, and available for maximum reuses throughout your company.
  • E-mail client apps are a far more productive fast-spam-deletion tool than browser interfaces. Which is why I'm not going full browser.3:14 PM Oct 1st from TweetDeck
    • JK2---Still, to date, nobody has really implemented “bulk mail” folders properly. None of the anti-spam services that I’ve used ranks the mail in that folder by degree of likely spamminess, so I can sift the false-positives with fewer keystrokes or mouse-clicks, and with a single rapid glance.
  • BTW, my Aweekstweets 3 wks ago, per tweet-volley with @jamet123, was kernel of today's Service-Oriented Analytics blog. My thinking builds.2:58 PM Oct 1st from TweetDeck
    • JK2---You have to reuse your own work to the maximum extent possible. I have to do analysis on the run, and lash it all together into a coherent structure at a moment’s notice. Try being an analyst in this industry, and you’ll see what I mean.
  • You start to click one window/object, but another pops up suddenly in its place, and you mistakenly click that. Whatcha call those moments?2:48 PM Oct 1st from TweetDeck
    • JK2---I call it “clickjacking.”
  • RT @SethGrimes @atmanes "Talmud [etc] are pre-electronic hypertext" JK--This is sorta like arguing who was the Bob Dylan of 15th century.1:58 PM Oct 1st from TweetDeck
    • JK2---Footnotes. Endnotes. Tables of contents. Sidebars. Indexes. Who invented those? That’s as much “hypertext” as anything in the Talmud. Essentially, they’re all cross-referencing navigation guides pointing to and/or containing material with some direct or tangential relation to the main text.
  • RT @atmanes "TimBL and co. at CERN developed a hypertext system in 1990. " JK--Thanks for listing milestones in Internet ongoing re-invention12:13 PM Oct 1st from TweetDeck
    • JK2---Let’s split hairs here. Was the “Internet” invented when ARPANET went live in 1969? When TCP/IP was defined a few years later? When the term “Internet” was coined? When DNS was established? You see what I mean? No one innovation defined this still-evolving creation we call “the Internet.”
  • New Forrester blogpost: "Service-Oriented Analytics: Tapping into the Predictive Smarts of Your Entire Organization" (http://bit.ly/aZlhE)12:12 PM Oct 1st from TweetDeck
    • JK2---Read it.
  • Internet keeps re-inventing itself. Newspapers keep on re-inventing themselves. Institutions accrete: don't spring entire from Zeus' head.8:58 AM Oct 1st from TweetDeck
    • JK2---Newspapers today are a far cry from the broadsides of the 1700s, the penny press of the 1800s, and the dense gray hard-news periodicals of the pre-WWII period. Most of today’s newspapers are in fact collections of feature stories heavy on analysis, graphical blandishment, and human interest. Everybody got the actual “news” hours before on the Internet, or continuous TV news channels.
  • "Bing Records First Monthly Decline Since Launch" (http://bit.ly/pMHYZ) JK--Bada BING bada boom? Or bada BING bada BUST? You decide!6:57 AM Oct 1st from TweetDeck
    • JK2---Today’s search engine wars feel so yesteryear. Just as yesteryear’s browser wars feel as ancient as the Punic Wars. The new battlefield for “eyeballs” is in social media.
  • TweetDeck has UI of one-armed bandit. Drop quarter/tweet into slot/box. Pull the arm/[enter]. Watch wheels [all friends:mentions:DMs) spin.8:07 PM Sep 30th from TweetDeck
    • JK2---Seriously. Pay attention to the entire experience of using this and similar Twitter clients. It almost feels McLuhanesque: it’s a new visual grammar for consuming streams of real-time intelligence from all over the planet. Just as TV shaped our global consciousness in the 50s and 60s.
  • RT @timoelliott "had to listen to 20 years of [tiresome definition wars re BI]" JK-Not as annoying as metaphysics of "data" v "intelligence"7:59 PM Sep 30th from TweetDeck
    • JK2---This is all ersatz armchair epistemology. Seriously. I don’t recall any actual, honest-to-God, pedigreed philosophers debating this “data” vs. “information” vs. “intelligence” vs. “insight” jazz. Give it a rest please, BI industry. Nobody justifies an investment in one BI tool over another based on one vendor’s ability to help you deliver any of these higher-order “i-words” any more efficiently. Likewise, I’ve yet to come across a DBA who boasts that their data warehouse packs more “insights” per terabyte or per-dollar spent than the previous one.
  • RT @tkgeorge @timoelliott @merv "debating "BI is not X" semantics = monumental waste of time." JK--Huh? 10-sec tweet-blurt = no time at all!3:15 PM Sep 30th from TweetDeck
    • JK2---Debating these never-ending definitional issues is something we do as small talk, out in the hallway over coffee between sessions with substance. It’s an ice-breaker. Not a waste of time. Just a way to fill otherwise idle time.
  • RT @JeromePineau "Is MSFT doing/saying anything about columnar re: Madison or are they still all about rows?" JK--Re Madison, still all rows3:13 PM Sep 30th from TweetDeck
    • JK2---Another one of these never-ending debates in the BI and data warehousing space: the query performance benefits of column-based data storage/access (i.e., what Sybase, Vertica, ParAccel, SenSage, Infobright, and others offer) vs. row-based (which is what most DBMS/DW vendors implement). Another one of those highly technical issues that one dare not mention in a social setting devoid of DBAs.
  • Synonyms abound for condition of "too much": surfeit, surplus, superfluity, excess, plethora. Really, it's all too much!2:14 PM Sep 30th from TweetDeck
    • JK2---All of these ones listed are in the working vocabularies of most people with a college education. A healthy command of synonyms gives writing texture and poetry. No—don’t consult Roget’s unless you already know a word, understand its meaning, and simply want a second opinion on some subtle shade of meaning. Nothing’s more obvious than a fool with a thesaurus who plops a slightly-wrong quasi-synonym—the verbal equivalent of a sour note—into their writing.
  • RT @JeromePineau "Isn't IQ columnar & ORCL claiming hybrid now?" JK--I said "most"; Sybase, Vertica, etc are; ORCL just partly in ExadataV21:41 PM Sep 30th from TweetDeck
    • JK2---Questions such as this presuppose that it’s somehow inevitable or especially virtuous for vendors--and by implication, users—to “go columnar.” There’s a lot of this assumption floating around the industry. If columnar were so much slam-dunk superior to row-based storage (which has partitioning, indexes, join strategies, materialized views, and the like to speed its queries), then why didn’t everybody adopt columnar en masse long ago?
  • Still this odd enduring notion that columnar database vendors compete primarily against each other. Duh--most top DW vendors non-columnar!1:25 PM Sep 30th from TweetDeck
    • JK2---Fortunately, the columnar database vendors themselves are not very parochial. Sybase, Vertica, and others duke it out—fairly effectively—with row-based RDBMS vendors on the criteria that matter: fast queries, fast loads, mixed-query workload management, and the like.
  • Is Twitter tracking which human being tweeters are the most hashtag-intensive in their tweets. My fingers stumble that high on keyboard.1:17 PM Sep 30th from TweetDeck
    • JK2---Maybe it’s just me, but I find that too many hashtags interrupt the flow of a tweet, and make it even less readable than it already is. A hashtag-intensive tweet looks like it’s more fit for machine-processing than human consumption.
  • RT @nenshad: "ironic if scoring a pa/dm ranking algorithm by hand." JK--Is fingers on keyboard "by hand"? Standard Forrester Wave scorecard.1:15 PM Sep 30th from TweetDeck
    • JK2---Not ironic at all. “Scoring” means one thing in the predictive analytics and data mining (PA/DM) arena, and something entirely different where vendor/product evaluation is concerned.
  • Local TV meteorologist on global warming this morning: "These are only models." Yep, myriad warring predictive models mining common history1:07 PM Sep 30th from TweetDeck
    • JK2---Scoring a global warming model, in PA/DM terms, involves marshalling a massive and growing body of scientific research data to validate the predictive power of a massive and growing (and contentious) range of scientific hypotheses. “These are only models” means “these are only the current ‘champion’ models favored by whoever in the scientific community has the upper hand at this time, and which may be eclipsed in the future by today’s ‘challenger’ models for various reasons, with economics, government funding, partisan politics, and shifting cultural zeitgeist among those reasons.”
  • Crunching thru vendor data gathered under Forrester PA/DM Wave. Yes, dear vendors--I'm pondering EVERY LAST DETAIL. Will start scoring soon.12:50 PM Sep 30th from TweetDeck
    • JK2---And yes, dear readers, I print out much of it to read on airplanes. Can’t stand craning my neck to watch in-flight movies. Much rather Zune in and zone out.
  • RT @merv: RT @nenshad: "Please repeat after me: BI ≠ reporting" JK--Please repeat after ME: advanced analytics ≠ BI.12:42 PM Sep 30th from TweetDeck
    • JK2---Please repeat after me: “BS ≠ BI.”
  • Interviewed by Jeff Kelly (http://bit.ly/8i3iw) on DW appl vdr "same-day deploy&value" claims. Yep, I said, it happens: for specific deploys12:40 PM Sep 30th from TweetDeck
    • JK2---My actual e-mail response: “A fair share of marketing hype, for sure. Yes, of course all of these DW appliances can be “up and running in less than a day.” But the issue is: can they be optimized for all of the customer’s queries and loads, and integrated with all the customer’s BI applications and data sources, in less than a day. To the extent that the customer, or the DW appliance vendor working on-site with the customer, can do that optimization and integration in a matter of hours, yes indeed it is possible to get “deep analytical insights” the same day from data processed through the appliance. This scenario is certainly possible, and I’ve heard many DW appliance users who say they’ve done it. It’s especially possible/feasible for the simpler DW appliance deployments: in other words, those that involve a limited range of queries, reports, ETL jobs, sources) where the DW appliance is already pre-certified to work with the customer’s existing BI and ETL tools. It’s also especially possible/feasible if the DW appliance you’re deploying incorporates the same DBMS (e.g., Oracle, IBM DB2, Microsoft SQL Server) as your previous DW environment—which means the migration of data, schemas, indexes, queries, ETL, and metadata is correspondingly simpler than if y you are migrating to a DW appliance that incorporates some DBMS you’ve never used before.”
  • What you realize re tweets, columns, analyst rpts, even books, it's not wordcount. It's the "call" you make. That's the payload, the impact!12:10 PM Sep 30th from TweetDeck
    • JK2---My shortest, sharpest tweets often have bulletforce impact.
  • I generate tweet fodder as I generated Network World op-eds for 21+ years. Milk the moment. Develop a POV. Bang 'em out. Don't look back.12:04 PM Sep 30th from TweetDeck
    • JK2---If you’re receptive and prepared, your freshest of-the-moment observations are often the truest. Tweet ‘em! Now!
  • Speaking of tweet-tight, that's not always best approach. I prefer "keeper" slides that I study in detail, hold for future reference.11:34 AM Sep 30th from TweetDeck
    • JK2---I often lose interest in purely eye-candy too-high-level business-value blathering presentations. I want substance.
  • RT @Victor_Sokovin "Exadata V2 cell equivalent in Teradata [AMPS]?" JK--DK. Larger issue: need industry-standard DW performance benchmarks.11:27 AM Sep 30th from TweetDeck
    • JK2---I’ve sketched out thoughts for those, but haven’t developed them to any degree. Yet. Personal action item for 2010.
  • RT @randydarthur "I am more than a MARCOM router." JK--To the extent one becomes a router, one's credibility suffers. Analysts add value.8:03 AM Sep 30th from TweetDeck
    • JK2---It’s an ongoing struggle, as an analyst, to not get manipulated into becoming an unwitting tool of vendor marketing apparatus. We all push against that. Vendors love us because we’re in a position to influence customers. Vendors love us, but it’s what the sociologists call a “secondary relationship” kind of love, not a “primary relationship.” A secondary relationship is where we “use” each other.
  • Thanks Twitter follow-me requesters for the personally empowering, search-engine-optimizing, Internet-marketing advice. But all full up!7:15 AM Sep 30th from TweetDeck
    • JK2---Speaking of being used. These sorts of twitterers want to extract cash from us (use their direct-marketing consulting services), so that we can, in turn, extract cash from others (direct-market whatever it is we peddle to others).
  • Noticing that reporters and other analysts fail to grasp Oracle Exadata v2 significance. Comprehensive in-DB analytics primary. OLTP 2nd-ary7:09 AM Sep 30th from TweetDeck
    • JK2---The Oracle’s marketing the latter angle. So it’s not just the reporters/analysts who are misrepresenting the significance of this release.
  • Dear Twitterers & Facebookers: Please don't send "EVERYBODY SHOULD READ THIS ARTICLE!" e-security blasts. Neverending alertstreams grow old!7:03 AM Sep 30th from TweetDeck
    • JK2---After a few such blasts from on individual or entity on a social network, I’m inclined to unfollow them. Cry wolf!
  • Dear Twitterers at Industry Events: Don't tweet the exact same promotional boilerplate that the sponsoring vendor handed you.6:59 AM Sep 30th from TweetDeck
    • JK2---Commercial vendor-driven friend-spam from otherwise trusted sources tends to strain relationships. Unless of course you’re plugging a Forrester conference, then it’s cool!
  • RT @CurtMonash "Workload mgmt is a necessary but not sufficient condition for OLTP/DW integration." JK--Yeah, of course. Other conditions?6:54 AM Sep 30th from TweetDeck
    • JK2---Example of Curt being curt.
  • Fear uncertainty & doubt are our fate. Change is big FUD generator, but so is stagnation. Arbitrariness of it all is what we can't abide.10:00 PM Sep 29th from TweetDeck
    • JK2---One friend loved this one. Thanks. I labored over it a bit more than I do most tweets.
  • RT @marcusborba: "integrating OLTP & DW w/ Exadata 2) @CurtMonash http://bit.ly/28tWI0" JK--Oracle workload mgt tools are key ingredient.9:40 PM Sep 29th from TweetDeck
    • JK2---The larger picture of Oracle Exadata v2 is that it provides an intelligent storage layer that enables any application logic—OLTP, OLAP, predictive analytics, data cleansing, etc.—to be executed, scaled, and accelerated through massive parallelism. This intelligent storage layer virtualizes across disk- and cache-based persistence, leveraging hybrid columnar- and row-based persistence. Oracle workload management, resource governance, and policy tools enable all of these applications to coexist in a distributed Exadata fabric with load isolation, high availability, and guaranteed SLAs.
  • NYC notorious 1800s slum "5 Points"' only remaining "point" is corner of Worth & Baxter: SW corner of Columbus Park: Scorsese "Gangs of NY"9:35 PM Sep 29th from TweetDeck
  • "Nobel vs. Ig Nobel" (http://bit.ly/N1dQ5): JK--Ig Nobel awards: like late Sen. Proxmire's Golden Fleece prizes, minus the federal funding9:27 PM Sep 29th from TweetDeck
  • "Talend Acquires Master Data Management Firm" (http://bit.ly/f9pcg): JK--OK, but Talend can't make serious go in MDM without consulting arm.9:19 PM Sep 29th from TweetDeck
    • JK2---Yes, MDM is a key component of a full data management “stack.” But the notion of a data management “stack” is starting to feel obsolete. The deeper the stack, the more likely you are to get crushed under the sheer complexity and cost of all that software. The more complex your requirements for a “single version of truth” data management capability, the more you want a strategic partner to help you assemble the lean subset of one or more vendor “stacks” to meet your specific needs.
  • "Way Over Yonder In the Minor Key" (http://bit.ly/VVCMZ): lyric: WGuthrie, music: BBragg. Woulda been cool to hear Woody & Billy harmonize.9:15 PM Sep 29th from TweetDeck
    • JK2---Both of them with homely voices, but Billy’s working class English in the lower register, Woody’s middle class Okie in the upper, somehow you know they would have merged into a righteous blend of singin’ transatlantic leftie cuzzins.
  • My initial macro-blog strategy in '04 was quick riffs on the day's news. My initial micro-blog strategy in '09 has been that + quasi-msging8:42 PM Sep 29th from TweetDeck
    • JK2---The day’s news keeps me off-balance and paying attention. Quasi-messaging keeps you that way.
  • @jowyang @rwang0 @charleneli "How Would You Define a Social Network?" (http://bit.ly/N0rnv): JK--Pub-sub P2P identity, content, & status net8:40 PM Sep 29th from TweetDeck
    • JK2---A social network is a) person to person (to person to person to person ad infinitum, building communities of greater or lesser extent), b) publish-and-subscribe (all the persons giving and taking all kinds of stuff), c) identity (all the persons publishing and subscribing their identities, profiles, peeves, passions, etc. to varying degrees), d) content (all the persons publishing and/or consuming content in the form of blog posts, status updates, tweets, podcasts, etc. to varying degrees), and e) status (all the persons publishing and subscribing to alerts re their new content, new identity attributes, new responses to other people’s content, etc.).
  • Try as I may, I cannot work hashtags into my Twitter style. Something about me doesn't want to spend my life stoking search engines.8:35 PM Sep 29th from TweetDeck
    • JK2---In the abstract, hashtags are cool: implementing a shared, user-extensible taxonomy/ontology (“folksonomy”) of key subject domains, to facilitate more efficient search, classification, and aggregation. In the concrete, hashtags are a pain: burdening tweeters with having to remember them or look them up, with having to type them into our tweets in the heat of the moment, and with having to grant them precious real-estate within the 140-char footprint. I rarely bother. There must be a better, more automatic way to implement folksonomies in Twitter.
  • "Mozilla denies it will 'ribbonize' Firefox" (http://bit.ly/3BXRSP): JK--Huh? Did I miss something? Unwelcome new verb? Clippy redux?8:34 PM Sep 29th from TweetDeck
    • JK2---I don’t mind the “ribbon” in Windows Vista or in Office 2007. I simply mind not being able to toggle to a non-ribbon interface.
  • "The multitasking myth" (http://bit.ly/gncCD): JK--My mind's only working on 1 task at time. I arrange physical self to rapidly mind-switch.8:30 PM Sep 29th from TweetDeck
    • JK2---No, I’m not some Star Trek mutant. What I mean is that I feng shui my physical space for max productivity. I pile things with exquisite attention to fast physical search and retrieval.
  • Everything blows over. Today's wind- & sun-stabbed early autumn bluster. Deck chairs. My late afternoon headache. Recollections of August.8:22 PM Sep 29th from TweetDeck
    • JK2---August augurs autumn.
  • Social media & BAM, 2-way visibility: you into customer pain, they into your processes #Forrester BTF (http://bit.ly/2F4Kky) Chicago Oct 8-98:14 PM Sep 29th from TweetDeck
    • JK2--- Visibility is the most disruptive force in the business worlds. Social media provide the most powerful new telescope since Galileo, providing consumers, resellers, business partners, and other stakeholders with deep visibility into how enterprises are serving, or failing to address, their needs. In the process, social media, and other visibility technologies such as business activity monitoring (BAM), are changing the balance of power across organizations everywhere. In this session, Forrester analysts Natalie Petouhoff and James Kobielus will discuss the transforming role of social media and BAM, which are fostering greater engagement among stakeholders, improving customer service, and optimizing value chains. Full visibility is key: into your customers, their pain. Learn how social media and BAM enable end-to-end process visibility and thereby fuel transparent engagement between executives, employees. and customers and other business stakeholders. Understand how social media and BAM contribute to superior business performance by using visibility to clarify points of responsibility and accountability for operational and change processes. Discover how social media and BAM provide the information--voice of customer (VoC) data for social media, key process and performance indicators for BAM-- needed to transform all business functions and foster cross-organizational collaboration, multidisciplinary teaming, and agile response to competitive pressures
  • "Vint Cerf Details Interplanetary Plans for DTN " (http://bit.ly/2Li116): JK--Disruption Tolerant Networking. Old ARPANET dreams don't die.4:21 PM Sep 29th from TweetDeck
    • JK2---Ground control to Major Vint.
  • Every time I refer to the "R" language, I put that letter in quotemarks. Helps me search for it later. Ought to be standard practice.10:08 AM Sep 29th from TweetDeck
    • JK2---Inadvertent steganography. Single-character names hide in plain sight. You humans can easily pick them out of the noise, but search engines can’t.
  • Social media->conversations->reveal hidden affinities->communities->customer loyalty. #Forr BTF (http://bit.ly/2F4Kky) Oct 8-9 Chicago.7:44 PM Sep 28th from TweetDeck
    • JK2---To change the subject just slightly: Natalie Petouhoff and I both graduated from Livonia (Michigan) Stevenson High School. I’m 2 years older than her. Her photo’s in the sophomore section of my senior yearbook. We didn’t know that affinity until the week before last, when we were rehearsing our joint preso long-distance. Idle chit-chat, one thing led to the next, to the “omigod!” revelation. We never actually “met” back then or heard of each other. Though, undoubtedly, we crossed paths many times in the hallways in the 1975-76 school year.
  • RT @lcecere: "A hotel with a view of the parking lot. So much for glamorous life of an analyst!" JK--Glamour? Ha! It's non-stop hard work.7:15 PM Sep 27th from TweetDeck
    • JK2---There’s fun too. But the fun’s always sandwiched into and around the work. At my age, I have no idea anymore what “glamour” is. Then again, nerd that I am, maybe I never did.
  • So far the fall colors are all in the sky. Deep cool orange sunset tonight over northern Virginia. Last week's cricket din a distant memory.7:05 PM Sep 27th from TweetDeck
    • JK2---Must bring sweater to Chicago.
Posted by at

Thursday, October 01, 2009

FORRESTER blog repost Service-Oriented Analytics: Tapping into the Predictive Smarts of Your Entire Organization

Service-Oriented Analytics: Tapping into the Predictive Smarts of Your Entire Organization

By James Kobielus

Advanced analytics is a key competitive weapon of companies everywhere. Visionary organizations are those that take a future-facing, analysis-driven perspective on new challenges. They do this by grounding management forecasts in solid historical information sets, leveraging and extending companies’ existing investments in data mining and predictive modeling.

To become truly future-focused, organizations must build out their predictive muscles through deepening commitment to these and other advanced analytics technologies, which also include interactive visualization, multivariate statistical analysis, text analytics, and massively parallel enterprise data warehousing. However, enterprises must be careful not to adhere to the common practice of implementing advanced analytics tools in tactical, application-specific silos. One of the downsides of this traditional practice is that diverse predictive modeling teams can find it difficult to share their deep domain expertise, best statistical approaches, and most powerful data exploration and visualization features.

How can companies realize the transformative potential of predictive analytics for business success? For starters, you must get rid of siloes that fragment your data mining initiatives into separate camps. You must also build a bridge between your data mining operations and the teams that manage your text analytics, business intelligence, complex event processing, and business process management efforts. And the key approach for silo-smashing is service-oriented architecture (SOA).

At first glance, SOA may seem like a foreign topic to many analytics professionals, but it shouldn’t be. In the broadest perspective, SOA refers to best practices for encouraging greater reuse, sharing, and cross-platform interoperability among key business resources. Typically, one associates SOA with reuse of one specific type of resource: application functionality that is distributed across heterogeneous, networked platforms. Nevertheless, key SOA principles—such as standards-based service virtualization, reuse, brokering, and governance—are as applicable to predictive models as to any other resource that lives online.

Predictive models empower your product managers, marketing specialists, risk managers, process analysts, senior executives, and other personnel with access to sophisticated forecasting, time-series analysis, and scenario-testing tools. Each model is a statistical encapsulation of your business’ current view of the future in some specific application, subject, or decision-support area. Incorporating the expertise of subject-matter experts, these models allow organizations to gauge the potential impact of future projects, campaigns, and other initiatives, and also to adjust execution of these initiatives in mid-stream.

Predictive analytics need not be a purely blue-sky planning tool. This technology can sit at the core of your SOA strategy, and leading-edge enterprises are in fact doing that. Companies in such verticals as financial and telecommunications are embedding predictive logic deeply into data warehouses, business process management (BPM) platforms, complex event processing (CEP) streams, and operational applications.

What will it take for your company to fully align your advanced analytics efforts with your SOA strategy? Most important, Service-Oriented Analytics requires an executive-level commitment to becoming a predictive enterprise on all levels. From the perspective of your existing predictive modeling teams, this will require an ongoing focus on collaboration across business, function, and subject domains. You should create a culture and offer incentives that encourage modeling professionals to reuse each other’s expertise on problems that cross multiple domains.

Reuse is everything. Here are some high-level guidelines for establishing a reuse-friendly Service-Oriented Analytics practice in your company:
  • Reuse modeling best practices: Start by consulting Forrester’s report on setting up a Business Intelligence Solution Center (BISC), which we defined as “an institutional steward, protector, and forum for BI best practices.” Given that predictive analytics is a key segment of BI, you will find it necessary to incorporate this technology into your BISC’s scope.
  • Reuse modelers: Cultivate a professional cadre of predictive modeling experts who are more than just wizards in advanced statistics and mathematics. Encourage subject-matter experts in all business areas to undertake predictive modeling projects and to team with modeling experts in other projects, applications, and business units. Provide incentives for modelers to regularly move between business units and subject domains, thereby spreading their expertise throughout the enterprise.
  • Reuse models: You should begin to investigate predictive-model governance tools, which support version control, check-in/check-out, and other controls over models created in and imported from diverse tools. Model governance tools—from vendors such as SAS, SPSS, and KXEN--facilitate reuse, consolidation, combination, and cross-synthesis among disparate models. And you should also investigate options for embedding predictive models within BI, BPM, and other operational applications, thereby leveraging your growing analytic asset into new deployments.
As you deploy predictive models into operational applications, you should provide other applications with SOA-based access to them through Web services, Web 2.0, and other standardized interfaces. In that way, you will be creating a critical bridge to your application development teams and be creating a thoroughly Service-Oriented Analytics environment.

Your first steps down this road are clear. Forrester will be happy to assist you in developing a detailed enterprise roadmap for Service-Oriented Analytics. In so doing, you’ll be better able to tap into the future-facing analytic expertise that exists throughout your business.
Posted by at
Subscribe to: Posts (Atom)