My connection with AR is not completely severed – I will continue to lean on Gartner, Forrester, IDC, etc., for market intelligence.  But strictly in terms of AR, I am off the beat. It has now been over 15 years since I shifted out of product development/management into the analyst realm. How are things different than they were 15+ years ago when I first stumbled into IDC as a development tools analyst?

The More Things Change, the More Things Change, 5 Things Specifically

1. The Digital Analyst:  The Internet and Mobility revolutionized so many things, including AR.  Today analysts’ phone numbers, emails and Twitter handles litter my smartphone, and tweets and blog posts pop up like weeds in an untended lawn during a dry July, and social web sites and blogs crawl with contacts out of the analyst ranks.  Hey, there were online communities back pre-1995; remember Compuserve?  It is still there!   Back then if you wanted to find tech influencers, there you went.  Today you go to FB, LinkedIn, or all kinds of “open” .orgs or other communities sites – and analyst sites of course which were not existent pre-1995.  This has led to the near death of analyst firm physical libraries.
2. Coverage Overlap:  Because Internet/mobile enable and reward convergence, the circles of coverage are not as discrete as they were during the “client/server” era.  Overlaps due to the nouvelle technologies have forced analysts to think more outside their historic boxes.  Analysts who stubbornly remained totally in a silo have lost some ground.
3. Analysts Doing Research, Researchers Doing Analysis:  Pre-1995 analysts were either IT advisors or statistical geeks, but those boundaries are blurred.  The qualitative analysts are somewhat more quantitative, and vice versa.  There is still room for improvement on both sides of that equation though.
4. Sprouting of Small Analyst Firms:  Pre-1995 there was the big two, but today Forrester is usually included as tier-1, so now there is the big three – plus a few thousand others.  Today there are many independent analysts and very small firms, enabled by social media and the ability to type quickly on a mobile device – I think small fingers are a help to indie analysts
5. PR Masquerading as AR:  Pre-1995 AR was more often aligned with product management, product marketing, sales and/or strategy/executive.  Today AR is more aligned with PR, and often treated like an adjunct to PR.  In my opinion, though, the best AR programs are balanced, and treat all of these constituencies with equal attention and budget.  Show me an AR program that follows in the trail of PR and I will show you a vendor who is unnecessarily losing battles of influence in the sales cycle.

But Some Things Stay the Same… 5 Things Specifically

1. Pay-for-Play Remains Unchanged:  Several analyst firms have addressed it,  but blatant quid pro quo has not disappeared despite the occasional hypocritical piece written about it by someone in the media.  Heck, some “analysts” are nothing more than 3rd party marketing sites for their vendor customers.   Some day maybe I will name names, but suffice it to say that money to say something nice about vendors is still a legitimate business model.  Gartner remains quite pure though, even if they treat vendor customers like customers today, more respect and less disdain – thankfully.
2. Relationships First:  A glass of wine still beats a Tweet, or email, or other electronic communication.  You want to form a trusted relationship with an analyst?  Spend some time face-to-face, make the human connection.  I don’t care if you have the most sophisticated social media listening technology in the world, it can’t compare to a handshake, a smile, and a genuine conversation.  Don’t forget we are in this together.
3. Research wins:  Not every day, but the analysts and analyst firms who really have a survey leg to stand on will stand the longest.  Those that do their homework and share their findings in easy-to-digest fashion were around pre-1995, are around today, and will be around when the “Cloud” is considered legacy.
4. Gartner is still #1:  Gartner’s imminent demise has been predicted for years. Some have said that peer networks or social media will destroy Gartner. Sorry, I don’t buy it.  Gartner seldom falls asleep at the switch, and they still possess, by far, the largest cadre of strong analysts.  Follow how Gartner has adapted over the past 5 years and you have to take your hat off to them.  But damn they are expensive.
5. Get a Haircut, Wear a Suit, Take a Shower:  Okay, after lauding Gartner I have to make a complaint that is 15 years old:  some Gartner analysts are poor consultants, so spend your SAS dollars carefully.  Some of them never acquired the chops that professional consultants possess.  Gartner Consulting has those chops, but many Gartner analysts lack those chops.  I think Forrester is in the big three today largely because they hired analysts who could truly consult; who could dress, converse, carry themselves professionally and socially.   Some of those excellent Forrester analyst/consultants burned out because Forrester realized how yummy the margins were on a $10,000 per day consulting gig.  Some of those analysts went out on their own, figuring they could charge around the same amount, and sop up the overhead for themselves.  Maybe that is why some Gartner analysts remain untamed as consultants, but if I were Gene Hall I would require my analysts to attend consulting finishing school.

And with that, buh-bye AR, and thanks to the few CMOs out there who actually appreciate AR – you are an unfortunate rarity.

Every AR person knows that many of the most influential analysts in the information technology industry work at Gartner.  But analysts are not the most influential influencers out there, peers are – IT buyers and practitioners most trust the insights of other IT buyers and practitioners who have been through similar buying and implementation processes. The historical blockades to peer-to-peer exchange, however, have been (a) finding qualified peers and (b) providing a safe harbor for peers that prefer to remain anonymous in order to participate.

Some influence management thinkers have recently opined that, on the back of social media, Gartner and other top industry analyst firms could be rendered obsolete by “peer networks” or “expert networks.” In fact, already several quasi-analyst and analyst entities are heavily using the digital networking-centric approach, most notably Focus and Wikibon. And other top tier analyst firms like IDC and Forrester have tapped into the community network effect, for example see IDC Insights Community.

Gartner, however, isn’t so easily leapfrogged. Witness Gartner’s Peer Connect, a tool available only to Gartner for IT Leaders customers offering a private, peer-to-peer information exchange environment, beyond the reach of vendors and consultants. Even Gartner’s own analysts cannot participate directly, though they may introduce topics of discussion.

Peer Connect leverages Gartner’s ability to tap into a plethora of qualified IT peers out of its existing customer base – at no additional cost to its IT customers. Gartner has dealt with the privacy concerns of some of the IT peers by offering an anonymous self-profiling option that still exposes industry, company size, job role, vendor experience, plus past and current projects. Thus, a Peer Connect user could search for a peer with a profile that matches “enterprise architect in healthcare with revenues over $250m currently or recently involved in an Oracle CRM project.” Once an appropriate match is found, the resulting interchange may be conducted email-to-email if both parties are not anonymous, but if either party wants to remain anonymous the interchange takes place through protected digital collaboration where the email addresses are blocked.

Gartner doesn’t hard-sell Peer Connect. There is a modicum of marketing collateral for Peer Connect on Gartner.com, and it is buried in the general concept of “Peer Networking.” The Gartner IT salesperson may certainly discuss Peer Connect during the sales process, but Peer Connect isn’t a prospecting feature – Gartner still leads with its research and analysts. The Peer Connect community is entirely self-selecting; there is no pressure for a Gartner for IT Leaders customer to join the Peer Connect community.

The genesis of Peer Connect goes back to 2006 when Gartner IT end-user surveys revealed the desire of Gartner’s IT customers to directly share each others’ experience and expertise. Specifically, users wanted to exchange tactical experience, and to obtain help answering detailed questions that an analyst would not typically be expected to answer. The initial Peer Connect rolled out during 2007 giving IT customers an exchange mechanism with which to register, profile themselves and to list projects they worked on – and to search for those individuals that could answer specific questions. Since then Gartner has constantly enhanced Peer Connect, adding what we now refer to as social media features, and has improved the search, profiling and security aspects. Peer Connect has not only been popular with Gartner’s IT users, it has also driven research topics for the IT Leaders platform.

What does this mean to an Influence Management or Analyst Relations professional? In a B2B community like enterprise IT, peer communities provide a layer of insulation for actual IT practitioners beyond advertising, blogs, research, media and other 3^rd party opinions. Like Gartner, vendors should consider embracing the concept of peer networks by giving customers and prospects direct access to one-another without interference by the vendor. Vendor-oriented user groups have been around for decades, and are the progenitors to peer networks. Perhaps the most notable example is IBM Share which goes all the way back to 1955 (!). The vendor’s attitude should be to support, facilitate and where it can, without being intrusive, learn from these groups – but providing a safe harbor for honest exchange and sharing of best practices counts as the primary goal.

What of other research and analyst firms that use a similar approach? To me there is quite a difference between “expert networks” and “peer-to-peer” simply because “experts” are difficult to qualify. How does an expert receive the title of “expert?” Is it self-proclaimed? Is there a hidden agenda to sell the “expertise?” Are “experts” blessed by some 3^rd party “expert of experts” who in fact has little expertise in any particular area of IT?

As suggested by the Enterprise IT Influence Mountain at the top, IT people trust other IT people first and foremost. A tip of the hat goes to Gartner who understands that IT expertise doesn’t begin and end with analysts, and for providing a safe harbor of exchange for its IT customers, helping overcome the “ease of access” hurdle that typically limits sharing between peers.

1. Early Stage Vendor Leans on Agency PR to do AR: The analyst relationships required to establish interest in an early stage vendor among the sometimes cynical lot of industry analysts requires trust – where the AR person obtains the proverbial “seat at the table” with the analysts. PR agencies often over-depend on social media, hard pitches, or analysts who only play to the media. The by-product is the start-up flies under the radar of the analysts with sway with real customers or potential partners. Emerging vendors are better off with a part-time AR contractor, or letting a product manager or product marketing handle the AR work; it more important to be real than over-hyped during the early stages.

2. Too many AR bodies in a single market: I know of a vendor that competes in a single market, but has 6 FT AR personnel on staff. They have splintered their market so far that they end up confusing analysts, while spending more on headcount than any AR ROI metric could support. For most IT markets a vendor vies in, you need one AR lead, and one or a half-time backup if it is complex or covered by many analysts, or if you believe in back-ups (I do). The revenue and/or geographic diversity has to be there to support that additional requirement for AR bodies for a single market.

3. The wearing of two hats – AR and PR:   Pretend an AR team has 9 full time bodies, but 6 of them have PR backgrounds, and the PR team constantly dips into that well of expertise to handle PR body shortages – which seem endless. On paper you have 9 AR heads, but in reality you might have 1 or 2! Why 1 or 2? The remaining AR types are so over-burdened by covering for those donning PR frocks, the AR types fry. I know of several cases, names omitted to protect the guilty, where this was rampant, and the AR program struggled even on blue sky days.

4. Drawing an exclusive correlation between revenue and AR headcount: Often AR team sizing discussions begin and end with relative revenue. However, a $30b vendor with a single product line sold primarily through channel might get away with a two decent AR people. A $10b vendor with several billion dollar product/service lines, mainly selling direct and taking AR seriously may legitimately require 7 AR people. Revenue is a factor, but it is not the only factor.

5. Forgetting about corporate: Pretend an IT vendor earns $5b in annual revenue, while competing in 3 distinct markets. One might guess that 3 to 5 headcount is needed, one lead for each market and maybe one or two junior folks acting as back-ups and handling operations. Ah, but what about STRATEGY?! What about the most influential analysts in the world who don’t just focus on a single market, like Crawford Del Prete of IDC, Ray Wang of Constellation, Rob Enderle, a bevy of the Gartner fellows, Ted Schadler or Chris Mines of Forrester, etc.? What about green, giving, R&D, sales, marketing, finance, operations and board structure and performance? An effective AR program has somebody, often the lead or one of the senior-most AR types, who handles the “corporate” role. Often that “vision” rating depends or is influenced by the effectiveness of a corporate AR practice. If the analysts never hear from the senior executives about strategy, vision and execution at the corporate level, the analysts will assume the worst (i.e. there is no strategy, vision or execution), which is never the right approach.  So the worst practice is to not factor in headcount to handle the corporate AR functions.

An AR friend of mine recently asked if there was a formula or guideline for right-sizing AR teams.  Unfortunately, there is no simple formula to the staffing question, but off the top of my head there are five factors that contribute to determining a reasonable headcount model for right-sizing an AR team, whether full-time employee or 3rd party agency or contractors:

1. Enterprise focus: Vendors that have more of an enterprise focus will tend to need to more AR staff per revenue dollar than consumer-oriented IT vendors. E.g., IBM, which is essentially 100% enterprise focused, would likely have more AR headcount/revenue than say Acer which sells plenty of PCs to consumers.

2. Sales model: Vendors with a higher percentage of sales through a direct sales force will tend to require more AR headcount/revenue dollar versus vendors that depend heavily on an indirect channel sales model. Why? Because direct sales is more directly impacted by IT buyer influencers, like Gartner and Forrester, than through the channel.  There is nothing like your own sales force calling the AR team and saying, “Hey, Gartner is in this deal, and…”

3. Product/service Diversity: Vendors with more diverse product lines tend to require more AR headcount/revenue than vendors with vendors with less diverse product/service lines. For example, say there was a pure-play storage vendor, and competing vendor who offers both storage and security.  Everything else being equal, the latter firm (storage and security) would tend to require a few more AR headcount because of a more diverse product line.

4. Geographic Diversity: Vendors with a true global reach will tend to need more AR personnel, just for geographic coverage purposes. E.g., say there were two 2 billion/year revenue enterprise-oriented vendors with similarly diverse product lines and similar sales models; but one has geo splits of 60% US, 35% EMEA, 5% APJ. The other vendor has 40% US, 35% EMEA, 25% APJ – you might expect another headcount to support APJ at the latter vendor.

5. AR Program Model: I like to think of AR programs as coming in “small, medium and large” models. That doesn’t indicate the headcount, but indicates the role of the AR team in the overall context of the vendor, thus:

· Small – An AR program that basically chases PR: AR has little budget, analysts are treated as an extension of media, there is little market intelligence buying; minor analyst event spending; little to no spending on analyst advisory or custom research projects. The “small” AR model usually has AR reporting into PR, or into a very PR-oriented head of Corporate Communications.

· Medium – An AR program focusing on core AR work versus media: MQs, Waves and market share are treated seriously.  However, analyst advisory and research projects are relatively few and far between, and AR plays a minor role in co-marketing, events, or market intelligence buying. Most of the spending with analyst firms does not come from the AR budget, though the AR team has a enough discretionary budget to capture the interest of analyst firm account managers.

· Large – AR is treated as a strategic element of sales/marketing: Perhaps more than 50% of the all spending on analyst firms emanates from the AR budget, and the AR team carries strong influence on other analyst-related spending like events and custom research projects. AR is directly involved with market intelligence buying, is front-and-center in analyst firm negotiation, and senior management and the strategy team at the vendor are willing to engage with top analysts.

Naturally the “large” model will yield more AR headcount/revenue than the small.

In summary, IT vendors that are (1) enterprise-focused, (2) largely use a direct sales model, (3) have diverse product/service lines, (4) are geographically diverse and (5) take AR seriously and empower it as a strategic function should carry more AR headcount/revenue than similarly-sized vendors that do not exhibit that criteria.

This simple 5 criteria analysis only scrapes the surface.  Plenty of other factors go into determining headcount, such as how important are metrics to the AR team?  Does the AR team maintain internal and external web portals, or participate actively in social media (versus just as tracking tool at vendor AR events)?  Does the AR team travel extensively?  Is the team heavily weighted AR experience-wise to one end of the scale or the other (in theory more experienced AR people should be a little more efficient and effective per headcount)?  Does the AR team work with a wider set of influencers and experts, or only strictly industry analysts?

Next post will focus on 5 gotchas for wrong-sizing AR teams.

I challenge you to find V3 on the Web:  The URL is actually not www.v3.com but www.valleyviewventures.com – like International Data Corporation goes quite strictly by “IDC” these days, but the URL just hasn’t been changed yet.  You will not be awed by the V3 web site, but that doesn’t matter one iota to Fred Abbott, V3′s founder, who says with utter sincerity, “It’s all about the analysts.”

Fred started V3 in 2002 out of his home after working in sales and sales management at places like Gartner, Giga, IDC and Hurwitz for two decades.  Fred didn’t have any grandiose business plan, didn’t crawl up Sand Hill Road with a Powerpoint and an open palm.  Instead, during the dot-com bust a few analysts who had gone independent were looking for someone else to handle the sales and contract management aspects of their fledgling businesses so they could concentrate on being analysts.  Ever-well-connected Fred agreed to help them out – simple as that.

Five years into V3s slow-growing but flourishing role of acting as the business development and contract management arm for independent IT industry analysts, Linda Ziffrin joined Fred at V3.  Linda’s background is in selling HR applications, and she spent seven years at PeopleSoft during the 1990s. Fred had the desirable problem of too much work to do, so Linda stepped in to help.  Today the amiable, easy-to-do-business-with yet highly effective Fred and Linda, a.k.a. V3, act as the sales arm, either wholly or partially, for some of the most famous and influential analysts in the world including: Rob Enderle who is the most quoted industry analyst in the world; Charles King who writes much of and facilitates the widely distributed and read Pund-IT, Roger Kay ranks as one of the most oft-quoted PC experts; semiconductor guru Jim Handy; and Altimeter Group which houses an exclusive club of top analysts such as Charlene Li, R “Ray” Wang, Jeremiah Owyang and Michael Gartenberg.  This is just a partial list of V3′s affiliates.

Unlike the Big Three who drive their brand at most opportunities, Fred, instead of looking for scale and geographic reach, looks to work with analysts “that could use a little help” and particularly with analysts who like to collaborate with other independent analysts.  Check out Pund-IT at www.pund-it.com and you will find many of the other analysts in the V3 sphere authoring IT industry perspectives.  Fred and Linda also prefer to work with analysts who prefer to market themselves rather than a brand.  Yes, Rob Enderle’s company is “Enderle Group” but everyone in IT has heard of Rob Enderle.  Similarly Charlene Li, Ray Wang, and Jeremiah Owyang made Altimeter Group, not the other way around.  Do many know that Roger Kay’s official research company name is Endpoint Technologies Associates?  Not sure, but many know of Roger Kay.

Just how important is V3?  Well, if your primary metric of “influence” are press citations by analysts, then we might have to change the “Big Three” into the “Big Four.”  Mind you that I believe that the realm of AR has become too PR-centric, probably because the other arenas of influence plied by industry analysts, namely IT buyers and investors, treat interactions with industry analysts with strict privacy.  It is virtually impossible to measure, for example, the number of short list placements an AR team won for a vendor by doing better work than the competition.

But regardless, the “analyst in the press” metric is certainly helpful in measuring analysts in the media communications channel, and here the V3 affiliated analysts shine.  According to tracking of “tech analysts in the media” by IT Memos’ reports (see http://memos.itdatabase.com published on April 6, 2010), V3 had the #1, #3, #6, #9 and #10 most quoted analysts over the past six months.  Gartner had three of the top 10, Forrester one of the top 10, IDC zero.  In fact, if you look at the 50 most quoted analysts from the IT Memos report V3 affiliated analysts reaped 1196 quotations, Gartner had 977, Forrester 612 and IDC 368.

I cannot vouch for IT Memos’ technique.  For example the analyst quote counts were associated with an analyst, and a firm like IDC gets quoted extensively for market share and forecast statistics without naming an analyst.  In fact, another metric IT Memos tracks called “Firm Mentions” resulted in a Gartner runaway with 6978, IDC was second with 3935, and Forrester third with 3308.  The next closest IT analyst/research firm in terms of mentions (IT Memos included firms like “Goldman Sachs” in their list – not sure that was intended, but I am ignoring those counts) was iSuppli way down at 1304.  So at least in terms of analyst firm branding as measured by firm mentions the Big Three are indeed the Big Three.

But in Fred Abbott’s model, who cares about firm mentions?  The V3 brand doesn’t show up at all but the affiliated analysts are at the top of list!  Fred and the independent analysts are doing things differently, and it is working quite well thank you.  Give credit where credit is due:  V3 is right up there with the Big Three in the named analyst quotes metric (note that IT Memos didn’t know that many of the top analysts they listed were V3 affiliates, or at least didn’t report it that way).

Hey industry analyst, tired of getting paid a straight salary for traveling extensively and performing pressure-packed consulting sessions that your firm charges out at $1000/hour or more?  Even if you are making a healthy salary, your cut is maybe $100/hour pre-tax.  These consulting sessions yield gross margins that would raise the eyebrow of even a Larry Ellison.  Analyst, think you have enough connections, credibility and savvy to go out on your own, and either gain better control over your time, your research agenda or how much of that consulting day fee you actually keep?  Fred and Linda will charge you straight 15% commission for your work, and you don’t have to stick your nose into too many of the business details, you get to primarily focus on being a superior analyst.

Do you think you have what it takes to impress Fred and Linda enough for them to take you on?  Fred doesn’t ask for a business plan either.  He will talk with you, help you figure out your approach for no fee, and he and Linda only get paid if you are successful.  As Fred said in my interview with him, “With a new analyst we can try it out for a few months and see if it works.  What we try to get them to understand is that it is all about them, not V3, that marketing themselves is what is important.”  I find it deliciously ironic that V3′s affiliates include some of the most sophisticated analysts at wielding digital marketing and social media, yet Fred keeps V3 down to a bare minimum – no V3 fan club on Facebook last I checked.

Just as open source bedeviled the purveyors of proprietary Linux and Windows, Fred’s antithetical approach has to have some folks at the Big Three scratching their heads.  Despite dubious attempts to tie-up analysts with non-competes (which do not apply in California so not coincidentally Cali happens to be the home of many of V3′s top affiliated analysts), to mix in events and consulting, to outsource some of the lower end research jobs to places like India, to acquire strong niche firms, the Big Three have not undercut the quiet march of V3 and independent analysts.  V3′s straightforward, honest approach has turned the industry analyst community 90 degrees.  Why?  How?

“It’s all about the analysts.”- Fred Abbott, Founder and President of V3.  How refreshing.

But I didn’t have the time or money to work with EDC, so I cheated instead:  The other constant at least here in the USA in terms of metrics for the developer community is DICE – http://www.dice.com which has been the most prolific job site for IT types since, well, also as long as I can remember.  In a fit of curiosity I developed a long list of IT related keywords around technologies, applications and roles mainly, and put them up into the DICE jobs search in mid-December, 2009.

The results showed a huge gap between hype and reality in the realm of common, everyday IT job demand:  Job counts were quite high for Java, SQL/databases, C-variant languages, Windows, Unix particularly Linux, .NET, Web development, enterprise applications, testing/QA and security.  Mobile and “Cloud?” job demand, however, is not anywhere in the same ballpark as the more ensconced technologies.  Maybe DICE isn’t the best place to post or look for mobile-related or Cloud jobs, but I am guessing it offers a pretty good pulse for the industry.

That doesn’t mean that eventually the hype won’t catch up and change reality, after all in terms of DICE jobs “mainframe” is now about the same as “mobile,” though COBOL still eclipses Cloud and roughly ties SaaS in my set of searches.  Between meet-ups, tweet-ups and analyst firms offering services to help analyst relations teams deal with the analysts (talk about the fox in the hen-house phenomena), well, I just like to throw some harsh light of reality into the miasma of IT silliness.  I figured an analysis of job counts from DICE would suffice on that front for now.

What types of IT personnel, mainly in North America, are being hired?  During that day in December DICE had 51,897 jobs listed, a healthy sample.  Note that DICE has jobs for all kinds of IT, corporate, SMB, consultancies and supply-side.

I used 38 technology oriented search works plus 15 role/lifecycle search words, and I limited the search to job postings over the past 30 days.  Note that I used the raw search, and did not depend on the tags associated with the jobs because the tags were woefully inconsistent.  I know that this “research” is not scientific and that there are all kinds of IT-oriented jobs sites out there, including many focusing on a particular locale, technology or ecosystem.  Still, DICE does have a strong cross-representation of IT-related job postings, the best I could find.

Below I offer a synopsis of findings based of the searches in various IT job categories, and if you really are interested I’ve included many of the actual job count lists at the bottom of this post.  Finally please note that in some cases where the job counts were very low I simply left the technology off the list.

Conclusions

Languages:  Java and the C-variant languages continue to the development languages/environments of primary demand.  In terms of “authoring” (not exactly “programming” but certainly counts as “development”) HTML and JavaScript remain pre-eminent with other scripting languages like PHP and Python filling out the roster.  There is still a reasonable smattering of demand for legacy language skills like Visual Basic and even COBOL.

Platforms/OS:  Windows still wins but barely against Unix, and in terms of Unixes Linux is by far the most popular variant.  Solaris of the proprietary Unixes shows up on the most job descriptions, and at the other end of the proprietary scale HP-UX has the lowest demand for DICE hosted jobs.  In terms of middleware .NET holds sway with more than twice the jobs listed than WebShere with WebLogic not too far behind in 3^rd; WebSphere + WebLogic almost balance the middleware job battle between J2EE and .NET therefore. Way below those behemoth platform categories, Force.com (Salesforce.com’s Cloud platform) showed rather well, but hasn’t quite caught up to the Mainframe for job demand.  Tibco followed, and the “Cloud,” Oracle’s Fusion, NetWeaver, Amazon (a Cloud entry) and LAMP were at the back of the pack.

Data:  Microsoft’s SQL Server, XML and Oracle’s database as a triumvirate, in that order, dominates this space for jobs at DICE.  It looks like there is still tons of demand for SQL skills out there.  DB2, mySQL and Sybase were seated next to each other in the second tier of job demand on DICE in the database arena.  There were other databases in the job pool but the numbers were quite low.  Similarly, while I didn’t include the various BI and data integration tools out there, that list was too long, fragmented and the numbers quite low, Informatica showed well with slightly over 1,000 jobs, and Business Objects and Cognos made the loudest showing in the BI space with both around 1,000 jobs posted.  At the low end of the database/BI scale was Ingres with 13 jobs listed, so in this case the first shall be last.

Applications:  While the general term “web application” beat any individual application brand, when you add it all up there was still more interest in “enterprise” applications than web apps per se.  We know that SAP leads in enterprise applications in EMEA, but that Oracle with its many acquisitions leads here in North America, and that showed up in the job counts.  In fact Microsoft Dynamics had more jobs showing up that SAP.  But regardless, IT looks for people with Oracle, SAP and Microsoft application skills in large numbers.  Perhaps shocking was the low demand in mobile related applications – the general “mobile” category landed well behind the “enterprise” and “web” categories, and Blackberry and iPhone barely made a dent in this IT job pool.  Guess all those purported iPhone developers are bootstrapping in Silicon Valley apartments.

Roles/Lifecycle:  Though the term “programmer” has fallen nearly entirely out of vogue, “engineer” and “developer” are roles that run strong on DICE.  However, “test/testing” and “quality/QA” weren’t far behind in aggregate.  Out of the over 50,000 jobs on DICE, ‘engineer’ or ‘developer’ showed up in aggregate over 33,000 times.  But those testing/QA types showed up over 30,000 times.  If there is one other surprising conclusion to reach from this exercise besides the fact that Cloud and mobile seem relatively stuck in the hype phase its that that quality-oriented jobs are very important to IT these days. In the next grouping comes “network” and “database” and “security” all accounting for slightly over 10,000 jobs using those keywords.  “Analyst” and “business analyst” are not far behind, not far below 10,000 jobs.  “Architect” and “Administrator” fall around the 5,000 job mark.  If your specialty is UI (user interface) design or development, there isn’t a raging demand for you relatively to these other role/lifecycle categories.

Job Counts

Development Languages/Environments (number of jobs – language or environement; note that a job can list more than one language).

1. 9,737 – Java
2. 4,797 – HTML
3. 4,682 – C#
4. 4,289 – C++
5. 4,289 – Javascript (actually tied for 4^th with C++)
6. 2,510 – AJAX
7. 1,638 – PHP
8. 1,298 – Python
9. 1,001 – Visual Basic
10. 831 – Eclipse
11. 478 – COBOL
12. Note:  Could not perform a legit search for “C” – it turned up too many other jobs that were not development oriented

Platforms/Operating Systems

1. 7,932 – Windows
2. 7,583 – Unix
3. 6,192 – Linux
4. 5,589 – .NET
5. 2,048 – WebSphere
6. 2,005 – Solaris
7. 1,400 – WebLogic
8. 1,176 – Mainframe
9. 968 – AIX
10. 803 – Force/Force.com
11. 601 – TIBCO
12. 387 – Cloud
13. 385 – Fusion (includes 184 Cold Fusion jobs, so Fusion alone would be 201 maybe)
14. 380 – LAMP
15. 280 – NetWeaver
16. 275 – HP-UX

Data/Data-related

1. 8,216 – SQL Server (Microsoft)
2. 6,061 – XML
3. 4,983 – Oracle database
4. 1,744 – DB2 (IBM)
5. 1,666 – mySQL
6. 1,148 – Sybase
7. A too long list of other databases and BI tools with generally low numbers, though Informatica, Business Objects and Cognos were above 1,000 jobs each

Applications:

1. 9,817 – Web application
2. 6,613- Oracle Applications
3. 4,163 – Dynamics (Microsoft)
4. 3,709 – SAP Applications
5. 1,453 – Mobile
6. 1,184 – PeopleSoft
7. 423 – Blackberry
8. 267 – iPhone

Roles/Lifecycle

1. 25,168 – Application
2. 19,438 – Test/testing
3. 18,850 – Engineer
4. 15,862 – Developer
5. 12,342 – Database
6. 11,564 – Network
7. 11,345 – Quality
8. 9,459 – Analyst
9. 7,054 – Business Analyst
10. 5,417 – Architect
11. 4,225 – Administrator
12. 2,682 – Programmer
13. 1,622 – UI/GUI

Check out the results at:

http://www-03.ibm.com/press/us/en/pressrelease/29168.wss

IBM has secured more patents than any other company in the world for the 17th straight year.  While certainly the numbers impress, 4914 patents, more than 1300 than the next closest firm, what seems most compelling to me is the breadth of patents and research initiatives as well as a wider perspective about research in general.  For example IBM also publishes a journal of its research efforts that do not result in patents, which is both (a) generous sharing of research investments but also offers IBM (b) limited protection of their IP.

So view the 4914 and 17th straight year as just some of the indicators of IBM’s level of commitment to innovation.  Frankly I am equally as impressed by IBM’s new program of sharing their approach on innovation as a service to clients.  If you have scored the most patents in the world for 17 straight years you should be sharing that “service” with clients and partners.

Aligned with the “Smarter Planet” macro-theme, IBM Research doesn’t purely focus on IT, but an entire range of initiatives that reflect IBM’s business orientation.   Thus IBM looks at the potential application of IT in a socially and business driven manner, not just from the technology out.  Look at examples such as automotive telematics and you will get the idea.  While IBM often touts its petaFLOPS recordbreaker Roadrunner computer with the press, and plans to actually rollout Roadrunner on a limited basis by 2011,  see here, for most of the industry and wider world IBM’s more vertically-focused efforts will provide actual commercial and human impact sooner.

Mind you “most patents” does not necessarily equate to “most innovative.”  Even on IBM Research’s own blogs do they bring into question whether the pressure to reach patent is useful or not, and they discuss the plethora of unnecessary patents in the industry which slows down the patent process for legitimate patent seekers.  However, it is difficult to argue with such a huge lead in number of patents.

These patent counts certainly bring into question Information Week’s 2009 most innovative company list, where Apple shows up as number 1, Google number 2, Microsoft 4, IBM 6 and Samsung (#2 in patents behind IBM) all the way down at 16.  Apple and Google may be the most visibly innovative at the consumer level, but in terms of IT industry R&D spending and patents, IBM (#2 in spending, #1 in patents) and Microsoft (#3 in patents, #1 in spending) should have been given a more in-depth analysis.  If innovation is a popularity contest, well, okay, Apple and Google. But innovation insinuates itself in all kinds of subtle ways, not just in your phone or music players or your Web search or mapping technology.

Regardless, congratulations and thanks to IBM who has been setting the bar for on-going commitment to innovation in the IT industry, and one could argue the world, for basically two decades.  And thanks to Microsoft who isn’t very far behind, who actually spends more on R&D now than any other IT vendor, and has, despite facing financial pressures really for the first time in its existence over the past years, actually significantly increased its commitment to research.

Yes, I am going to express my view on Cloud and Virtualization.  Everyone else has, why not me?  Common wisdom is that the big enterprise computing innovations for the twenty-tens will ride on the backs of Cloud and Virtualization.  Fortunately, I have been exposed to the Cloud and to Virtualization longer than most, all the way back in the Cro-Magnon Computing era of IT, during which I was born as programmer.

In fact, I figure I was one of the first to actually use the Cloud.  Was I an early beta user of Salesforce.com?  Nope, that happened 10 years ago, late-to-the-game of the Cloud era.  I am talking Cloud circa 1973.  We used National CSS, at least I think that was the name, a timesharing facility out of Stamford, CT or near there.  On our remote terminals we would load a program, load data, and run the workload (we called it a job).  It would spit out a report that we could view online or we could print.  The NCSS “data center” was 200 miles away roughly from our downtown Boston office.  Is this a whole lot different than what you can do with Amazon EC2?

Was it elastic?  Who knows.  It ran the jobs fast enough for our satisfaction.  If you needed more “elasticity” you would work overtime and run the job at night, the increase in speed was palpable.  Or you pay a higher rate for more “real memory.”  We ran I/O heavy jobs at night.  In fact for the super heavy I/O jobs somebody would drive down to Stamford with tapes.  Or you could mail the tapes, or courier them if you needed it done fast (FedEx before FedEx).  Elastic? Yes.  Dynamically Elastic?  No.

Was it secure?  We thought so.  There wasn’t much of a public network 36 years ago, and not nearly as many disturbed geeks who took pleasure in attempting to break through security walls, so the hacking risks were way lower.  But I suppose there were some risks, and I believe some kind of encryption was a choice if you were worried, you would pay extra for it.

Was it multi-tenant?  Sure was.  All kinds of people used that computer at the same time.  We were running batch jobs (though viewing output online wasn’t exactly batch, we could do that too).  But for the era, sure enough, it was multi-tenant.

Was it pay-per-use?  Sure was.  At the end of every job it would tell you how many CPU and I/O units were used, and there was a rate card.  You could check it out on the monthly bill.  They had a nice sliding scale on the rate, the more you used the less you paid per unit.  Seemed fair enough.

Eventually I found out the mainframes and operating systems NCSS was using, but we ran our jobs for months before I knew what was in the “glass house.”  After I knew I didn’t care much because that computing power was in the “Cloud” though I didn’t have a name for it.  Our job was to run those jobs.  It worked, and apparently worked for others.  In fact they had a library of programs (called utilities in those days) that you could use for free, of course the CPU units and I/O weren’t free.  But you could load up from that utility library and use the utility was free.  Web Services?  Not exactly, but not far removed.  There was zero inter-network program communication going on, at least not at the application layer.

There you go, Cloud, 1973, read it and weep you Web 2.xers.  Tweet that if you feel so compelled.

On to virtualization.

My 2nd real IT job was at an electronic financial services company during the early-to-mid 1980s that Chase acquired.  Not only did corporate banking apps for Chase customers run on our systems, but nearly 80 other banks’ apps, and their thousands of customers’ client-sides, ran on our systems.  It wasn’t core banking apps mind you, no DDA for 80 banks and Chase, but they were important apps for treasury/cash management types. Remember the concept of “ASP” (Applications Service Provider) that was so hot in the late 1990s and early 2000s?  Remember Corio, RIP?  Chase was an ASP 15 years before Corio was a glimmer in their VC’s eyes.  All those customers, at Chase and other banks, accessed THEIR applications and data through a virtual user environment – yes, a form of virtual desktop if you will.

On the server side we used virtual machines too, running horizontally and vertically – vertical in the sense of isolating business functions from one another, horizontal to provide failover, load balancing, and more protected environments to house the most sensitive data and transactions.  We had a management layer to track and resource balance that miasma of server virtual machines that ran on, you guessed it, its own pool of virtual machines.  Oh, and we had dedicated, secure lines to various money moving facilities, so we routed and balanced and secured the incoming traffic from customers, and did the same on the outbound side to SWIFT and the other money/transaction switches.  Edge routing anyone?

This was 1984, I swear.  Not Orwellian, but Huxley-esque, a brave new world of virtualization.  Hey, how about that, the mid-1980s full blown user-side and server-side virtualization.

We didn’t know what to call it, there were no marketing departments pushing Cloud or Virtualization.  We didn’t think of it as a revolution, or a movement, or a trend.  It was a means to a solution’s end.

Don’t get me wrong.  I think Amazon and Microsoft and Salesforce.com are entirely on the right track.  VMware and Citrix have developed some amazing software that offer solution architects at all levels of the stack an opportunity to optimize (and yes, also to screw up) solution designs and implementations.  When you have a massive, pervasive network that can be made secure – usually -the concepts of Cloud and Virtualization, have more reach, more impact.  Heck more people are using computers than anytime in history, and the people who use computers often have several computers, and they are more connected than ever.  The scale, relatively speaking, is far greater than in the 1970s and 1980s.  To me, THAT is what is different.

It is similar to the global warming situation.  As Al Gore has so elegantly depicted, the problem isn’t exactly the use of fossil fuels, or the destruction of forests, or the stamping out of drylands.  It is the scale that is different.  1 billion people burning wood to stay warm is not nearly the same problem as 7 billion people burning all kinds of things, wood, atoms, petroleum, coal, water molecules, methane – each person burns more than one person did 100 years ago, and there are 7x more people.  The challenge is scale.  And that is the case in our world of computing – more people using more computing devices that they expect to be pervasively connected, and we have several more billion people to go yet.  Yes, yes, bring on the Cloud and Virtualization, heaven knows the world needs it.

But is it new?  Is it a revolution?  Hmm, sounds like we are AT LEAST on Cloud 2.0 and Virtualization 2.0.  Maybe 3.0, or maybe 3.x.

To this day there are “workstation apps” that are best left running standalone on a workstation.  There are giant batch jobs that run their best race on z-Series mainframes.  And there are even some client/server designs that still work well – in fact most implemented “enterprise apps” are nothing more than client/server designs with a Web face on them – to this day.  Excel in the Cloud?  Sure, Office Live makes it work.  The Cloud lets ALL of these modes of computing potentially play with one-another.  Now that is a revolution.

I think SaaS and Cloud and Virtualization, overlapping ideas and subjects, are great.  It is about time everyone else caught up to some of the sophisticated solution designs and implementations of the 1970s and 1980s.

Now mobility… that is something pretty new.

Or is it?

I did something a little crazy over the holidays:  I read through the latest couple of 10-Ks and/or annual reports from the last few years from roughly the twenty largest IT vendors in the world.  I did this in search of R&D spending – was it still happening, who was doing it.

I came up with a list that includes 16 large vendors that disclose R&D expense in their respective “Statements of Operations.”  In order by 2008 fiscal revenue, largest-to-smallest, the list includes HP, IBM, Dell, Microsoft, Cisco, Intel, Apple, Oracle, Google, EMC, SAP, Seagate, Symantec, AMD, CA and NetApp.  I also looked into Accenture, CSC, Asustek, Lenovo and Sun and a few others but for a variety of reasons, such as the lack of disclosure of R&D expenses, I decided not to include them in the analysis mix.

In fiscal 2008 those 16 very large IT vendors reported nearly $556 billion in revenues.  They spent over $44 billion in R&D, so spending on R&D at a rate of 8% of revenues.  Pre-tax income for those vendors was $100 billion.  We can’t say that “therefore they spent 44% of pre-tax income on R&D” because R&D expense is accounted for before pre-tax income is calculated.  In addition, one could not say that “If the vendors spent $0 on R&D they would have earned $144 billion in pre-tax income” because it is virtually impossible to reasonably calculate the impact of R&D investment on pre-tax income in the scope of a single fiscal year – though I bet IT vendor CEOs wish it were easy to calculate a trustworthy number.

We all know 2009 was an economic debacle, so I am looking forward to making a comparison of R&D spending for fiscal 2009 versus fiscal 2008 for these 16 very large IT vendors.  So far nine of the 16 vendors have reported 2009 fiscal full-year financials, and of those nine 4 increased R&D spending, 5 decreased R&D spending.

Also of interest is that some of the vendors most associated with “being innovative” like Apple and Google do not actually spend, on a percentage basis, a particularly high amount on R&D.   Google does spend more than the 8% average of the 16 vendors, spending 12.8% of R&D expense to revenue in 2008, but in fiscal 2008 Microsoft, Cisco, Intel, SAP, Symantec, AMD and NetApp spent more on a percentage basis than Google.  Apple, perhaps shockingly, spent only 3.6% of revenue on R&D, but given their investments in retail stores the way the numbers work out is really not that shocking.  Still, only HP and Dell had lower rates of R&D spending in fiscal 2008 on a percentage basis than Apple in 2008.

Many have bemoaned the idea that Sun’s R&D brilliance and commitment will set beneath the Oracle portfolio of acquisitions.  Perhaps this is true, but Oracle did spend 12.2% on R&D in 2008, and 11.9% on R&D in 2009, roughly 4% above the index 8% rate.  CA likewise counts as a bit a of surprise – CA also spends nearly 12% on R&D.  When CA comes up with compelling concepts like “Mainframe 2.0″ there is proof that R&D has made a comeback in Islandia.

Who cares?  I guess the point is this:  $44 billion isn’t chicken feed.  Definitionally not all of this is R&D is research, not PhD driven, and not all of it is patch fixes either.  We can assume some kind of normal curve, with the bulk of R&D spending around upcoming product releases (with often a blush of patches soon thereafter), but some important but relative minority of spending on research or big ideas, and a roughly even amount on plugging holes found further into production.

But given the relative lack of VC spending on IT compared to a decade ago, the established vendors, those that have survived or thrived during a decade of massive consolidation, have assumed the primary reins of IT innovation.

I love sunsets, most of us do. My front deck in Belmont, CA looks due west and every night if we aren’t fogged in or it isn’t cloudy, and here in the hills of Belmont that is about 2/3rds of the days, we are blessed to witness the sun dipping over the low but lovely Santa Cruz Mountain peaks that separate Half Moon Bay from Silicon Valley – plenty of Creamsicle skies. Since I haven’t purchased a Cisco Flip yet and so haven’t recorded a sunset from my deck yet (note to family:  a missed holiday gift opportunity), I am sharing a short sunset I found on YouTube of one of the loveliest spots on earth, Gay Head on Martha’s Vineyard.

But I do not at all enjoy Sun Microsystems dipping over the the horizon of the IT landscape. When the Web exploded during the mid-to-late 1990s Sun was a darling of the new echelon. Their commitment to R&D resulted in a best-of-breed trifecta that included their own chips (Sparc), operating system (Solaris) and programming environment (Java). Add the fact that they “got” the Web in terms of vision evident by the prescient “the network is the computer” tagline and Sun basked in several years of notable success.

But Sun missed or was late covering several “left turns” in the industry:

1. Proprietary UNIX shifting to Linux
2. Proprietary boxes shifting to x86 servers
3. The monetization of Java – imagine if Sun had spun-out JavaSoft around 1999, talk about a cash war chest that could have carried through during the bursting of the bubble, hindsight is 20-20, alas.
4. The growing importance of middleware platform; Sun invested but didn’t execute well (remember iPlanet?). Imagine if Sun had acquired BEA and let them be. Instead, Oracle came out of nowhere, under the prod of Thomas Kurian, and became IBM’s alternative in a linchpin market.  Note that Sun wasn’t the only has-been in middleware, but the lack of any significant software platform north of Solaris left them quite exposed.

So be it, the big fish eats struggling fish phenomena is nothing new to the IT industry. What bothers me though will be the loss of R&D. Sun’s contributions of ideas and innovations had to be among the top in the industry on a per/revenue$ basis. Oracle, despite a legacy of organic R&D, is relatively speaking more of portfolio company than an R&D company, and will likely not maintain the level of R&D investment that Sun maintained in Sun’s product areas.

The golden era of Sun creativity centered around the likes of Bill Joy, John Gage, James Gosling and Eric Schmidt – and their many innovative compatriots inside and outside of Sun – their ilk may never be seen again in a commercial IT vendor.  Even in terms of marketing innovation the Java then JavaSoft team, with the likes of George Paolini, Jonathan Schwartz and Kim Polese, made things happen that had not happened before in the IT marketing space. Microsoft, not too shabby at marketing (yes, an intentional understatement) had all they could handle with the JavaSoft team.

But now we wait for the final setting of Sun Microsystems, and this is one sunset I wish would hide behind a cloud. What are the lessons to be learned from Sun? Two perhaps:

1. The balance between R&D and the realities of the marketplace must constantly be maintained and reviewed. Creativity cannot be buried under market tactics, but unbridled creativity can cause missed market opportunities, and eventually lead to a hubris and arrogance that can knock the pins out from the “execution” side of supplying IT.

2. Sun was often faulted, often by IBM, for partially hiding their light under a bushel (in an Eclipse of sorts). Make the R&D lines of demarcation extremely clear to your co-creators and partners: Project A will be utterly off base for anyone external (and most people internal). Project B will be an open book of participation and learning, even in a coopetitive environment. Trying to share your R&D but eat it too leaves one exposed to counter-movements, and you spend cycles covering your tracks.

It looks like the Oracle-Sun deal, after a scare with the EU, will actually close.  While the largest 16 IT vendors spent about $44b in R&D collectively during their respective fiscal 2008s, nothing to sneeze at and an average rate of 8% of revenue, Sun, spent more in the 13-14.5% range during the 2007-2009 fiscal years, and this while their financial fortunes were already well waning.  There are those, myself included, that think 2010 will be a return of investment for IT innovation after a dour 2009, even from VCs.  Thus there will be plenty “new” as Sun goes under.  But it is unlikely we will see anything quite akin to Sun during its heyday, a top vendor able to craft and nurture a culture of creativity that typically slips away from large vendors as they reach into the billions and tens of billions of revenues.