• Home
• About
• Writers
• Landscape
• Wizards Index
• Fraud
• Wanna Be Like Larry?

HP, Autonomy, and a Context Free Expert Output about Search: The Bet on a Horse Approach to Market Analysis

I don’t think too much about:

1. Azure chip consultants. You know, these are the firms which make a living from rah rahs, buzzwording, and pontification to sell reports. (I know. I labored at a non-azure chip outfit for what seems like decades. Experience is a good instructor. Oh, if you are a consultant, please, complain about my opinion using the comments section of this free blog.)
2. Hewlett Packard. I recall that the company used to make lab equipment which was cool. Now I think the firm is in some other businesses but as quickly as I latch on to one like the Treo and mobile, HP exits the business. The venerable firm confuses my 69 year old mind.
3. Autonomy. I think I did some work for the outfit but I cannot recall. Age and the lifestyle in rural Kentucky takes a toll on the memory I admit.

Nevertheless, I read “HP’s Autonomy Could Face Uphill Battle In Data Market.” There were some gems in the write up which I found amusing and illustrative of the problems which azure chip consulting firms and their experts have when tackling certain business issues.

The main idea of the write up for “investors” is that HP faces “challenges.” Okay. That’s a blinding insight. As you may recall, HP bought Autonomy for $11 billion and then a few months later roiled the “investors” by writing off billions on the deal. That was the mobile phone model, wasn’t it?

The write up then pointed out:

HP wanted Autonomy to jump-start its move into software and cloud-based computing. Autonomy is the No. 1 provider of search and retrieval software that companies use to find and share files and other information on their websites and document management systems.

Okay. But that too seems obvious.

Now here comes the kicker. The expert outfit providing inputs to the reporter doing the bull dog grip on this worn out bone is quoted as saying:

“Software license revenue (in this market) isn’t growing at the same rate as before, and we are beginning to see the rise of some new technologies, specifically content analytics and unified information access,” Schubmehl said. These new types of software can be used with types of business analytics software, business intelligence software and other software to help enterprises do a better job of locating specific information, he says, which is the job of search retrieval software.

I don’t know much about IDC but what strikes me from this passage is that there are some assertions in this snippet which may warrant a tiny bit of evaluation.

Will context free analyses deliver a winner? Will there be a Gamblers Anonymous for those who bet on what journalists and mid tier (second string) consultancies promulgate? For more about Gamblers Anonymous navigate to http://www.gamblersanonymous.org/ga/

Here goes:

First, the growth in “this market isn’t growing at the same rate as before.” Okay, how about a fact or two. Based on my research which has been chugging along for more than 20 years, the market for search has been a problem for a long, long time. The poster child for growth was Autonomy, which achieved its $800 or $900 million via a combination of hard nosed sales, acquisitions, and even harder nosed management of the company. Endeca, by way of contrast, generated about $150 million in revenue from search, ecommerce, and business intelligence related services and licenses. The company sold for about $1.5 billion, based on the information I had at the time of the deal. The rest of the 250 search vendors which I have tracked since the late 1990s were far less successful in generating revenues. Examples (which did not seem to affect the azure chip crowd) range from the spectacular implosion of Convera to the whimpers of Entopia. Fast Search & Transfer was a candidate poster  child, but the company seemed to be caught in a combination of financial inquiries, product line pruning, and conversion to a snap in for SharePoint since Microsoft took over the company in 2008. Outfits like Brainware, Exalead, ISYS Search, and Vivisimo were acquired but for fractions of the amounts paid for Autonomy and Endeca. Who even remembers Entopia, iPhrase, or Siderean? When azure chip outfits talk about growth of the search market, my research suggests that the crazy numbers generated by some of the research firms come about because search was not going anywhere.

Second, how about those new technologies? Well, how about them? Consider analytics. Are analytics new. Most of the companies which are scurrying around with next generation analytics, predictive analytics, or real time analytics are constrained by what is sometimes called the Big O problem. I will be discussing this minor hitch in the git along in my talks in London and New York in May 2013. If you are not into the Big O thing, the basic idea is that there may be fancy math, but the systems are constrained by computational scale. So shortcuts are needed. As a result, what’s new is the naive assumption that analytics will extend search in a massive way. Sorry. –What’s being hyped by the azure chip crowd is reports which assert that something really new is happening. The article mentions Palantir. So what is Palantir doing in addition to consuming hundreds of millions of venture funding, working to sidestep the i2 Group legal matter, and generate eye popping graphics for those who do not understand math? The answer is, “Everyone is chasing revenue.” Are the technologies “new”? In my view, the fancy math is, like Autonomy’s method, not exactly a great leap forward. Leaps are tough to achieve because of Big O and other constraints such as cost.

Third, what’s with the reference to ElasticSearch? My recollection is that the experts responsible for one of the azure chip outfit’s analyses of the company were contractors. I don’t have the list of authors in front of me, but I suspect that at least one of these people is probably less well informed than some of the others on the authorship and research team. Since I was an author, I guess I can be the big numbskull, but I would hazard the suggestion that another person might be the candidate wear the dunce cap. Fact is there are dozens of open source “search” options available. Most of the proprietary search vendors use open source software. The adoption of open source search software is strongest in established, large organizations. This is an extremely important point. It helps explain why the best customers for search vendors are looking at open source options when from LucidWorks, Sphinx Search, or one of the dozens of other providers. The costs for traditional search solutions are high and the payoffs from search are not what licensees want. As a result, outfits with findability problems are looking for a silver bullet. Azure chip poobahs are, therefore, offering up nostrums, snake oil, and jargon to sell reports and jump start sales.

Here’s one of the slides which I will use in my upcoming talk. This is a list of some of the baloney terms azure chip satraps are using to help their clients and prospects think they can solve their findability woes:

Notice search is in read. The reason? Search has been around for decades and still does not work. Google is not a search vendor. Google is an ad company. Vivisimo, to take an outfit at the other end of the technology scale, is not a search company. Vivisimo does connectors for Big Data. Search is just one term. The revenue comes from selling something about one of these other “terms”, or “trends”, or new things.

What action should a person take when confronted with write ups which are floating fact free and without context. Three suggestions:

1. Find someone to help you ask and demand answers based on verifiable, factual information. Assertions are okay, but those assertions are designed to sell consulting work, not help those who need solid, verifiable information to make a decision.
2. Remember that search is not working particularly well despite decades of innovation, investment, and effort. A new term slapped on something which does not work may not get the problem solved. In fact, taking azure chip outputs at face value may be adding to one’s information and cost woes.
3. Azure chip outfits are struggling to generate revenues. As a result, marketing is more important than any other factor. Azure chip consultants often have spotty work records, know how to flip burgers, and maybe use LinkedIn to get a part time job. None of these work experiences can take the place of knowing about the companies, technologies, and contexts of information processing.

Well, it’s Derby Day? If you cannot go to Churchill Downs, why not bet on an azure chip consultant? The likelihood of winning the findability race are similar to what the horse crowd will face today.

Stephen E Arnold, May 4, 2013

Sponsored by Augmentext

Comments

Copyright © 2008 Beyond Search · ArnoldIT.com · Disclaimer · Revolution News theme by Brian Gardner · Log in