Oracle Endeca, More Oracle Than Endeca
January 27, 2013
In December 2012, Boston.com reported on the “Endeca exodus.” You can get the scoop by reading “As Endeca Exodus Continues, Trio of Former Employees Start Salsify to Help Manufacturers Distribute Better Product Info.” The content marketing play is not the part of the article which I found interesting. Here’s what I wanted to capture:
Co-founders Steve Papa and Pete Bell have both left Oracle as of this month. Former Endeca SVP Chris Comparato is now at Acquia, the Burlington company that peddles web content management software. Others have left for PayPal Boston, Silver Lining Systems, Sqrrl, Hopper, Internet advertising company DataXu, and Lookout Gaming, a new startup.
Why not cash in and check out? It is tough work making a search system generate revenue and even more difficult to achieve revenues and make a deal to sell the company to a larger firm.
The big outfits who buy search engines have the opportunity to learn first hand exactly how difficult it is to:
- Build revenues and turn a profit
- Find the resources to keep the software working
- Figure out how to market in a way that does not end in a flame out.
With these changes, Endeca is now Oracle. Can Oracle become more like the pre-acquisition Endeca which caught the attention of Oracle in the first place? Worth watching.
Stephen E Arnold, January 27, 2013
If you are interested in gourmet food and spirits, read Gourmet De Ville.
Computer Automation Is Making Researchers Obsolete
January 26, 2013
In archives and libraries around the world, piles of historic documents are sitting gathering dust. One of the problems librarians and archivists have with these documents is that they do not have a way to historically date them. The MIT Technology Review may solve that problem, says the article, “The Algorithms That Automatically Date Medieval Manuscripts.” Gelila Tilahun and other people from the University of Toronto have created algorithms that use language and common phrases to date the documents. Certain words and expressions can date a document to a specific time period. It sounds easy, but according to the article it is a bit more complex:
“However, the statistical approach is much more rigorous than simply looking for common phrases. Tilahun and co’s computer search looks for patterns in the distribution of words occurring once, twice, three times and so on. “Our goal is to develop algorithms to help automate the process of estimating the dates of undated charters through purely computational means,” they say. This approach reveals various patterns that they then test by attempting to date individual documents in this set. They say the best approach is one known as the maximum prevalence technique. This is a statistical technique that gives a most probable date by comparing the set of words in the document with the distribution in the training set.”
Tilahun and his team want their algorithms used for more than dating old documents as well. It can be used to find forgeries and verify authorship. The dating tool opens many more opportunities to explore history, but the down side is that research is getting more automated. Librarians and scholars may be kicked out and sent to work at Wal-Mart.
Whitney Grace, January 26, 2013
Sponsored by ArnoldIT.com, developer of Beyond Search
Forrester Fills the Gap in Search Market Size Estimates
January 25, 2013
I used to enjoy the search market size estimates of IDC (the time it takes to find info group), Gartner (the magic quad folks), Forrester (yep, the “wave” people), and Ovum (we do it all experts), among others.
I read “Growth of Big Data in Businesses Intensifies Global Demand for Enterprise Search Solutions, Finds Frost & Sullivan” and found several items of interest in the brief news story which arrived via Germany. Is Germany a leader in enterprise search? I heard that 99 percent of Germany’s search means Google. The numerous open source players are not setting the non-German world on fire, but I could be wrong. Check out GoPubMed, for example, of an interesting system which has a modest profile.
Now to the size of the search market.
The first thing I noticed was the nod to Big Data, which is certainly the hook on which many dreams for Big Money hang. With enterprise search vendors looking for a way to gain traction in a market which has been caught in awkward positions when licensing and deploying “search,” new words and new Velcro patches are needed. I won’t mention the Hewlett Packard Autonomy matter nor the Fast Search & Transfer matter nor the millions pumped into traditional search vendors with little chance of paying back the investments. No. No. No.
I want to quote this statement from :
The growth of Big Data across verticals presents the enterprise search solutions market with further opportunities. Since newer data types are not confined to a relational database within an organization, solutions that can search information outside the scope of these relational frameworks are widely accepted. Demand for personalized search tools that operate in a pool of unlimited data from internal servers, the Internet, or third-party sources is also growing.
Ah, but how does one crawfish away from exaggeration? Easy. I noted:
However, the disparity between customer expectations and actual search outcomes could dissuade future investments. Customers expect a single query to retrieve the right results immediately. Therefore, search providers must offer timely and relevant results, taking into account the continuous addition of new data to repositories.
But “How big is the market? my inner child yelps. The answer:
Apache Lucene Solr Updates
January 25, 2013
The DZone Big Data/BI Zone has let us know that a new version of Apache Lucene Solr has hit the internet. Apache Lucene Solr 3.6.2 has been unveiled, and it will roll into many other products that build upon the open source code. Read the details in, “Apache Lucene Solr 3.6.2.”
The gist of the release is in the first few lines:
“Apache Lucene and Solr PMC recently announced another version of Apache Lucene library and Apache Solr search server numbred 3.6.2. This is a minor bugfix release concentrated mainly on bugfixes in Apache Lucene library.
Apache Lucene 3.6.2 library can be downloaded from the following address: http://lucene.apache.org/core/mirrors-core-3x-redir.html?. Apache Solr 3.6.2 can be downloaded at the following URL address: http://lucene.apache.org/solr/mirrors-solr-3x-redir.html?”
Two products sure to be affected and improved by the update are LucidWorks Search and LucidWorks Big Data. LucidWorks chooses to use Lucene Solr as its foundation because of its dependability, agility, and strong developer and user communities. LucidWorks and any product that builds on open source and is going to be strong, secure, and continuously updated, just by its nature, and therefore a better choice than a proprietary option.
Emily Rae Aldridge, January 25, 2013
Sponsored by ArnoldIT.com, developer of Augmentext
Enterprise Search by White Is Vital Tool in Business Search Management
January 24, 2013
Martin White, information management consultant and Managing Director of Intranet Focus Ltd., is one of the leading experts on enterprise search and information access. White has published seven books on topics surrounding information consultancy and enterprise search applications. His most recent publication, Enterprise Search: Enhancing Business Performance, focuses on how to plan and implement a managed search environment in your corporation. The book explains how to meet both the needs of your business and your employees.
White makes a clear business case for search, emphasizing the need to evaluate current search systems and the creation of a support team. The book is well organized and easy to read, with a thorough preface giving an overview of chapters and topics as well as simplified summaries at the end of each chapter. This style makes White’s recent book a great tool for the busy professional.
Chapter 12 is a recommended starting place, listing twelve critical success factors. White states that if you don’t meet at least eight of these twelve, which include investing in a search support team, getting the best out of your current investment in search, and providing location-independent search, then you definitely need the contents of this book.
In Chapter 10, titled Managing Search, White expands on the idea of managing a search support team:
“Implementing search should never be ‘a project’. The work of ensuring that users continue to have high levels of search satisfaction will never come to a close. Each week, and perhaps even most days, there will be something that needs attention. The role of the search support team is not just to be reactive but to anticipate when changes to the search application need to be made, or to identify a training requirement that will address an issue that is just starting to show up on the search logs and user satisfaction surveys.”
Most organizations are not prepared for the rate of growth of information that they are experiencing. White does a great job dissecting the need for enterprise search and then giving you the tools to successfully manage your system, based on far more than just available technology. The section on the future of enterprise search, Chapter 11, stood out to me. White makes an excellent case for why this topic can no longer be ignored.
Additional features include a thorough glossary, lists of books and blogs on information retrieval and enterprise search, and resources for further reading. The book is available here from O’Reilly Media in eBook and print formats. Highly recommended.
Andrea Hayden, January 24, 2013
Sponsored by ArnoldIT.com, developer of Beyond Search
Search Technologies Success Lies in Corporate Retreats
January 23, 2013
While many companies may see corporate retreats as an obvious place to cut spending, co-founder and chief executive of Search Technologies believes retreats are some of the most valuable investments made by the company. In the Washington Post article “Value Added: This Herndon Search Company Found its Perfect Retreat in Costa Rica,” we learn about how Kamran Khan of Search Technologies believes corporate retreats are crucial to the success of his growing business. The most recent off-site cost $100,000 in company money and took place in highly-educated and tech-savvy Costa Rica.
The article explains the importance:
“Khan, who started Search Technologies in 2005, said it’s the only time when everyone in the company — including the management team — can be in one place. Khan uses the chance to address his 100-person staff, informing them of how the company is doing and outlining the goals for the next year. ‘I prefer to get people together and . . .clarify our strategy, which is very simple: We are going to be experts in the search space.’”
Khan and his team at Search Technologies may be onto something with this plan. Launched in 2005, the company was on track for $18 million in revenue for 2012, and the company’s net profit margin is about 5 percent. The IT services and search implementation software company services the Daily Mail newspaper’s Web site portfolio in Britain and helped Amazon.com launch its new cloud search product. Apparently the secret to success lies in Khan’s philosophy of hiring “good people” and taking beach trips. We have learned that Search Technologies is hiring in anticipation of further growth during 2013.
Andrea Hayden, January 23, 2013
Now You Are Talking: Can a Company Make Money with Enterprise Search?
January 22, 2013
I have better things to do that capture my immediate thoughts about “Inside H-P’s Missed Chance to Avoid a Disastrous Deal.” You can find the article in a dead tree version of the Wall Street Journal on page 1 with a jump to Page 16, where the “would not comment” phrase appears with alarming frequency.
The most interesting point in the write up is the quote, allegedly crafted by a Hewlett Packard Big Dog:
Now you’re talking.
Like much of the chatter about search, content processing, and Big Data analytics, on the surface these information retrieval software companies are like Kentucky Derby hopefuls on a crisp spring morning. The big pay day is two minutes away. How can the sleek, groomed, documented thoroughbreds lose?
The reality, documented in the Wall Street Journal, is that some companies with sure fire winning strategies can win. Now you’re talking.
How did HP get itself into the headline making situation? How can smart folks spend so much money, reverse course, and appear to be so scattered? Beats me.
I have, however, seen this before. As I read the Wall Street Journal’s story, I wrote down some thoughts in the margin of the dead tree instance of the story at the breakfast table.
A happy quack to Lubrisyn.com
Herewith are my notes to myself:
First, name one search vendor in the period from 1970 to the present which has generated more than $1 billion in revenue from search. Acquisitions like IBM’s purchase of iPhrase (er, what happened to that outfit), Vivisimo (now a Big Data company!), or SPSS’s Clementine (ah, you don’t know Clementine. Shame on you.) Don’t toss Google and its search appliance into the mix. Google only hints at the great success of the product. When was the last time you searched using a Google Search Appliance?
Second, didn’t Microsoft purchase Fast Search & Transfer for $1.2 billion in January 2008. How is that working out? The legions of search add in vendors for SharePoint are busy, but the core system has become a little bit like dear old Clementine. Fast Search was the subject of a couple of probes, but the big question which has not yet been answered as far as I know is, “How much revenue did Fast Search generate versus how much revenue Fast Search reported?” I heard that the revenues were, to some degree, inflated. I thought search was a sure fire way to make money.
Third, after more than a decade of top down marketing, why did Endeca need cash infusions from Intel and SAP venture units? How much did Oracle pay for Endeca? Some azure chip consultants have described Endeca as the leading vendor of enterprise search. Endeca added ecommerce and business intelligence to its line up of products. What was the firm’s revenue at the time of its sale to Oracle? I estimated about $150 million.
Fourth, Dassault, the company with the “system”, bought Exalead. What has happened to this promising technology? Is Exalead now a $200 million a year revenue producer for the prestigious French engineering firm? Perhaps the “system” has been so successful that Exalead is now infused into Dassault clients throughout the world? On the other hand, wouldn’t a solution with this type of impact make headlines every week even in the US. Is it more difficult to to cultivate information retrieval revenues than other types of software revenue? The good news is that Dassault paid a reasonable price for Exalead, avoiding the Autonomy, Endeca, and Fast Search purchase prices.
These examples reminded me that even if my estimates are wide of the mark by 20 or 30 percent, how could any company generate the astounding growth required to pay the $11 billion acquisition cost, invest in search technology, and market a product which is pretty much available for free as open source software today? Answer: Long shot. Exercise that horse and make sure you have what it takes to pay the jockey, the stable hands, the vet, and the transportation costs. Without that cash cushion, a Derby hopeful will put a person in a financial hole. Similar to search dreams of big acquirers? Yep. Maybe identical?
Two different points occurred to me.
On one hand, search and its bandwagon riders like Big Data analytics must seems to be a combination of the Klondike’s mother load and a must-have function no matter what a professional does for a living. The reality is that of the 65 search and related vendors I have written about in my books and confidential reports, only three managed to break the $100 million in search revenue ceiling. The companies were Autonomy, Endeca, and Fast Search. Of the three, only Endeca emerged relatively unscathed from the process. The other 62 companies either went out of business (Convera, Delphes, Entopia) or stalled at revenues in the millions of dollar. If one totals the investments in these 65 firms to generate their revenues, search is not a break even investment. Companies like Attivio and Coveo have captured tens of millions of venture dollars. Those investors want a return. What are the odds that these companies can generate more revenues than Autonomy? Interesting question.
On the other hand, search and its child disciplines remain the most complex of modern computing problems. Whether it is voice to text to search and then to predictive analytics for voice call intercepts or just figuring out what Buffy and Trent in the sales department need to understand a new competitor, software is just not up to the task. That means that money pumped into promising companies will pay big dividends. Now the logic may make sense to an MBA, but I have spent more than 35 years explaining that progress in search is tough to achieve, expensive to support, and disappointing to most system users. The notion that a big company could buy software that is essentially customized to each customer’s use cases (notice the plural of “cases”) and make big money is a characteristic of many firms and managers. The reality is that even governments lack the money to make search work.
Don’t get me wrong.
There are small firms which because they focus on quite specific problems can deliver value to a licensee. However, big money assumes that search technology will be a universal, easily applied to many situations. Even Google, with its paid search model, is now facing innovation challenges. With lots of smart people, Google is hiring the aging wizards of search in an attempt to find something that works better than the voting methods in use today.
What do my jottings suggest? Search is a tough business. Assumptions about how much money one can make from search in an era of open source options and cost cutting need to be looked at in a different way. The current approach, as the Wall Street Journal write up makes clear, is not working particularly well. Does this search revenue track record suggest that the azure chip consultants, former middle school teachers, and real journalists miss the larger message of search, content processing, and Big Data analytics? My tentative answer is, “Yep.”
Stephen E Arnold, January 22, 2013
The Question Drives the Search
January 22, 2013
Over at Chiliad, an article called “Search Vs. Correlation Vs. Causality-What Do Your Goals Require?” discusses how different types of questions change search results. Business intelligence and search are different aspects of the same end result and together they can generate more useful results. Correlations provide analytics, thus turning up unexpected and often useful relationships. The value is not in observations, but rather connections between data, which then influences decision making. The “why” factor is also a big part, because it explains how the data will be used and what the end result will be.
It involves more legwork than anything else:
“Iterative Discovery—understanding “why”—requires a different approach. Not only does digging in deliver more information, it suggests new inquiry and allows you to dig deeper. It helps you understand—across all your sources—what matters most. Although Chiliad named this approach Iterative Discovery, we didn’t invent it. Great researchers and analysts did. We simply observed them—and created a tool tuned to figuring out…’What does it mean?’”
If the why question cannot be answered than search, business intelligence, and everything else is useless. Users conduct these actions to find an answer and if an answer is not provided the action are worthless.
Whitney Grace, January 22, 2013
Sponsored by ArnoldIT.com, developer of Beyond Search
Vaporware Does Not Make You Rich
January 20, 2013
HP bought Autonomy in hopes to turn a profit from the company’s software, but upon delving into Autonomy’s records HP discovered they had invested in vaporware. Read Write focuses on the “Vaporware Allegation Latest HP/Autonomy Twist.” Stanley Morrical is suing HP, because he does not believe the software exists and all HP has to do is prove that it bought $10.3 billion worth of marketable software. To cover a possible blunder, HP claims that Autonomy fooled them with creative accounting and information misrepresentation. Morrical states that HP is doing this to cover its own tracks for making a foolish purchase or a nonexistent purchase.
“While claiming to have IDOL 10 ready, HP actually had nothing to sell, Morrical is accusing. Essentially, he claims, IDOL 10 was vaporware.
‘You go out in the market and say it’s available and it’s not,’ Aron Liang, an associate at the San Francisco law firm Cotchett Pitre & McCarthy, which is representing Morrical, said. ‘So either they knew it and they’re lying or they don’t even know what they’re selling, which in some ways may even be worse.’
David Schubmehl, a tech analyst for International Data Group, said he was briefed on IDOL 10 in June. However, Schubmehl says he hasn’t talked to any companies using the software.
‘I can’t confirm that anyone is actually using IDOL 10,’ Schubmehl said. ‘However, I have had briefings about that back in June and it certainly seemed to be part of their big data offerings.’”
Nobody has used IDOL 10 it seems, so how could a company have $900 million in revenues from vaporware? Somebody here is lying, but HP and Autonomy are pointing the finger at the other person. Whose nose is really growing?
Whitney Grace, January 20, 2013
Sponsored by ArnoldIT.com, developer of Beyond Search
Yandex Creates Powerful Facebook Search App
January 19, 2013
We know that Facebook is very protective of its services/ products and that their practices concerning user data are questionable. What will Facebook do, however, with Yandex’s new search app? Tech Crunch announced, “Russian Giant Yandex Has Secretly Built A Killer Facebook Search Engine App Codenamed ‘Wonder’.” The search engine app allows users to ask what content and businesses friends visited. Facebook prohibits search engines to use its data without permission. A spokesperson from Yandex was not able to comment on Wonder, but did confirm the company as interested in mining social data and building social products.
Wonder works by allowing its users to vocally search for information and it lists whether their friends have searched for it as well. Yandex so far has limited themselves to the Russian market, but Google and other competitors have eaten away at its revenue and so they are turning to other areas. Some areas are mobile, maps, and app discovery for services/products.
What does Facebook think about this? Facebook tried to allow its users to search friends’ content with Nearby. Also Wonder might use too much of Facebook’s user data and Facebook does not volunteer user information to search engines, which Wonder might do. Facebook is taking its own steps to get into search:
“CEO Mark Zuckerberg himself explained at TechCrunch Disrupt SF that Facebook is getting into search:
‘Search is interesting. I think search engines are really evolving to give you a set of answers’’ I have this specific question, answer this question for me.’ Facebook is pretty uniquely positioned to answer the questions people have. ; What sushi restaurants have my friends gone to in New York in the last six months and Liked?’ These are questions that you could potentially do at Facebook if we built out this system that you couldn’t do anywhere else. And at some point we’ll do it. We have a team working on search.’”
There are various options that Facebook could do with Wonder: buy it, make a joint partnership, grant permission, etc. but we will have to wait and see what will happen. We do know that users are demanding Facebook create a better search engine and Wonder is making them work faster to develop it.
Whitney Grace, January 19, 2013
Sponsored by ArnoldIT.com, developer of Beyond Search
-
- Subscribe to Beyond Search
Feature archive
News archive
-
