Are the search engine wars back?

In the late 1990s and early 2000s large web-based search engines fiercely competed for the market share. The big names of that period, such as Magellan, Excite, Infoseek, Inktomi, Northern Light, and AltaVista, completely faded from the collective memory, and Google emerged as the sole victor.

Google offered at least two advantages: a simple user interface clean of excessive advertising and search results clean of promoted links. Google also came up with an effective business model for a search engine.

A decade later Microsoft challenged the king of the search with Bing.

The fuss started in June last year when Google introduced changes to their classic look and feel: background graphics, left side bar toolbox, new image search results display… The problem was that the users immediately noticed how similar to Bing these “novelties” are.

Then this week Google punched directly back. This is how the sting operation worked:

  1. A meaningless string – “mbzrxpgjys” – is submitted as a query in Bing. Initially Bing does not show any search results.
  2. The same meaningless string is then manually associated with a webpage – RIM’s home page – in the Google’s index. However, the RIM’s homepage does not contain the string “mbzrxpgjys”. If Bing indexed the RIM’s home page, it did not find “mbzrxpgjys” and it could have not associated it with the RIM’s home page. This is why Bing initially shows no search results for “mbzrxpgjys”, but Google does.
  3. However, after a while, Bing starts showing the RIM’s home page for “mbzrxpgjys” just like Google.

The conclusion? Well, according to Google: Bing copied Google’s search results.

Hmm… This claim is perplexing: what does it exactly mean? One way to interpret this is that Bing copies Google’s ranking. This does not seem to be true – not even judging from the examples provided by Google. Another way to understand this is that Bing does not use its own index, but rather displays Google’s search results. This also does not seem to be the case, since results seem to be fairly diverse.

What this claim really seems to mean is that Google is panicking. This Google’s claim reminds me of those high school dorks that complained to teachers that the guy next to them is copying their test. – They worked so hard while the cool guys partied. Why would the cool guys now profit from their hard work? In my high school these disputes were quickly resolved with some bitch slapping in the school yard – and guess who always won.

It doesn’t seem likely that the dork strategy will work well for Google either. The Google guys seem somehow less and less cool. They also seem currently to be into picking fights everywhere.

And now Bing strikes back: Google manipulated Bing search results through a type of attack also known as click fraud. This means that Bing is now accusing Google of artificially creating a click stream pretending that they are Bing consumers, in order to trick Bing into believing that some webpages, such as the RIM’s home page, should be in Bing’s index, even if it does not make any sense to the indexing algorithms.

Google’s lashing out certainly created a lot of buzz, but judging from the current responses, it seems rather that a dork is getting bitch-slapped in the school yard again. For us, the spectators, this could mean that the search engine wars are coming back. The outcome of the previous round was that the search user experience drastically improved. Let’s hope for the similar outcome again. In any case, judging from the current skirmishes, we can expect a good show.

Visit our new product website (in Polish): Zarządzanie informacjami.


By TwitterButtons.net


Share

Multilingual Bookmarking

In the age of global economy and Internet-based media multilingualism plays an increasingly important role. Corporations, workgroups and communities are becoming multinational and transnational. The need to share knowledge between multilingual speakers is becoming a necessity. However, multilingual bookmarking is no simple matter.
The following are requirements that every enterprise bookmarking application should satisfy.

User Interface Localization

Users should expect to share and manage bookmarks through an interface with the full native language support.
Although localization of user interface is the simplest aspect of the multilingual bookmarking, in practice it is more complex than what it may seem at the first glance. Translation of labels and messages, for example, can easily run into the problem that the translated text is too long to fit into the application layout. Even worse, localization to the right-to-left languages, such as Arabic and Hebrew, requires inversion of the entire user interface (including subtle graphic details, such as on which side of a combo box the arrow that extends the list is).
Even the Latin alphabet varies between languages. There are actually multiple Latin alphabets – each with a specific sorting order. In addition, these differences also necessitate separate alphabetic indexes (for alphabetic browsing).
To make things worse, localization is never simply translation. For example, the English computer desktop is actually a lectern in Polish (“pulpit”). Often, terminology that makes perfect sense in one language simply cannot be translated into another.

Multilingual Bookmarks

Users should expect to bookmark multilingual content and data. The application should provide means to clearly identify the languages of the bookmarked content by merely inspecting the bookmark (i.e. without retrieving the actual content).
This early language identification is actually only one aspect of a broader problem. A significant deficiency of the browser bookmarks is that besides the title they do not provide any further description of the content. Tags and categories are useful for finding bookmarks, but they contribute little to understanding the content. Search engines, like Google and Bing, feature a short excerpt from the document text, in order to enable users to decide if the link is worth clicking or not. Bookmarks should allow the same: a short free text description of the content, which could then include a specification of the languages used in the text.

Multilingual Tagging

Users should expect to use multilingual tags. The application must provide means to indicate that one tag is a translation of another. In such case, even if a bookmark was tagged in only one language, users should be able to retrieve this bookmark by specifying the translation.
The simplest approach is to allow tagging in different languages. For example, the same web page could be tagged with “decathlon” (English) and “Zehnkampf” (German). Accordingly, the page could be retrieved in both languages. However, the problem arises when such page is tagged only in one language, for example using the English tag “decathlon”. – A German speaking user seeking pages tagged with the German tag “Zehnkampf” would simply miss that page.
A better approach would be, therefore, to provide tag translations. Accordingly, “decathlon” and “Zehnkampf” would be actually one and the same tag, only their display labels would be localized to specific languages. This solution is more problematic than what it may seem at first.
The English term “education” is broader than the French term “éducation”, and includes also certain aspects of the French terms “enseignement” and “formation”. On the other hand, the French term “enseignement” is broader than the English term “education”, and includes also certain aspects of the English terms “teaching” and “instruction” (for which the French also have the term” instruction”).
These difficulties involved in using multilingual keywords are studied and well understood in the library science and knowledge engineering. Serious efforts are made to build multilingual thesauri used for annotation in corporate and academic libraries. However, this experience was never applied to the social tagging.
Share

Do you need Enterprise 2.0 or Enterprise Bookmarking?

“[C]an someone explain to me the problem Enterprise 2.0 is trying to solve?”

Dennis Howlett, Enterprise 2.0: what a crock, ZDNet August 26, 2009

At least two Enterprise 2.0 suites include enterprise bookmarking as a feature (Lotus Connections and Jive SBS). Many others provide functionalities that could be (arguably) used to replace bookmarking. – If your team needs to share bookmarks, wouldn’t it be better then to adopt a full-fledged Enterprise 2.0 application?

Here are four reasons why you should not do that:

  1. Enterprise 2.0 is a philosophy. – You must adopt the whole Weltanschauung and change your lifestyle if you want to make it work. Enterprise 2.0 is not just another software. It requires a different corporate culture. It demands that you change your work habits, internal protocols and procedures, workflows… Enterprise 2.0 is commitment. Are you ready for the commitment? Do you need it?
  2. Not everybody wants to be exposed at work in the same way as in private life. –Enterprise 2.0 assumes that social software can be seamlessly transferred from the Internet to the enterprise. This is not necessarily true. Somebody told me recently that Facebook is for friends and family – LinkedIn is for everything else. Apparently, people – including the Tweeter generation – do make difference between work and life. The resistance to change that Enterprise 2.0 brings is real.
  3. What are you looking for: knowledge management or social networking? – Better knowledge management does not necessarily imply social applications. The question is where the focus should be. Do you want to improve knowledge sharing in the enterprise, or you want to transform the corporate culture? Enterprise bookmarking is knowledge management software. Enterprise 2.0 is social software.
  4. Prefer specialist software. – Even when the bookmarking is included in an Enterprise 2.0 suite, compare it to a specialized Enterprise Bookmarking application. Bookmark sharing as a feature in an application with a wide scope is different than a full-featured Enterprise Bookmarking application. Here is one simple benchmarking criterion: does bookmarking allow you also to share your use experience? If not – keep looking for the application that does.

Out of the 14 reasons why Enterprise 2.0 fails, #3 is choosing the wrong software. – Choose your software carefully. Make sure that you are buying the best software for your needs.

Share

Design Principles for Enterprise Bookmarking: Simplicity and Affordability

It is difficult to imagine a team of, say, 5 people or more who work primarily with information that would not need to share their bookmarks, bibliographies or some other kind of content references. – This is the business case for the enterprise bookmarking. But what should these people expect from the applications that provide such useful functionality? – There are two basic requirements: simplicity and affordability.

Simplicity

The web usability guru, Jakob Nielsen, views simplicity as one of the human rights (Usability: Empiricism or Ideology?). He also points out that this is not merely an ideological point of view: the users simply won’t struggle with the design complexity, and instead they will seek simpler alternatives that do not waste their time.
But simplicity is not a design principle that is easy to follow. In a highly competitive market there is always an urge to provide more features, more options, a richer user experience… Every product wants to outperform its competition by offering more of everything, and this obviously runs counter to the principle of simplicity.
It takes courage to declare that “less is more”, and to seriously follow it as a design principle. 37signals did exactly that with BaseCamp (Getting Real). – And they did it with a great success. They looked into their competition and found out that relatively simple project management activities, such as a to-do list, were enriched with so many nice-to-have features that they became excessively complicated. This is where they decided to break with this trend of outdoing the competition. “Our products do less than the competition – intentionally. We build products that work smarter, feel better, allow you to do things your way, and are easier to use.” (Ibid.)
Just like the project management activities, the basically simple and intuitive bookmarking can be also easily transformed into elaborate, intricate operation. You don’t believe it? – Here is one example: introduce configurable tagging metadata!
Here is how this works. You decide that since everybody uses tags, you need to come up with something else, something more – you need an innovation. So, you decide that your tags are not going to be simple keywords as elsewhere, but rather a data array consisting of multiple attributes (fields) and their corresponding values. Moreover, you even provide a facility for users to define the fields in this array, set restrictions on data types and data formats, etc. – As the result, that simple one-click bookmarking that everybody uses in their browsers now becomes a struggle with protracted multi-field forms.
These configurable annotation metadata may certainly appeal to about 1% of the potential enterprise bookmarking clients. Out of that tiny group maybe even 10% would be even willing to actually continuously and consistently bookmark their content using such complicated methodology. But this would also leave the rest of the users – who are the vast (99.9%) majority – stuck with an absolutely unnecessary complexity in usability and configuration.
This kind of “innovations” is certainly better left for customization projects paid by specific customers, and tailored to their specific requirements. Bookmarking is simple, and this is how it should stay for the vast majority of users. Here – less is certainly more.
Enterprise bookmarking users must demand a simple application that does not oblige them to change their current bookmarking habits – or any other work habits. The standards generally adopted on the web should be adopted and maintained, and the application should build its usability on the user’s familiarity with such standards. The innovations are good – but not for the sake of outdoing others. Innovations should further simplify the use of an application, and not do the opposite: innovations should not make an application more complicated and more difficult to use.

Affordability

Do you know what is the real cost of your software?
When you ask for the price of a software product, you typically hear the purchase price only, the price of the license. But this is never the whole cost.
Start with the infrastructure: how much additional software and hardware will you need to run your enterprise bookmarking application? Is this included into the price of the software, or is it that bill on you? Just imagine an enterprise bookmarking solution that requires an expensive database license!
And how about installation, configuration, deployment, intranet integration, integration with other applications – are these included into the price, or you will need to provide that too? – This is usually the honey trap of the free software: what starts as free soon turns out into a full-fledged “do-it-yourself” R&D project. Just like the IKEA furniture: some critical parts will always be missing, the holes are in the wrong place, and you will put it together wrong. – Twice. Eventually, to get somewhere, you will have to seek professional assistance and open the wallet.
Next, add training and implementation, support and maintenance, security and updates… Corporate software is usually like the notorious stone (or nail) soup: it starts with a pebble in a pot full of water on the fire. But then, to make it right you need a carrot, a potato, some oil, salt, few spices… Eventually, you end up paying for the whole soup.
Enterprise bookmarking users must demand an affordable application. The real, total price of your enterprise bookmarking solution – the whole bowl of soup – must be affordable. – Very affordable.

Share

5 Reasons Why Internet-based Social Bookmarking is not Suitable for Enterprise

You may wonder why you should invest in an enterprise bookmarking application when you can simply use Delicious, StumbleUpon or any other of numerous Internet-based social bookmarking services. – But, do you use Gmail or Hotmail as your corporate email? Of course not!
Here are 5 reasons why the Internet-based social bookmarking is not suitable for an enterprise:

  1. You want your proprietary, sensitive data behind your corporate firewall.
    No enterprise can risk keeping its business data on a public server. –Not even when the public host claims to be well protected. Even the leading networks, such as Google’s Gmail, Microsoft’s Hotmail or Facebook, are often hacked and their security is breached with catastrophic consequences. In such cases apologies are issued, but there are no guarantees, there is no compensation for damages. The only place for the corporate knowledge is safe behind the corporate firewalls.
  2. You want full control over your data.
    Security and privacy are not the only issues concerning data protection. You also need to back up your data regularly, archive it and then retrieve it from archives, ensure its integrity, maintain its usefulness… Some day in the future you may even need to merge it with additional data or federate it with additional data sources. You may need to modify its data schema, allow other applications to access it, even mash it up to create new applications… No doubt, you need a full control over your data.
  3. You want guaranteed performance.
    Public servers don’t provide any guarantees of their availability or performance. Outages hit every public network, no matter how big, strong and smart it is. Can you risk losing availability of your knowledge resources when serving a client? Can you depend on the performance of public networks during the peak time? – For business you need control over workloads, you need to carefully load balance the available resources, you need to guarantee the availability and the performance when they are required.
  4. You need more than tags and folksonomies.
    The issues with the Internet-based social bookmarking are not only technical, but also essential. Tagging is not sufficient for the corporate knowledge that requires more organization, definition and structuring. Even Delicious introduced the so-called tag bundles that intend to bring more structure into the chaos of folksonomies. But business needs more. Tags cannot be mere keywords. – They need to reflect the corporate vocabulary, encapsulate corporate knowledge, even promote corporate culture. In order to be useful for an enterprise tags must be knowledge enhanced.
  5. You need knowledge management.
    Bookmark sharing is not sufficient in a corporate context. Corporate users need to know why their colleagues bookmarked a resource, how did they actually use it, did they find it useful and for which purpose, do they know other similar resources, could they contribute ideas, solutions, opinions on related issues… Even who bookmarked a resource may have significance. Corporate users seek expertise, want to ask questions concerning a specific resource, want to learn from others’ experience… In short: in the corporate context bookmark sharing is just a tip of an iceberg. – A whole world of knowledge management is hidden below the water surface.

Here is a related video on YouTube: Social Bookmarking vs. Enterprise Bookmarking

Share

What is “enterprise bookmarking”?

Since the term enterprise bookmarking emerged a year-or-so ago, there has been a great deal of confusion regarding what it really is and how it is related to the social bookmarking.

The term social bookmarking was coined by Delicious (at the time known as del.icio.us) in 2003, but the idea of sharing bookmarks online was not new. The first one to offer shared online bookmarks was itList (http://www.mail-archive.com/scout-report@hypatia.cs.wisc.edu/msg00038.html) in 1996! During the dot com bubble era numerous bookmark sharing websites appeared: Backflip, Blink, Clip2, ClickMarks, HotLinks…

Nevertheless, these early attempts did not work out, which is often blamed on the burst of the bubble. However, Ari Paparo – the founder of Blink – offers a different angle on the lack of success of the early bookmark sharing applications. He believes that “it all came down to product design, and to some very slight differences in approach” (http://www.aripaparo.com/archive/001456.html).

Indeed, Delicious invented tagging. – Tagging radically changed user behavior and (arguably) could even count as the major conceptual innovation of the Web 2.0. Annotation of content using keywords was certainly not new. Nevertheless, it was always considered to be something that only librarians do, and the keywords were drawn from a controlled vocabulary (thesaurus). Tagging and the emerging folksonomies opened the world of library classification and indexing to the masses.

The simplicity of organizing, managing and searching bookmarks that was enabled through the use of tags may just be that “slight” difference of approach that Ari Paparo credits as the decisive factor in the success of Delicious. Soon, Delicious was followed by numerous similar websites featuring slight variations of the basic online bookmark sharing: Furl, Simpy, Citeulike, Connotea, Stumbleupon, Ma.gnolia, Blue Dot (later renamed to Faves), Diigo… Social bookmarking became a recognizable feature in the Web 2.0 landscape.

Surprisingly, it was IBM who first raised the idea of using social bookmarking in the enterprise context. Their research introduced Dogear in 2006 (http://queue.acm.org/detail.cfm?id=1105676) that soon became a part of the Lotus Connections.

Dogear simply copied the concept of Delicious and put it behind the corporate firewalls. Its inclusion into the Lotus Connections gave it a distinct Enterprise 2.0 flavor. This trend was soon followed by the ConnectBeam’s Spotlight, Jive’s SBS and FastNeuron’s InfoFlow. What is common to all these products is that they are primarily Enterprise 2.0 applications designed for the intranet and extranet social networking, and that they only marginally include the enterprise bookmarking functionalities. According to this trend, enterprise bookmarking is to Enterprise 2.0 what social bookmarking is to Web 2.0.

A different approach was taken by Jumper 2.0 and our own BrightLight. These products view enterprise bookmarking primarily as a knowledge management discipline, and place a greater emphasis on the creation and management of the corporate knowledge bases. The social aspects of the enterprise bookmarking – sharing, collaborative filtering, feedback and recommendations – are seen not as a goal in itself, but rather as the means for an effective transformation of the so-called tacit knowledge (i.e. knowledge that exists only in the heads of the employees) into the explicit knowledge. This approach also justifies products that are entirely dedicated only to the enterprise bookmarking, and do not cover additional Enterprise 2.0 standard features, such as blogs and social networking.

In this respect BrightLight is actually unique: BrightLight integrates discussion forums and wikis into the concept of tag. Tags are called topics in BrightLight, and every topic is automatically also a discussion forum and an entry in the knowledge base that is edited as a wiki. Users can hold discussions, ask questions and seek for expertise about every individual topic. From this point of view, topics are discussion threads. At the same time, users can also collaboratively compile and edit a description of a topic in the form of an encyclopedic article – like the articles from the Wikipedia. In this respect, topics are wiki pages. Nevertheless, the main role of topics is still to tag bookmarks, so that bookmarks can be browsed and searched using topics.

Accordingly, two trends, maybe even philosophies can be distinguished in the enterprise bookmarking. The first generation of products viewed enterprise bookmarking merely as a supportive functionality of the corporate social networking. The new generation of products views enterprise bookmarking in the context of the corporate knowledge management.

To learn more about BrightLight visit http://enterprise-bookmarking.net/.

This is the Technorati claim token: 8Q6H4XTGHU89 (please, ignore it).

Share

Follow

Get every new post delivered to your Inbox.