“People use Google at home, but at work it’s a different matter,” Google’s product manager Cyrus Mistry tells CBR. “CEOs are increasingly asking why search at work can’t search be the same as at home.”
That is certainly something Google would like. Figures from Internet market research firm Nielsen Online for April 2009 suggest that of the 8.6 billion searches carried out online, Google accounted for 5.5 billion. Yahoo! was second with 1.4 billion, with Microsoft trailing in third on 852 million. But how easy is it to transfer success in the consumer space to the enterprise market, and how much can the two learn from each other? Where does innovation in the consumer space leave the enterprise market?
“A lot of search vendors haven’t positioned why they exist,” says Martin Atherton, research director at Freeform Dynamics. “It’s all about ‘our product is better than yours’. That should be the last point raised, not the first.”
Microsoft’s Bing engine, launched in June 2009 to replace the much-maligned MSN Live Search, claims to be a “discovery” engine rather than a traditional search engine. The company’s own research found that a large percentage of search queries are not answered by the current generation of search engines.
“40% of search queries go unanswered. There is something missing here and a big consumer need. We can see it in the logs [of searches]. When searching using existing search engines I have to keep re-querying things - adding more words, clicking on a site, going back because it is not the right site, and ultimately abandoning their queries,” Paul Stoddart, Microsoft UK search lead, says.
Microsoft says it is hoping to make search results more relevant. So for example a search query for Wimbledon will return different results during the tennis season than at Christmas time. A search for flight information will reveal schedules and times as well as results for local hotels and weather information.
Just before Bing’s launch another Google rival emerged. Wolfram|Alpha, developed by British physicist and mathematician Stephen Wolfram, is a computational knowledge engine rather than a traditional search engine. Wolfram|Alpha aims to calculate answers to direct questions, rather than trawl through information that is already available on the web. The information it uses has been collated and verified by its own experts, so users are said to be presented with more relevant and accurate results.
Google has fired back with its Google Squared engine. Rather than presenting a user with a list of websites that contain the search query, Google Squared pulls together information from various websites and presents the results to the user. “It’s a way of getting more information without doing that many searches,” says Anthony House, Google’s UK head of communication and PR. “It’s trying to bring the results to you and add structure to unstructured data.”
It is easy to see why Microsoft has been pushing its search strategy recently. Nielsen Online put Google’s share of the US search market at 57% in January 2008, rising to 63% in December of that year. Microsoft, by way of contrast began at 12% and slid to just under 10% by the end of 2008.
It is worth pointing out that Nielsen’s figures are skewed by the inclusion of results from Microsoft’s Club Live searches. Club Live (now known as Club Bing) is a programme that rewards users for searching within Live by clicking on the Search and Win link in the results. Search Engine Land estimates that including Club Live searches in its results bumps Microsoft’s figures by as much as 3%.
Google’s domination of the consumer search space has helped it turn in impressive financial results. The Mountain View, California-based firm’s figures for Q1 2009, released in April 2009, revealed revenue of $5.51bn, a 6% year on year increase. Google-owned sites generated revenue of £3.70bn during that period, up 9% from the same period a year previously. UK revenue totalled $733m. Net income for Q1 2009 stood at $1.42bn, compared to $1.31bn for Q1 2008.
Google’s financial results for the full year 2008 show just how reliant the company is on advertising revenue. Total yearly revenue stood at $21,795,550, with advertising revenue contributing $21,128,514.
FAST track to success?
Microsoft has long being trying to improve its search offering, in the consumer space as well as the enterprise market. In February 2008, the company acquired Norwegian enterprise search firm Fast Search and Transfer (FAST) for $1.2bn in cash. The move was seen by many as an attempt by Redmond to improve its capabilities to search for non Microsoft-based unstructured data.
At the time of the deal, Jeff Raikes, president of the Microsoft Business Division, said: “Enterprise search is becoming an indispensable tool to businesses of all sizes, helping people find, use and share critical business information quickly. Until now organisations have been forced to choose between powerful, high-end search technologies or more mainstream, infrastructure solutions. The combination of Microsoft and FAST gives customers a new choice: a single vendor with solutions that span the full range of customer needs.”
In February 2009, the company released FAST Search for SharePoint, hoping to improve the search capabilities of its collaboration suite. As well as enabling users to search within the company’s firewall, FAST Search for SharePoint also enables search on externally-facing websites. This opens up the search results to include not only internal, structured data, but also externally-held, unstructured information gathered from the web, one of the developments that has moved enterprise search towards the consumer space.
This leads to another element of enterprise search – having control over the data that you’re searching.
“Companies are caught between a rock and a hard place – they don’t know what they don’t know,” says Atherton. “A search query could return 10, 50, 100 results – but are they right? Lots of search products assume that the house is already in order.”
Finding the data is only half the problem, Atherton argues. A much clearer understanding of what, where and how valuable the data is, is also critical. “Companies need to see the bigger picture with information management,” he says. “Search can be the answer, but you have to establish the problem first.”
In January 2009 UK search firm Autonomy acquired US-based content management vendor Interwoven for $775m. At the time of the deal, Autonomy CEO Mike Lynch told CBR that the deal would help the company improve its offerings in the legal sector as well as help companies to understand the context and meaning of content.
“We can understand the meaning of content while the CMS [from Interwoven] is good at putting information around the content, like who checked it in and checked it out; as well as put a nice front end on it to make it intuitive to use,” Lynch told CBR.
With search vendors moving beyond their traditional field of expertise – “Autonomy is now a content management company, albeit a very good one” says Simon Bain, CTO at Simplexo, reflecting on the Interwoven acquisition – are companies better off going with a larger company with broader capabilities?
“Vendors with broader remits can be of use to customers seeking to learn about how a topic, in this case search, fits into the broader landscape, and also useful for vendors with narrower remits, meaning smaller product portfolios, to ride on the coat tails of.” says Atherton.
But with such importance being placed on finding the correct results – especially where legislation, regulation and law are concerned – companies may feel more comfortable dealing with a specialised search vendor such as Recommind or Endeca.
The legal sector, Recommind’s traditional stronghold, is one industry where the ability to quickly and correctly search through data is vital. “Information is increasingly recognised as the most valuable business currency – but with the sheer volume plaguing organisations today, it can actually create major challenges for companies if employees are struggling to locate the information they need for their jobs,” Simon Price, Recommind’s European director, said recently.
Google’s latest enterprise search offering, Goole Search Appliance 6.0, claims to be able to search billions of documents. The product contains what Google calls a “dynamic scalability feature” which enables companies to connect multiple GSAs across a number of locations. This enables the search platform to cover billions of documents if needed.
“This is the largest scale of search a company will need,” says Mistry. “Google itself has close to a fifth of a billion documents. We are talking really high-end enterprise search.”
Lynch’s assertion that products should be ‘intuitive to use’ is another example of the enterprise-consumer crossover. GSA 6.0, includes a number of features borrowed from the consumer space. These include Query Suggestion, which auto-fills suggestions as a user begins to type, and the ability to remove answers that are less relevant from the results listing. It also features Ranking Framework, meaning a user can tailor results depending on their relevancy, Google’s Mistry says. “It is like a star system or a bookmark system,” he tells CBR. “The user can dictate which results are most relevant and over time the Ranking Framework will learn the user’s preferences.”
The move to ‘consumerise’ elements of enterprise search is perhaps where the market should be looking for its next evolution. “It’s the ultimate goal,” says Mistry. “Every worker truly wants to be able to find everything at work just as the would online at home.”
This is potentially a dangerous move, believes Alistair Handyside, CEO of Simplexo, a UK open source search enterprise start-up. “Enterprise search is not about the marketing, it’s about getting the right answer. We wouldn’t manipulate the data or answers because someone has paid $2 for that query to be at the top. Enterprises need to know that the results are totally accurate. Companies need to find information as fast as possible due to compliance, freedom of information requests and so on. The correct responses need to be at the top of the list.” he tells CBR.
Atherton echoes Handyside’s view: “If you do a Google search for my name, for example, you’ll get lots of answers but only some of them will be useful. In the enterprise, particularly law, you need confidence that the answer is what you’re looking for. You may do a search for a certain case and find five files that you didn’t know existed.”
The is much that the enterprise space can learn from consumer search, especially concerning ease of use and user involvement. However, consumer searches that return a large number of vaguely relevant results is not good enough for businesses. They need to know that what they are searching for is relevant and accurate.
An enterprise search engine will only be useful if a company is in control of its data – how much they have, what it means and where it is. If a company is aware of this then there is great potential for search to be the starting point for a wider information management strategy.
Carousel image courtesy of Just Us 3, Flickr, CC licence.