Despite the economic downturn venture capitalists are still looking at investment opportunities in digital media. New Media Knowledge spoke to one to get the low down on what VCs want.
moreWeb publishers and broadcasters always have to look for new and innovative ways to maintain existing audiences and win new ones. When BBC Radio 5 Live realised it was losing listeners to other sources of football-based content it launched a new service to win them back. New Media Knowledge met the people responsible.
moreTelevision is increasingly embracing new media to reach new audiences and add interesting applications. New Media Knowledge talks to the people behind a new online conservation series that aims to raise wider awareness of the plight of endangered species.
moreSemantics expert Professor David Crystal OBE has been analysing linguistics since the 1960s. He has spent the last 12 years applying his studies of semantics to good effect on the Internet. He spoke to New Media Knowledge’s Chris Lee about his career with words. more
African nations were urged this week to engage with new media to attract more visitors. New Media Knowledge looked for examples of new media success in the travel sector. more
Semantic search is poorly understood and leading to claims for its powers that lie beyond the bounds of what computers are able to do, says Charlie Hull, MD of Lemur Consulting.
moreBusinesses are alienating a large section of the online market by failing to optimise their sites for disabled Web users. more
Microsoft’s continued pursuit of Google has taken a positive step, following the announcement that the Redmond-based company will provide the search tools for Facebook’s 90 million users. more
On 8 September leading practicioners and strategists in the UK and global search scene came together to address the latest trends. Commercial, technological, and research innovations were explored and debated. Read the full report...
On 8 September leading practicioners and strategists in the UK and global search scene came together to address the latest trends. Commercial, technological, and research innovations were explored and debated.
By Deirdre Molloy
[Register and post your own comments on this report below...]
Introduction and opening remarks - Mike Grehan, Smart Interactive
Mike Grehan noted in opening that in 1995 there were 75 websites and everybody knew each other, and there was no businesses online at all. Then Infoseek and Alta Vista appeared and businesses started to ask why they weren't appearing on the websites. In the forthcoming third edition of his book Search Engine Marketing: The Essential Best Practice Guide, Mike explained that he has added coverage on systems and processes in search that don't really work, and he took 'page rank' as an example. Page rank doesn't work, he asserted. As a formula it's pretty simple - but now that so many sites are database or dynamically driven it just can't be done. Google says that page rank is the rank of the importance of the page to Google. This, Mike commented, is a symptom of 'Google Anxiety Syndrome' (GAS).
New Research Findings - John Myers, Latitude
Jon noted Latitude's latest research, which revealed the simultaneous declining costs-per-click (CPCs) in the UK market and the UK market's increasing competitiveness. So why has there been a CPC downturn? He made an interesting point in regard to this question when he addressed the degree of warehouse knowledge differences.
Google is different from Overture, which is different from Miva. They all have different characteristics and areas where they best deliver.
The US is more competitive in e-commerce and travel, but overall the UK is a more competitive market, especially in finance and loans. With the US market being that much bigger, you would expect keyword bids to be higher there, but the word 'loan' is $4 dearer in the UK.
Mike Grehan interjected that contextual advertising doesn't work. AdSense isn't very 'ad sensitive' - citing the example of when a story about the murder of someone by stabbing is accompanied on the right hand side with a selection of ads for knives.
Jon concluded that there is now an overall marked trend of a shift from direct marketing to search marketing, as the cost-per-aquisition (CPA) is much cheaper. I-Prospect have conducted research that found that even though search budgets are increasing, clients still don't know how to measure the value of search.
Search Engine Marketing: UK Trends In Investment & Return - Amanda Jones, iLevel
Amanda began by looking at some figures for UK Search engine marketing spend - current and forecast. At the start of 2005, it totalled $400m. By 2010 it's predicted (by Forrester Research Inc) to reach $800m. 2005 will see a 57% growth rate but growth will slow down dramatically for 2006 as the market reaches maturity. SEM as an overall percentage of online ad spend will peak in 2006 at 67%, tailing off to 49% by 2010.
Then she turned to investment growth challenges. How do we secure standalone budgets for search she asked, while preventing the cannibalization of other digital ad spend? Getting buy-in for natural (or organic) search optimization (SEO) is one way. Another is using a variety of placements and buying models - through Google, Overture, Yahoo!, and Miva, who are being joined by MSN and Ask Jeeves. Securing brand budget is a further challenge, as is working with the service providers (media agencies, SEM agencies, search engines or in-house).
While growth in supply has been steady, historical growth in supply will come from drivers that will moderate. But as with many historical trends of supply shortage, innovation will create new supply. In terms of the investment drivers, Amanda saw the drive for verticalisation as a key trend. Retail and travel are in decline in terms of the usage in traditional search engines, but the growth is in verticalisation, eg. Expedia is a niche verticalised search engine. If we are to increase supply, it's not just about getting page views but about extending and supplying offline search and using new technologies and features online.
What then are the conclusions for intelligent investment allocation, Amanda reflected, adding that local search will spur more growth in the overall search sector. The future investment driver will be heightened commerciality, including targeting and driving customer away from the web, as seen in Miva's recently launched Pay-Per-Call initiative. This will bring more non-ecommerce advertising into the sector.
Understanding searcher's behaviour is another driver she explained, as is integration; that is, understanding how search fits into the wider marketing mix, helping to create synergy across your sponsorships and events marketing. Convergence of media devices will also spur growth. Monitoring and protecting your brand also matters, Amanda explained, for preventing and counteracting negative PR in search results. Some brands have a lot of issues with their results.
Mike Grehan added that Bill Hunt & Mike Moran have just written a good new book on SEM Search Engine Marketing, Inc. The more general or shorter the search query, the more difficult for the search engine to deliver relevant results, Mike continued. By analogy, he said, you don�t stick your head round the travel agents door and shout 'holiday', or 'cheap flights!' but online is seems that people have expected this to work.
Local & International Search: the implications of search engine geotargeting - Edward Cowell (Teddie), Neutralize
In terms of local search, Teddy began, what works? At this point in time, Forester calculates that 73% of US site visitors come from search engines; 93% of whom, according to Jupiter Media Metrix, don't go beyond the first page of results. Overture estimates the local-search Web advertising market in the US will be a $1 billion market by 2008. While Hitwise says that 69% of UK searches currently come through Google.
Global considerations he flagged up started with targeting a country, sorting your own domains, taking note of the physical location of your servers and covering your key phrases. At the enterprise level approach, you need to be thinking strategically about results that are organic in terms of locality, and of 'brand scale', delivering long-term strategic presence and protecting your brand rather than spamming search engines. You can then act tactically in your bids for Adwords.
In terms of regional and local considerations, you could do worse than start by having your address and postcode on every page. Dialing codes with the phone number are also good. Teddie examined the service of Google Local. The paid local placement listings on the left hand of Google are distinguishable by the compass icon. But on Google Local they depend on proximity to the centre of the local conurbation, as most people don't put in the postcode. But even when they do, the ads are often still generic. Teddie noted that while proximity to the centre of town gets you to the top on Google Local, it doesn't make sense for the user. In Pay Per Click terms, he concluded, regional local targeting isn't yet effective.
Moving onto Google Mobile, Teddie demonstrated how organic listings get a better spread on mobile. There's lots of potential and volumes could be huge, but the way it works is far from perfect for now. Mike Grehan concurred that most sites aren't optimised for local search, and stressed that's another area we need to invest in.
Strategic Search Planning - Graham Hansell, Sitelynx
Graham began by asserting that most people see search as a technical area, not a strategic one. But TV search and peer-to-peer (P2P) search is going on now to a massive extent in order to download entire series of television programmes. These are big areas for advertisers to get into, but as yet they have hesitated, given the legality issues surrounding P2P.
Big wars are breaking out between the big search companies - but getting absorbed in these machinations is a waste of company resources for clients. People are going at it from a purely technical perspective. Instead you need to look at it from your business culture perspective, Graham argued.
Addressing search engine optimization (SEO), Graham charted the changing drivers of page relevance and asked who are we to trust? In 1994/95 authors described their own pages using metadata; then people got their fingers out and started spamming the 'keyword optimisation' process. Google then came up with the 'peer review' system of rating a page for relevance depending on what links you have pointing to it, and this formed the basis of Page Rank. Links became a judgment of relevance. But it's about the quality of the links, rather than the quantity, Graham added, ie. the traffic from and relevancy of the referrer counts for more.
The third driver for page relevance is the commitment rating - how long the site has been online and how often is it updated. Finally Graham considered personalised search. Here, instead of concentrating on what the web thinks, it�s about what the users think. Amazon's search engine A9.com, Google (with their messenger and VoIP services) and Yahoo are working hard in this area. Their add-ons help them understand what you want and what you want to find. So these are the four areas of note from an SEO perspective of what you want to be found from.
If you stick to these processes you'll generate a lot of traffic, Graham stressed, but if they are not built-in for in advance, agencies often get pushed into spamming keywords and search engines by clients. Planning and forethought are required, rather than reaction and post-mortems.
In terms of search marketing strategy, Graham said it must contain your aims (infomed by the usual demographic and market research that would go into an offline business), an understanding of the search opportunity in cash terms, and adoption of measures and key performance indicators. This strategy must be owned high up in the organisational structure, ie. Senior-level buy-in is key.
Centralised management of keywords was the next area Graham looked at, and he summarized this with what he termed the '3 Gs of Search Operation': guidelines, governance and good practice. In this last category, he illustrated his point with the following example. While on his bank's website searching for his local branch, they sent him to the local betting shop instead (ie. they hadn't updated the address - deliberately?) This shows both a lack of commitment, Graham said, and them trying to promote telephone and online banking at the expense of consumer expectation. But you can still promote these without giving people a negative experience, he explained. Why not give them a landing page instead?
Being Relevant About Relevance - David Crystal, Crystal Semantics
If you type the word "depression" into a search engine, hoping to find information on the economic sense of this word, you will be lucky if you get anything other than hits to do with the psychological sense, David commented in opening. And this word is typical. The average number of senses per word in a college dictionary is 2.5, and words are becoming more ambiguous as time goes by because of the proliferation of ordinary words as domain names. There are probably 3-to-4 senses per word on the web now, he surmised, with language becoming increasingly ambiguous.
To solve the problem of ambiguity on the web, and to obtain more relevant results, you have to get the linguistics done properly. This means, in the first instance, jettisoning the notion of 'keywords' and replacing it with the much more powerful notion of 'key senses'. There are two sides to the task. The words in the language have to be analysed into their senses, and these senses have then to be assigned to their appropriate knowledge categories. For example, 'depression' has to be assigned to the categories of economics, psychiatry, meteorology, geology, and so on. It is this combination of dictionary and encyclopedia which drives his notion of a 'sense engine'.
David started on this task in 1997, using a team of 40 lexicographers to work their way through the dictionary and assign senses to knowledge categories. It took three years and a considerable investment. The knowledge categories had evolved through classificatory work on the Cambridge family of general encyclopedias (now the Penguin family), which he edited.
The taxonomy which evolved was a combination of his own thinking and that of the various expert contributors (such as the Natural History Museum and NASA) to these volumes. When the approach was applied to the web, it needed expansion, and this took a further three years. The sense engine became functional in 2003, and is now used to provide improved relevance in such areas as contextual advertising, search engine assistance, e-commerce enquiriers, automatic document classification, and online security analysis.
The contextual solution, he commented, provides semantically relevant and useful results - results which can only be obtained by harnessing the power of human intuition to identify critical words and senses in the various domains and to assign them values based on their identifying and discriminating power. Computer-generated statistical algorithms - simply looking for frequent words, for instance - are not the solution. There are always surprises: for example, in a football report, the word 'football' may never even be mentioned. What's more, most webpages and blogs are not mono-thematic, they are multi-thematic. You certainly can't trust the headline or the first paragraph only, if you want to work out what a page is 'about'.
The Crystal Semantics approach has been fully developed only for English, but already its principles have begin to be applied to the analysis of other languages, such as Chinese. About 75 per cent of the analysis of a new language can be carried out by a process of translation from English; but the remaining 25 per cent is unique to that language, and requires fresh analysis. Identifying the nature of this uniqueness is one of the goals of their current research.
Mobile Search - Sean Walker, Overture
Sean Walker's presentation was set to cover 'Driving targeted traffic to your mobile content'. Unfortunately, Sean was unable to attend the event, but has kindly allowed that we provide a summary of his planned presentation.
In terms of the migration of search from web to mobile, Sean began from an understanding of the market. So how how does sponsored search work on mobile? First, Sean considered how Overture works on the web.
A searcher conducts a search on an affiliate site (Overture affiliates include: Yahoo, Ask Jeeves, Wanadoo, Lycos, Guardian Unlimited, MSN, Loot & FT.com). On the Overture site, businesses set up an account and submit their listings including Search-terms; Titles; Descriptions; Cost per lead. The listings are dynamically ordered by �bid price� (min 10p). The user clicks on an Overture listing, which triggers the payment of your bid price to Overture UK and a lead to your site.
Sean�s presentation then outlined a broad picture of the market in the context of the migration of search from web to mobile.
The UK and European mobile network operators (MNOs) are moving away from a walled garden approach to search and will allow users to search off their mobile environment. They are moving from being full content providers to being a transport infrastructure that provides others' content.
This is driven by the need to supply consumers with what they are searching for and to reduce cost by outsourcing advertiser management. Overture have developed a WAP Free Text Search service, to meet this demand, for the UK and European market. They estimate that by 2006 70% of mobile users will have Internet enabled phones, and by 2008 48% will be using mobiles as a part of their internet access.
Sean then outlined the different search models for WAP. First is the Restricted Access 'walled garden' model. This gives access to portal content only, but limited content often results in frustrated consumers. Within this context, the 'off portal search' function is often a bad user experience, with a big percentage of inaccessible or unopitmised-for-WAP Web results coming through on the WAP service. Such lack of control equals a risk to brands in this space.
Then the 'Controlled Access' model was considered, which allows controlled access by partners but gives priority for portal content (eg. consumables). This offers higher bandwidth revenue and an assured WAP-only experience.
In terms of what users are searching for on mobile some headline figures include: Adult - 38%; Mobile consumables - 19%; Search engines - 10%; Info services - 5%. Interestingly, searches for music downloads were only 1%.
He then explained how sponsored search on mobile works. Currently on Overture you can bid for: travel across 3 networks (Vodafone, Orange and Yahoo! Mobile). Other sectors poised to launch in October 2005 include mobile consumables, ring tones and wallpapers.
With mobile search, the advertiser uses Overture's Direct Traffic Center to manage their listings and bids. Their chosen keywords and max bid will determine which category they appear in and what their position is on the order of listings.
In closing, Sean reminds us that users pre-qualify themselves through their search, Overture will supply what they want, not push them to what we think they want. All of their advertisers are editorially checked for relevancy, legality, and adherence to codes of practice and quality standards. The model has already been a huge success for as the monetising engine for web traffic and has easily transferred to the mobile market.
Discussion:
In regard to vertical search engines, it was asked how much would the sector grow and how much importance should we place upon it. Mike Grehan said a lot of search engines are doing work on 'focused crawling' - they go out and put together a search engine that only deals with sport apparel or cosmetics.
Graham Hansell reckoned that vertical search engines are extending the use of search because they are delivering more relevant results and people therefore use them more often, such as the shopping comparison site Kelkoo. But there's also a growth, he added, in vertical search engines purely just to provide more contextual results or content for the advertisers - so a question arises here as to whether they are providing real value or just capturing real-estate twice. Well done vertical SE�s extend inventory and trust in search, he said.
Vertical search engines on the rise
Amanda Jones said that from planning perspective Google s putting a lot of investment in because in 5 years you will have Google Travel, etc. ilevel have discussed with Pricerunner how to benefit strategically from search, even though they were born from search themselves. David Crystal commented that vertical search engines haven't put a lot of effort into understanding how you will use the service. He recently searched for 'cellphone' in a vertical SE, and got nothing. The he tried 'mobile phones' and still nothing. But a lexicographer never gives up! Eventually he found that he got results for 'modular phones' and who would think of searching for that.
Teddie Cowell said that there are more local search engines as opposed to just local search. But the results on Google Local are negating its usage to begin with. This will be a big and not necessarily fragmented market, he surmised, because it will have a regular user base no matter how big or small. Also that many vertical search engines already existed and would continue to be developed, but because they would be for a specialist market may not ever by seen by general internet users, so their impact on mainstream search engines would be negligible.
Corporate versus consumer-led results
Jackie Danicki of Latitude asked about intent-driven search and spoke of the Yahoo Mindset product where user can select their intent through choosing commercial results (corporate pages, usually selling items) or research results (lots of customers discussing opinions and journalists' views). Mike responded, noting that 80% of searches online are non-commercial. David explained that there was a lot of research done on this in the 1970s and 80s, and two basic types of communication were discerned: informational and transactional. He contrasted Google's top-down approach to Crystal Semantics' bottom-up approach, doubting that the two approaches would ever meet.
Jon Myers said that from the informational point of view, people don't just go for the first or top results in their holiday search. Paid search gives you a lot of deals, natural search gives you more details on the location, etc.
Click fraud: gremlin or foe?
The extent of fraud was then raised, and it was asked if people in companies question it and is anyone doing anything about it. The question of whether anyone had received a refund from a search engine was raised and Teddie Cowell and Jon Myers were the only people to raise their hands. Teddie Cowell reasoned that for smaller firms it can be very time-consuming to research and put these cases together. But also sometimes with what looks like blatant click fraud, often the money hasn�t been debited from your account, or has already bee repaid by the search engines who are monitoring these things, so it's worth checking first before assuming something has gone wrong. Jon Myers added that Google are getting very good at detecting impression fraud. Indeed, Latitude has a dedicated Loss Prevention Officer, and collects more than �25,000 each month in click fraud refunds from the search engines on its clients' behalves.
A Latitude delegate said that all fraud should be reported to the search engines, but Graham Hansell countered that 'fraud' is a very media-emotive term, and we need to put this into perspective. Competition between companies has existed since day one, and online is no different. This is just another way of doing it. It's an issue with every media type. You have to try and control it but it's also a cost of doing business. Mike recommended looking at www.alchemistmedia.com for further information on click fraud.
----------------
About the Speakers:
Chair: Mike Grehan - VP, SMA (Search Marketing Association) UK & MD, Smart Interactive
Founder and CEO of Smart Interactive Ltd., a wholly-owned European division of WebSourced Inc., the world's largest search engine marketing company, Mike is recognized as one of the foremost SEM experts and is the author of multiple books and white papers on the topic. His best-selling second edition of Search Engine Marketing: The essential best practice guide received more plaudits than any other on the subject from the industry's leading players. In 2004, Mike was voted one of the UK's top 100 influential people in Internet marketing of the previous decade, in a poll of online marketer e-Consultancy's 22,000 UK members. Mike played a key role as founding member and promoter of the global Search Marketing Association (SMA).
Amanda Jones - Head Of Search, i-level
Amanda is Head of i-level Search, a �10m business handling the full range of Paid-for and Natural Visibility disciplines in one place. Her online life began in 1998 (when Google�s was born) at 24/7 Europe, global digital advertising sales network and technology provider. Working in London and then at their European headquarters in Amsterdam, Amanda was responsible for the coordination and implementation of pan-European advertising campaigns on client websites. Before joining i-level, Amanda was Head of Search at digital marketing agency Outrider, one of the pioneers of Search Engine Marketing services since 1995. At i-level her current client portfolio includes: Orange, Yell.com, Interflora, COI Communications, Specsavers, Starwood and Gillette.
Edward Cowell - Technical Director, Neutralize
Edward heads up the Neutralize (*\*) team of enterprise search engine optimisers. Introduced to the Internet in 1996 while working for MSN during the launch of Internet Explorer 3 in the UK, he now heads up one of the UK's leading enterprise search engine marketing agencies. An Overture Accredited SEM (Search Marketing Company), UK Premier Affiliate to Atlas OnePoint (GoToast), and member of both Nominet UK and SEMPO, Neutralize (*\*) are also active participants in the working group for the formation of SMA the Search Marketing Association UK and recently launched the Search Engine War blog. They have both highly skilled Pay Per Click and search optimisation teams. The founders of Neutralize (*\*) Teddie and Lucy were two of the earliest members of NMK back in 1999.
Graham Hansell - Founder & Head of Strategy, Sitelynx
Graham�s early experience of the Internet came whilst part of the innovative team at one of the UK�s earliest web development companies Britnet in 1995, where he worked on groundbreaking projects such as the Eurotunnel. In 1997, having identified the vast commercial opportunity surrounding SEM/SEO, he founded Sitelynx. Graham sits on a number of influential panels and boards responsible for moving the industry forward including SEMPO, SMA-UK, national and local government bodies including Greenwich Enterprise Board. In early 2005, Sitelynx was reported as one of the most up and coming agencies in the industry. Graham�s philosophy that search engines are the new audience-focussed utility of the twenty first century has enabled his accounts to accomplish swift return on investment.
Sean Walker - General Manager, Vertical Markets, Overture Europe
Sean joined Overture Services as a founding management member and Head of Sales in September 1999. Three years later was named General Manager, Vertical Markets for Overture Europe. During his tenure at Overture, the company has grown from one European office to a total of 12. His responsibilities include the development and execution of the entertainment vertical sectors of the business across Europe, and more recently, has been appointed to spearhead the company�s mobile strategy for Europe, allowing Overture to be the first global search provider to deliver paid search listings on mobile. Sean is also a Fellow of the Institute of Sales & Marketing.
Dominic Trigg - Vice President, Infospace Europe
Unfortunately due to illness Dominic was unable to attend the event. Dominic joined Infospace - owners of metasearch sites Dogpile, WebCrawler, MetaCrawler, WebFetch� and the power behind Excite� - in May 2005. He has worked in media (including the BBC) since 1991. His most recent role for Yahoo! as the Advertising Operations Director (Europe) saw him delivering advertising implementation across the various European territories and managing the portfolio of adverting sales products. He entered new media in 1996 with BT Internet and Multimedia group. Hired by Microsoft in 1997 as Advertising Director for Expedia.co.uk on launch, he went on to become Adverting Director for MSN UK. In 2001 he joined a joint venture owned by Sky, Warner and Sony called Music Choice as Media Director, moving to Yahoo! in 2003.
David Crystal - Chairman, Crystal Semantics & Honorary Professor of Linguistics, University Of Wales
David Crystal is honorary professor of Linguistics at the University Of Wales and formerly professor of Linguistic Science at Reading University. Author of over 100 books on linguistics and English language studies, such as Language and the Internet and A Glossary of Netspeak and Textspeak, he was awarded an OBE in 1995. He is Chairman of Crystal Semantics, co-founded with managing director Ian Saunders. A context targeting company, they have developed Textonomy, the first Sense Engine to deliver significantly more accurate and powerful Internet search results. The Textonomy suite of products includes solutions for search and navigation, e-commerce and contextual advertising.
Jon Myers - Director of Search Performance, Latitude
Jon Myers has worked in paid search since its inception, and has been with Latitude for more than five years. Recognised for his unparalleled experience, he was one of ten experts selected by Overture to create their SEM Accreditation scheme. He has worked in new media for more than seven years, in marketing, site development, consultancy, and design. At Latitude, Jon and his team ensure that clients� campaigns continually deliver results, monitoring and auditing the performance of all search campaigns.
See the original EVENT PAGE
Comments
You must be logged in to comment.