false

econsultancy_xml

For those of you who haven't already ready Dr Dave Chaffey's article on the "Long Tail" which he wrote for WNIM, then I would urge you to read it - <a href="http://www.wnim.com/currentissue/current/emarketing.htm">http://www.wnim.com/currentissue/current/emarketing.htm</a> The theory is interesting, but as Dave points out has actually been around for ages, however the practical applications and implicationsfor online that Dave outlines in the second half of the article are what I find most interesting. Partly because I hadn't quite thought through just how many applications there are, across a wide range of online marketing areas (e.g. e-mail, content management, search engine marketing etc.), but largely because, as a site owner, I recognise them all to be true. For example, if we look at our site traffic, measured by unique users, then it very closely resembles the Zipf’s law curve. There are a relatively small number of die-hard users and a very large number of occasional users. You might think this a bad thing were we trying to sell advertising, or as we are trying to sell subscriptions, as loyalty does not appear as high as one might want. However, we’ve not found this to be the case. You can have a highly motivated (= high propensity to subscribe or click on an ad) occasional user who has an acute need to be met, and you can have a frequent forum user, generating lots of unique usage and page impressions, who has no interest in subscribing or clicking on ads. We’ve recently been looking at measurement approaches / tools / audits (like ABCe, Hitwise, Panels etc.) largely in order to give 3rd party credibility to our own site stats. But it strikes me that these tools are still very crude because they cannot give any true measure of a site user’s intent, or behavioural propensities. And it is really these that I am interested in as an advertiser, or a site owner selling content. Metrics like page impressions, or even apparently more sophisticated sub-sets like average page views per session, are at best a mildly useful benchmark and at worst totally meaningless in terms of actual value. And yet it is these metrics which are the bedrock of commercial value exchange for online publishing. I don’t think this is actually as much of a problem as it might appear because, ultimately, everything is measured in terms of performance against other commercial metrics e.g. number of resulting sales, shifts in brand metrics. The metrics used to trade online inventory are really just proxies for other more important metrics. Certainly, when we’ve bought online advertising space in the past we’ve as good as ignored any site traffic, or even demographic, data we’ve been given and waited to see what actual results we got. As publishers ourselves, we know that there is a world of difference between quantity of traffic and quality of traffic, with ‘quality’ being understood as the degree to which that traffic is likely to do what we want them to. Dave’s analysis of how the ‘long tail’ applies to search engine marketing is also very insightful and, indeed, lends some weight to the argument that natural search and PPC need to be managed by the same agency, or team, or at the very least need to be very carefully co-ordinated. Looking at how SEO and PPC can be used tactically and strategically to target the head or the tail seems to me an excellent way to look at one’s SEM strategy overall. For example, we use SEO to target the long tail (e.g. the job title search ‘<a href="http://www.google.co.uk/search?hl=en&rls=GGLD%2CGGLD%3A2004-42%2CGGLD%3Aen&q=head+of+e-commerce+job&meta=cr%3DcountryUK%7CcountryGB">head of e-commerce job</a>’) as well as the head (e.g. the search ‘<a href="http://www.google.co.uk/search?hl=en&rls=GGLD%2CGGLD%3A2004-42%2CGGLD%3Aen&q=online+marketing&meta=cr%3DcountryUK%7CcountryGB">online marketing</a>’) but we often have holes which we need to use PPC to plug. Or, particularly at the head, there are terms that it is just not economic for us to compete on for PPC. Take this a stage further and look at the value of the traffic delivered from SEM at either ends of the tail and we find, perhaps not surprisingly, that the long tail traffic is much more valuable (more likely to convert) than the head traffic. It costs less in terms of actual click costs (for PPC) but then it costs more in terms of creating the site content and managing the PPC campaign, and the volumes are very low. The head end delivers volume but not value. Although again, with reference, to my points on the quality of traffic made above, many advertisers are (wrongly) impressed with sites that can deliver “head” end traffic whereas I think they should be advertising on sites that are strong on the tail. Obviously every site will have different levels of value across the ‘tail’ (e.g. insurance sites probably need to concentrate on the head to get volume and value) but I do think the whole concept is very powerful in making site owners (and advertisers) realize, if nothing else, that not all site visitors are equally valuable and that this curve can be used to help define one’s online marketing strategies. Ashley Friedlein CEO, E-consultancy

<FormattedContent xmlns="http://www.e-consultancy.com/schema/formattedContent/"> <Paragraph>For those of you who haven't already ready Dr Dave Chaffey's article on the "Long Tail" which he wrote for WNIM, then I would urge you to read it - <Link URL="http://www.wnim.com/currentissue/current/emarketing.htm" Window="New">http://www.wnim.com/currentissue/current/emarketing.htm</Link><LineBreak /><LineBreak /> The theory is interesting, but as Dave points out has actually been around for ages, however the <Emphasis>practical applications and implications</Emphasis><Emphasis>for online</Emphasis> that Dave outlines in the second half of the article are what I find most interesting.<LineBreak /><LineBreak /> Partly because I hadn't quite thought through just how many applications there are, across a wide range of online marketing areas (e.g. e-mail, content management, search engine marketing etc.), but largely because, as a site owner, I recognise them all to be true. </Paragraph> <Paragraph>For example, if we look at our site traffic, measured by unique users, then it very closely <Emphasis>resembles the Zipf’s law curve</Emphasis>. There are a relatively small number of die-hard users and a very large number of occasional users. You might think this a bad thing were we trying to sell advertising, or as we are trying to sell subscriptions, as loyalty does not appear as high as one might want. </Paragraph> <Paragraph>However, we’ve not found this to be the case. You can have a highly motivated (= high propensity to subscribe or click on an ad) occasional user who has an acute need to be met, and you can have a frequent forum user, generating lots of unique usage and page impressions, who has no interest in subscribing or clicking on ads. </Paragraph> <Paragraph>We’ve recently been looking at <Emphasis>measurement approaches / tools / audits</Emphasis> (like ABCe, Hitwise, Panels etc.) largely in order to give 3<Superscript>rd</Superscript> party credibility to our own site stats. But it strikes me that these tools are still very crude because they cannot give any true measure of a <Emphasis>site user’s intent, or behavioural propensities</Emphasis>. </Paragraph> <Paragraph>And it is really these that I am interested in as an advertiser, or a site owner selling content. Metrics like page impressions, or even apparently more sophisticated sub-sets like average page views per session, are at best a mildly useful benchmark and at worst totally meaningless in terms of actual value.</Paragraph> <Paragraph>And yet it is these metrics which are the bedrock of commercial value exchange for online publishing.</Paragraph> <Paragraph>I don’t think this is actually as much of a problem as it might appear because, ultimately, <Emphasis>everything is measured in terms of performance against other commercial metrics</Emphasis> e.g. number of resulting sales, shifts in brand metrics. The metrics used to trade online inventory are really just proxies for other more important metrics. </Paragraph> <Paragraph>Certainly, when we’ve bought online advertising space in the past we’ve as good as ignored any site traffic, or even demographic, data we’ve been given and waited to see what actual results we got. As publishers ourselves, we know that there is a world of difference between <Emphasis><Quote>quantity</Quote></Emphasis> of traffic and <Emphasis><Quote>quality</Quote></Emphasis> of traffic, with ‘quality’ being understood as the degree to which that traffic is likely to do what we want them to. </Paragraph> <Paragraph>Dave’s analysis of <Emphasis>how the ‘long tail’ applies to search engine marketing</Emphasis> is also very insightful and, indeed, lends some weight to the argument that natural search and PPC need to be managed by the same agency, or team, or at the very least need to be very carefully co-ordinated. Looking at how SEO and PPC can be used tactically and strategically to target the head or the tail seems to me an excellent way to look at one’s SEM strategy overall. </Paragraph> <Paragraph>For example, we use SEO to target the long tail (e.g. the job title search ‘<Link URL="http://www.google.co.uk/search?hl=en&rls=GGLD%2CGGLD%3A2004-42%2CGGLD%3Aen&q=head+of+e-commerce+job&meta=cr%3DcountryUK%7CcountryGB" Window="New">head of e-commerce job</Link>’) as well as the head (e.g. the search ‘<Link URL="http://www.google.co.uk/search?hl=en&rls=GGLD%2CGGLD%3A2004-42%2CGGLD%3Aen&q=online+marketing&meta=cr%3DcountryUK%7CcountryGB" Window="New">online marketing</Link>’) but we often have holes which we need to use PPC to plug. Or, particularly at the head, there are terms that it is just not economic for us to compete on for PPC. </Paragraph> <Paragraph>Take this a stage further and look at the value of the traffic delivered from SEM at either ends of the tail and we find, perhaps not surprisingly, that the long tail traffic is much more valuable (more likely to convert) than the head traffic. It costs less in terms of actual click costs (for PPC) but then it costs more in terms of creating the site content and managing the PPC campaign, and the volumes are very low. <LineBreak /><LineBreak /><Emphasis>The head end delivers volume but not value</Emphasis>. Although again, with reference, to my points on the quality of traffic made above, many advertisers are (wrongly) impressed with sites that can deliver “head” end traffic whereas I think they should be advertising on sites that are strong on the tail.</Paragraph> <Paragraph>Obviously every site will have different levels of value across the ‘tail’ (e.g. insurance sites probably need to concentrate on the head to get volume and value) but I do think the whole concept is very powerful in making site owners (and advertisers) realize, if nothing else, that <Emphasis>not all site visitors are equally valuable</Emphasis> and <Emphasis>that this curve can be used to help define one’s online marketing strategies</Emphasis>. </Paragraph> <Paragraph>Ashley Friedlein<LineBreak /> CEO, E-consultancy </Paragraph> </FormattedContent>

2005-08-05T10:02:48+01:00

true

1356

3619

102571

2005-08-05T10:02:48+01:00