| LED Digest 2252: Testing Web Metrics Accuracy? |
|
|
|
Shaun says, What is the gold standard for traffic analysis? I pay for web metrics. Wanting to check on its data I realized there is nothing to compare it to. Is there any independent check on web metric accuracy? ================================================== The LED Digest Moderated Discussion List "Effective Online Advertising, Since 1997" Data > Information > Knowledge > Wisdom pair Networks: The LED's Web Host Hosting and Domain Registration from a Trusted Leader pair.com for Hosting | pairNIC.com for Domains ================================================== List Moderator: Published by: Adam Audette LED Digest adam, led-digest.com http://www.led-digest.com .............................................. September 22, 2006 Issue no. 2252 .............................................. .....IN THIS DIGEST..... ====== NEW ====================== --== Testing the Accuracy of Web Metrics Data ==-- ~ Shaun Johnston "Is there any independent check on web metrics accuracy?" ==== CONTINUING ================= --== Creative Linking ==-- ~ Marty R. Milette "...the real problem is having long, nasty, parameter-filled URLs." --== Are Meta Tags Unnecessary? ==-- ~ Roy Williams "The Title tag is a meta tag, and the title tag is not a meta tag. Confused?" ~ Ivan Jimenez "...there IS a 'title description' within the meta tags but you shouldn't use it." ~ Bob Sheridan "I recently re-created our company Home page..." ~ Mike Banks Valentine "Convinced yet that keyword meta tags are pointless?" --== Small Biz & Search ==-- ~ Mark Whitman "...you can also get top 10 index placement without a single HTML character..." ==== BILLBOARD =================== --== Update: Site Hijacking ==-- ~ Peter D'Aprix ========== NEW =================================== From: Shaun Johnston Subject: New Topic - Testing Accuracy of Web Metrics Data What is the gold standard for traffic analysis? I pay for web metrics. Wanting to check on its data I realized there is nothing to compare it to. It claims to be more accurate than log file analysis because it avoids the possibility of caching. Dispatch of a graphic should match very closely the viewing of a page by a real person. This issue matters to me because I charge my customers for visitors delivered, PPC. The question arose because I have log data for two ends of a single stream of visits. So I can compare visits to the sending page on one site, and referrers of visits to the receiving page on the other site. The web metrics figure for referrers of visits received is greater than that site's log files show, which is OK -- that's what you'd expect if that page was being cached somewhere, those visits wouldn't register on the hosting server. But the web metrics figure is also greater than the number of visits apparently being sent -- visits to the dispatching page recorded in the dispatching site's log files. Conceivably that page could be cached, too. But that is a PHP page that is fed a string of parameters from which the address of the receiving site's home page has to be compiled. What I'm counting is number of visits to that PHP page with the correct parameters. Could more visits be actually sent from that page than are recorded in my log file as visits to it having the right parameters? Can you help me with this aspect of server behavior? For a PHP page that has to compile a destination address from parameters, can such a page be cached? My web metrics service says it can, and guarantee that their data is right, even though their figures range from 85% under to 130% over what log file analysis shows for visits. Is there any independent check on web metric accuracy? Shaun Johnston ======== CONTINUING =============================== From: Marty R. Milette Subject: Spidering Database-Generated Pages > I have a client who got a link from a company > with a high pagerank site that is very relevant. > The problem is that the link is on a page that > is accessed through a database query... - Dave Roberts, LED 2248 Database-generated pages aren't a problem -- the real problem is having long, nasty, parameter-filled URLs. Many content management / page generation systems will generate URLs in links like: http://www.mysite.com?sec=123&pg=456&session=4BDE1234545A3D3893A The parameters tell the spider the page is database-generated and it can easily be ignored. The real killer is the session ID which is different every time the page is viewed. The solution is pretty easy (get your favorite geek onto this) -- store session IDs in session cookies, and use a rewrite function to get rid of the parameters from the URL string to make it look something like: http://www.mysite.com/hotels/grand-hotel-europe/ This not only gets the site spidered, but also puts your keywords in the URL -- never a bad thing for humans OR spiders. I have many sites that use this technique and all the pages are spidered regularly. Regards, Marty R. Milette -------- new post - new topic -------- From: Roy Williams Subject: Meta-tags > Don't forget that title tags aren't meta tags... - Nathan Holley, LED 2251 The Title tag is a meta tag, and the title tag is not a meta tag. Confused? <title> the good one<title> <meta name="title" content="blah blah"> Best forget the 'meta' one, then.... Real gone, Roy Williams Nervous Records www.nervous.co.uk -------- new post - same topic -------- From: Ivan Jimenez Subject: Meta Tags > ... implementing meta tags - more specifically the keyword > meta tag - is not going to have any affect on your rankings. - Jill Whalen, LED 2248 To add to Jill's post and possibly answer a few follow-up questions, there IS a 'title description' within the meta tags but you shouldn't use it. I used to use it but some search engines now view it as over saturation and may penalize you if they find it within your code... Ivan Jimenez Smarter Clicks : Search Marketing Solutions http://www.smarterclicks.com -------- new post - same topic -------- From: Bob Sheridan Subject: Meta tags > ... implementing meta tags - more specifically the keyword > meta tag - is not going to have any affect on your rankings. - Jill Whalen, LED 2248 Hi Jill, I recently re-created our company Home page using a Template I use for new pages (no title, description, etc.). After creating the new Home page I published to our website. A day went by and I searched Google for "Restaurant Software" and instead of seeing my site listed in its usual 5th position, it had moved up to number 4, with Google substituting "RestaurantPlus" in the Title. MSN has me in top 10 with Title "Template". I immediately made fixes to our Home page Title and Description coding but thought it was interesting. Apparantly "content" rules! Bob Sheridan, Webmaster RestaurantPlus www.restaurantplus.com -------- new post - same topic -------- From: Mike Valentine Subject: Meta tags > ... implementing meta tags - more specifically the keyword > meta tag - is not going to have any affect on your rankings. - Jill Whalen, LED 2248 Both Kym McLaughlin and Cynthia Copenhagen quoted Jill Whalen's above comment suggesting they believe that keyword meta tags still have an effect on ranking. But then both backpedal on that statement by saying other factors are more important to ranking well - but meta tags still matter? Title tags still matter critically, description tags may, sometimes matter ... a little, but keyword meta tags?. It's long past time to stop wasting energy on that tag and I hope you'll both spend your energies elsewhere in the future. It's old news that keyword tags are worthless. I wrote a keyword meta tags tutorial article in 1999 or 2000 listing the top 8 ingredients to ranking well in the search engines. http://website101.com/metatags.html I went back today to review it and found that remarkably few things have changed. So I massaged and edited a few things, added a small type disclaimer ;-) and linked to some online tools from that page, including a "Search engine simulator" tool and an "HTML source code viewer" and a previous article I'd written in 2002 called "SEO Keyword Voodoo: Invisible Meta Tag Mumbo Jumbo" because, as you might guess from the title, I am 100% with Jill Whalen on this one. http://website101.com/Search_Engine_Positioning/keyword-voodoo.html Keyword Meta tags went south in about 1998 or so when webmasters began keyword stuffing. They are not used, nor are they trusted at all by any search engines. If they are not purposely abused, they are completely misunderstood by webmasters still. Don't waste your time fussing with them or tweaking them. They're an old and badly abused element of HTML code that no longer does a thing for ranking. To further prove my point, I've created a Rollyo search tool (at the bottom of both articles pages linked above) which allows you to search only the top 25 top SEO bloggers, and includes the Official Search Engine Blogs, for the search phrase "keyword Meta tags". At the time of this writing, only 9 results came back on that search. After visiting each of them, I discovered that most of those results are either derisive comments about "keyword meta tags" from the SEO blogger or are derisive comments left by visitors about the folly of meta tag massaging! My favorite blog post from among the results of that Rollyo SEO Bloggers and Search Engine Blogs search are from Aaron Wall of SEObook (sorry another guru) when he said, "an old client wants me to rewrite their meta tags. A total waste of time, but if it makes them happy, oh well...Meta tags? ... hehehe" Convinced yet that keyword meta tags are pointless? Mike Banks Valentine http://realityseo.com -------- new post - new topic -------- From: Mark Whitman Subject: Small biz search > ... creating a web page that is graphic only or > Flash only is instantly a rather poor judgement > to make when considering the WWW. - Rick Gortatowsky, LED 2250 As your long summary of website design 101 indicated, a site should be targeted to a specific purpose and sometimes demographic group. Using graphics only and Flash only has gotten excellent results when used for the purpose of being viral marketing tools. Viral marketing techniques, for certain products, are great (if successful) since you don't have to bow down to SEs but you can also get top 10 index placement without a single HTML character on a page. How? Well, that's something I had to figure out myself and leave to you to do the same however I've seen plenty of other people using what appears to be the same or similar techniques. The point is, poor judgement regarding website design would be dependent on whether a site meets or exceeds its goals, period. Conforming to simplistic design dogma for it's own sake can be limiting and counterproductive which in my opinion is a matter of poor judgement. M.Whitman ==== BILLBOARD =================================== From: Peter D'Aprix Subject: Update - Site "Highjacking" Conclusion to the "Porn Parasitic Attack" LED 2225: I know this thread has been overlaid by many other threads since I posted my problem back on 15 August this year. Essentially, the space I had with my host server, seanic.net, was hacked or entered by someone who had either used a password cracking program or had gotten my user ID and Password illegally (not from me) and placed several folders of many hundreds of porn link pages (no images) on my servers space. Some in upfront folders, other tucked deep in the recesses of my own folders for my site gourmetvoyageurs.com, a food and travel site - sensous, yes; porn, no. I had cleansed the host server of all files replacing the entire site with an FTP upload of my reworked site files (I chose this time to restructure my site since I had disappeared from Google anyway). I notified seanic.net about the problem and received a tepid reply that my site was hacked (definitely not an inside job they said) and I should remove all the unwanted files and folders, change my password to a random collection of characters and sit and wait for Google to come back again. They did not keep access logs for more than 2 weeks; it had been 6 weeks when I discovered the problem. I had already removed the porn folders, re-uploaded my entire site and had changed the password. But Google did not come back for a visit. I posted this experience on LED and received sympathetic and very useful suggestions from many LEDers like Cheryl Berry, but the most fruit bearing was from Janell Vasquez (thank you Janell!) which was to go to http://forums.searchenginewatch.com/showthread.php?t=12924 and post my story, which I immediately did. A Google guy, Brian White, replied. You can see his reply at this link. He had checked and saw that while Google still had my site indexed, it was blocked. He recommended applying for "reinclusion" with Google. The first time I tried that (it is deep in member only section of 'webmaster tools" of Google AdWords or AdSense account pages) I got nowhere. I waited a month and nothing happened. I remained blocked. One day I was uploading some new stories (I was still listed with Yahoo and the other search engines, but Google has been responsible for 85% of my traffic) and discovered that when I uploaded a folder / directory filled with pages and images which should have replaced the older folder / directory contents, it only added the new or changed content leaving all the rest. This was new to me since in the past on the same server, same account, uploading a folder replaced all contents of that folder on the server. So having thought I had replaced all the old, possibly invaded / hacked folders / directories on my server, I had the uncomfortable feeling I had lost control of my server space. I dug as deep as I could into the maze of folders and sub folders but could not find anything lurking. But I was convinced that Google had found something it did not like when it came back to re-spider. During this time, I received an email from another small company who had seen my post on www.searchenginewatch.com and said they were with seanic.net as well and had also been hacked just like myself but a month earlier in May. Now what are the chances of that? So I took the loss (I had paid a year in advance with seanic.net) and switched the site to a different host server (www.addr.com). Then re-applied for "reinclusion" and within 2 days my site was back up with Google. Hmmmm! I am not technically trained enough to search the internet and access logs to prove that my host server had someone who gained entry to my space to place the porn. Actually I can't prove anything about anything. Like the other party who had the same experience with the same host server, we are both in publishing in niche markets and neither of us has the time to spend digging or to train. With the help of a WISIWIG site builder, we can design and publish a site, do some modest SEO (thanks to all the tips on LED I might say!), but we are working 7 days a week now and cannot train to do the technical things. Actually our minds do not work in that direction anyway. I have a lot in common with the post from the store front owner who makes the time, but does not have the cash to pay the experts. I too do my own plumbing and remodeling. When I don't, I always have to finish off what the pro's did wrong regardless. Anyway, how can you pick an "expert" out there when the bloody battles here on LED show that there are as many opinions on how you should do things as there are "experts" to express them? One wonders, especially with this debate in Congress that has been slightly referred to on LED, whether the little guy with shallow pockets may be pushed out of what has been an equal opportunity virtual space for almost 10 years. It will be a shame if small, specialized niche sites are squeezed out. But for the moment, I guess we should make random passwords, no whole words. Then we have to store them on our computers where some hacker can find them. But we will need to change all of them every few weeks. Hang out at frequent searchenginewatch.com to keep up on the latest goings on. I suppose we can do that instead of watching CSI. Then if we have a problem, make sure we frequent searchenginewatch.com again to find solutions. Brian White of Google sure helped me and there are others! I hope my experience can help anyone else who is unfortunate enough to have this happen to them. One person who posted had had his sitemap hacked and links to porn pages inserted. At least that is what I understood the post to say. And then I think I understand a variety of people to say that whoever it is that "attaches" themselves to other peoples sites or hosting, then visits blogs, forums etc. and posts links to their "parasitic" hack pages and these links on long after the affected sites have corrected the problem(s) thus continuing to victimize the person who thinks they had done what they needed to do. Do any of you have ideas or experience with this? Can you shed more light on the subject? OK. Enough. You are probably wanting to get back to Click Fraud or Tags, far more interesting subjects. But thanks for bearing with me here and for your valuable advice. Peter D'Aprix ------------------------------------------------------- The LED Digest is sponsored by pair Networks: pair.com for Hosting | pairNIC.com for Domains Copyright 1995-2006 Orange Wheel, LLC. All Rights Reserved. ----------------------------------------------------------------- "The more sand that has escaped from the hourglass of our life, the clearer we should see through it." - Jean Paul |
