Found this to be a really interesting article on some sleuthing regarding how Netflix recommendations work and a peek inside some of the insights from their VP of Product.
Found this quote from a data researcher at Microsoft interesting:
I don’t make decisions from my gut. I make decisions based on as much information as possible, and then I make them with the least emotion I can
Here is a list of WordPress plugins I’ve found useful:
- Adsense Explosion
- Contact Form 7
- Creative Commons Configurator
- Exclude Pages from Navigation
- Google Analytics for WordPress
- Google App Engine for WordPress
- Jetpack by WordPress.com
- Memcached (and no this doesn’t activate!)
- Meta Tag Manager
- Pushover Notifications
- Redirection (nice, not problem free but allowed me to 301 my old URLs)
While doing some SEO research this morning I ran the Google search pictured below…
Having spent considerable time building sites/businesses that rely on Adsense as part of their revenue stream I can think of a number of Google algorithmic penalties this ad/content ratio seems to violate.
Relevant article from Search Engine Land:
@rustybrick What I said (“even if we didn’t block our results, the algo wouldn’t fire”) is true, b/c most Goog searches have no or few ads.
— Matt Cutts (@mattcutts) June 14, 2013
Matt goes on to add this disclaimer:
— Matt Cutts (@mattcutts) June 13, 2013
Hmm, guess I’m a cynic though I just tried lots of searches, most with brand names as I’m working on a shopping site, all had ads just like what’s pictured above. While percentage wise there are probably large numbers of searches that do not have ads that doesn’t mean ones that do should so blatantly violate the ad/content ratio.
[Updated: Dec. 6, 2013] Moz Blog has an interesting article speculating on Google’s next steps in 2014 with a nice analysis of these relatively new ads. Peter goes on to add:
This new format has been running on mobile browsers for a while now, and Google’s widespread testing makes it look like a foregone conclusion for desktop search. This change will have huge implications on both organic and paid CTR in 2014, regardless of the final form.
Dr. Peter J. Meyers – November 2013
At a banquet this evening our VP of Engineering Bob Clark picked up Wanderful’s
Innovate North State Tech Company of the Year Award!
Ok, so it’s a really nice review so just had to share. Also, testing out the Link Post Format on WordPress (need to Google what that does).
Having recently migrated all of my blog/site content to Google AppEngine away from my prior $10/month hosting I was curious about what the monthly cost of running the site would be. My blog went live on AppEngine on Nov 13th and with the default instance settings was running ~$2.00/day at roughly 50-55 instance hours the vast majority of which is servicing requests from crawlers.
Tuning Google AppEngine, For Cost
As mentioned above I started with the default settings and the first significant change was to set Idle Instance Max to 1 as running multiple instances chews through the free 28 hours at an xN multiplier rate. This change alone had a significant impact on reducing overall instance hours though clearly impacts the performance of my blog. I’ve tested the performance and for a blog I feel it’s suitable so I’m comfortable leaving things at this level for awhile.
robots.txt and dos.yaml
I’ve been running another site on AppEngine which has both significantly more pages and traffic than my blog though I’ve been able to optimize it to stay under 28 instance hours and one key factor has monitoring bot traffic and blocking via either robots.txt or dos.yaml.
Google Cloud SQL Costs
So far for the month of November which is missing roughly 12.5 days worth of actual traffic to my blog I’ll be around $15. In addition to normal site traffic (crawlers & users) I’m currently running JetPack which appears to ping back to my domain for tracking stats etc. I’ll need to evaluate the possible benefits of using dedicated memcache over the Cloud SQL expense as most of the site’s content is static. I’ve configured WordPress as recommended by Google using Memcached plugin as well as Batcache.
I’ll admit I mistakenly hadn’t thought too much about SSL prior to finally switching my domain over to Google’s servers. At that point, I faced the issue of having to buy SSL support for AppEngine which would certainly add to the monthly cost of running my blog. Seeing as how I’m the only administrator this wasn’t exactly an attractive option given the traffic challenges my blog has faced coupled with the desire to keep costs down. I’ve instead opted to use the free SSL on the appspot.com version of my site to manage WordPress administration which works *fairly* well though there are places (like the Themes Editor) which expect to load all content from the hosting domain which in my case doesn’t work. I suspect there are other issues here as well that I simply haven’t run into yet but time will certainly reveal those issues too.
At this point, I’m fairly happy with the switch as WordPress is far more advanced that BlogEngine.NET 2.X not to mention it has clients for iPad and iPhone which is likely where I’ll be posted from the vast majority of the time. That said, I’ll continue watching the hosting costs and learning more about the SSL issues.
Updated Results (Dec 6, 2013)
I’ve been running with the above modifications since Nov. 21, 2013 and my blog is running about 24-26 instance hours per day, within Google’s free 28 hours. It looks like my associated Cloud SQL instance is running about $14-16/month which is roughly comparable to where I was hosting before. Prior to these changes my instance cost was nearing $2/day and typically exceeding 50 instance hours.
Wanderful has just released an updated version (v1.0.3) of our iPad Discovery Shopping App Find&Save for iPad on the App Store on iTunes.
Now for the bits interesting to the iOS dev crowd regarding the approval process.
Submitted Nov. 20, 2013
Reached review status 6 hours later on Nov. 11, 2013
Reached Process Status on Nov. 24, 2013 9:37pm
Available for download on Nov. 24, 2013 10:48pm
Means we got a 5 day turnaround time which is not bad headed into the Holiday week.
My team at Wanderful Media has built two iOS applications for our Find&Save Discovery Shopping experience. Our iPad app is being featured in the AppStore and our iPhone just became available for dowload:
If you don’t want to mess with an app install or are on an Android phone check out our latest Find&Save geo-local Mobile Web implementation.
Now that I have my blog moved I’m looking to post more about how we’re building a discovery shopping experience for over 400 of the nations newspapers which looks a little something like this:
In 2004 I started blogging on my own domain using discountasp.net as they offered a competitive hosting package based on Microsoft technology. Fast forward nearly a decade and I haven’t used Microsoft technology for nearly 5 years and maintaining my Blogengine.NET blog had become an increasing challenge. Upgrades were very difficult and I found myself stuck on a tweaked version that couldn’t have easily be upgraded. Recently, Google added support for PHP on Google Appengine and posted steps to get WordPress running so I applied for and got early access and I’ve been slowing migrating all of my content.
I’m writing this post prior to completing the migration as the number of steps involved has been growing and I figured it might be helpful to track what was required so I don’t forget and to share what I found. The list here is not exhaustive by any stretch and it’s likely written in such a way for me to remember than perhaps general utility though nonetheless here it is.
Once I got a basic install working which was rather straight forward I spent several days getting reacquainted with WordPress which I’d used many years go though only locally. Needless, to say it’s come a long way and with an iOS client on both the iPhone and iPad it looks really promising.
- Memcached plugin (and no this doesn’t activate!)
- BlogML pluggin
- Redirection plugin
- Creative Commons
- Google Analytics for WordPress
- Adsense Explosion
- Pushover Notifications
Importing My Blog Content
First was the task of researching tools to import BlogEngine.NET’s BlogML format for which I found a plugin. The plugin wasn’t just drop in and go and I ran into issues with it processing my BlogML.xml file which required manually editing the XML as well as finding and patching XPath.class.php to v1.148 as it was outdated in the sources I’d found.
I also had to patch function parse_blogml in blogml-importer.php to read my BlogML.xml file from my WordPress dedicated Google Cloud Storage bucket like this (and yes, figuring out this URL took a fair bit of Googling where wpsdt-blog-files refers to my Google Cloud Storage bucket):
$sXml = file_get_contents('gs://wpsdt-blog-files/BlogML.xml');
Next I found my BlogML.xml file was too large so I had to chunk it down into several different smaller files in order for the import to proceed a process which took a long time as AppEngine repeatedly timed out importing the .xml file though it eventually worked. Oh yes, it took quite a bit of trial and error to figure out the various file permissions that had to be granted to get access from a WordPress plugin to Google Cloud Storage including the fact that I found myself cursing Google’s new Cloud Console with buttons to change permissions that only appear once you check the bucket you want to change. Perhaps needless to say it would have been awful nice if they were just disabled so you at least you knew they existed.
Setting up Redirects
I exported BlogEngine.NET to BlogML.xml and extracted all post URL’s to build a CSV mapping to import into the Redirect plugin as WordPress and BlogEngine.NET don’t have compatible slugs. (I’m still manually correcting the dash “-” placement in over 600 slugs as no simple pattern would have worked).
Setup redirect for BlogEngine.NET dynamic image serving /blog/image.asx?picture=(.*) -> /AppData/files/\1
Setup redirect for blog archives /blog/Archives.aspx -> /blog/2013 (close enough)
Redirect BlogEngine.NET syndication link /blog/SyndicationService.asmx/GetRss -> /blog/feed
Fix up FeedBurner RSS source link /blog/feed
Set Site Address (URL) to /blog so I could continue to use appspot.com’s support for SSL for site administration without paying for an SSL cert on www.stevetrefethen.com and saving myself the cost of the cert and what looks like $9/month for SSL support on AppEngine.
Of course, I learned a hard lesson along the way by changing both the WordPress Address and Site Address URL’s in such a way that created an infinite redirect loop. To fix this I had to install Google’s CLI tools for accessing MySQL on Google Cloud SQL and manually update the wp_options table to correct my mistake.
School Blog Content
While I was working on the the Scotts Valley Parcel Tax effort I setup a secondary BlogEngine.NET install (before it supported multiple blogs) and all of that content had to be exported, imported, and redirected as well so repeat many of the above steps.
I was a fan of ScrewTurn wiki and had been running a few installs of that as well so I moved a number of those pages here manually largely but cut/paste as there is/was not export/import available.
Handling Static Image Files
Over the years I’ve had various image files scattered around my site thus I had to copy these paths from my old site into the new WordPress install which included:
Upload various media files to my WordPress Google Cloud Storage bucket.
Domain Name Transfer
I’d not done a domain name transfer before so this was my first such experience and while it all went smoothly the only problem was a very poorly written email from OpenSRS (whose website looks like it’s from the early 90′s) that mislead me thinking the transfer would occur after 7 days. The email read (emphasis mine):
If you wish to cancel the transfer, please contact us before Tue Nov 12 04:13:22 2013 by going to our website, <link removed>
Needless, to say I didn’t click the link until after the deadline passed only then to realize the link had a checkbox I needed to click and submit.
Years ago I’d gotten a Google Apps for Domains account for my Gmail hosted personal email address so once the domain was transferred I just needed to edit the zone file and get it pointed to my WordPress AppEngine project.
After running into a few problems I found his issue with app.yaml
# 09/29/2013 # https://code.google.com/p/googleappengine/issues/detail?id=9972 # missing handler - url: /wp-(.+).php script: wordpress/wp-\1.php - url: /(.+)?/? script: wordpress/index.php
New Contact Form
After a bit of digging I found a plugin called Contact Form 7 which I’ve used to replace my old BlogEngine.NET contact form.
All in all I’d say this was a pretty damn painful process with a lot of trial & error that took me a few months of late evenings to get going. I’ve been blogging since June 2004 and accumulated an awful lot of cruft over the years that I simply didn’t remember which didn’t help.
The site has now been running for a few days and I’m in the process of trying to reduce the cost of running it on AppEngine as it’s currently over $2/day whereas discountASP.NET I was running about $20/month with SQL support. Just applied a few changes based on this post.
Found this post which I totally agree with.
[Updated: 12/15/2013] Added Contact form info
[Updated: 01/07/14] Updated Google AppEngine plugin to 1.2 running on 3.6.1