Random Acts of SEO

Today’s broadcast:


Website details: Bank of the Poor
Hangout Notes by Alistair Lattimore from Convergent Media

URL Canonicalisation
The domain you asked for review on was bankofthepoor.org, however they want to be using bankofthepoor.com. This is evident throughout the internal pages of the site that carry rel=”canonical” tags reflecting bankofthepoor.com. Incorrectly configuring their rel=”canonical” tags has resulted in the home page being indexed under bankofthepoor.org, while all internal pages are indexed under bankofthepoor.com.
Errors – 404
I got a lot of 404 errors while crawling their site, primarily:
  • query string variations that errored
  • invalid URLs within CSS files
  • Fix the invalid URLs within CSS files, they are slowing the site load time down and could be making the site visually look incomplete
  • Fix the internal links with invalid/missing query string values to avoid users stumbling into a 404 error while exploring the site
<title> tags
Found half a dozen URLs with no <title> tag at all.
They don’t look like high visibility pages, so likely won’t effect anything.
  • Check URLs for traffic, adjust if appropriate
<h1> tags
Lots of URLs don’t have an <h1> tag but use <h2> instead, probably isn’t going to effect their rankings.
  • Fix WordPress template when fixing rel=”canonical” issue above
meta description
Meta description tag was missing from over 100 URLs, many were important.
  • Add custom meta descriptions to all high traffic URLs
Invalid HTML
I found lots of invalid HTML throughout the site, isn’t going to effect search efforts but if less tolerant browsers it could cause content to look malformed.
  • Review HTML on high value pages to correct any rendering issues if present
Missing <img> alt attribute
Lots of images that are missing alt text, could be a missed opportunity.
  • Add descriptive alt attribute to all editorial images
Non-descriptive Link Text
Found a few ‘click here’ style link text scattered through the site.
  • Rework content to use a more descriptive link text
Sign Up
Signing into bankofthepoor using facebook results in a programming error, which is going to be effecting the volume of donations they are receiving.
After filling out the sign up form & submitting it, no account was created. I didn’t receive an error stating that the form had an error, no email was sent asking me to confirm my account. The form reloaded and was blank, so I’d need to fill it out again if I wanted to proceed with registration.
Sign up process requires to much personal information, which they don’t need to know. Instead they should streamline the registration process to email & password, then ask for other information after they’ve made their initial donation to remove friction/effort.
They are asking to donate money but I couldn’t access their site using HTTPS. It’s possible they use a third party payment system or a sub-domain for payments, which makes the comment less relevant. However given that they are asking for money, I think it is important that they have HTTPS available even if they aren’t taking the money themselves – so users know they take their security seriously.
Multiple references to the same script with different URLs
No gzip compression
No CSS sprites
No merging of script resources
No HTTP caching headers for static resources
Serving large images scaled in HTML
  • Install W3 Total Cache as a starting point
  • Remove duplicate references to the same resources
  • Generate CSS sprites where appropriate
  • Tune additional items with .htaccess where possible
It is growing like a weed, they need to have a responsive design.

Random Acts of SEO
Website Details: http://bankofthepoor.org/

Previous Sessions

Helping charity and non-profit websites reach more visitors through SEO, user experience and strategy.

  1. http://www.deafness.org.au/
  2. http://www.starlight.org.au/

Hangout Panel

Karol Dziedzic's profile photoDavid Amerland's profile photoJim Munro's profile photoArthur Radulescu's profile photoRob Maas's profile photoDaylan Pearce's profile photoAlistair Lattimore's profile photo

Hangout Notes

by Alistair Lattimore

The following list shows the number of indexed pages per domain:
513 site:starlight.org.au
467 site:www.starlight.org.au
40 site:starlight.org.au -inurl:www
1 site:volunteers.starlight.org.au
32 site:forms.starlight.org.au
7 site:starlightday.starlight.org.au
http://www.starlight.org.au & https://www.starlight.org.au are indexed
http://www.starlight.org.au & https://www.starlight.org.au are completely different sites
http://www.starlight.org.au & http://www.starlight.org.au/Pages/default.aspx are indexed
www/non-www are both crawlable by search engines
using a 302 redirect from http://www.starlight.org.au to http://www.starlight.org.au/Pages/default.aspx
broken URLs return a HTTP 404, however the 404 error document is horrible
Severe HTML validation issues with duplicate <head>, <title> tags throughout website
A lot of short <title> tags that could be more descriptive/optimised
home page meta keywords tag keyword stuffed
pages that have attempted to be de-indexed using meta robots tag are indexed due to the tag being invalid. They are using a value of NOHTMLINDEX, which appears to be the tag used to remove a page from the Sharepoint indexing service. However that obviously isn’t valid for search engines. this issue opens them up to their search results page, as an example, producing thousands of indexed URLs (it hasn’t yet but it could).
no robots.txt file
no XML sitemap
rel=canonical tag not implemented
both the static & animated favicon URLs lead to a 404 error
both jquery & js carousel have a lot of 404 errors throughout the site, so functionality is broken
pressing <enter> on search box instead of clicking Search, redirects to http://starlight.org.au with an error instead of to http://www.starlight.org.au
facebook widget used in right sidebar showed me an ad leading to widgetbox.com
facebook widget includes <noscript> follow links to widgetbox.com
Google Analytics is still using synchonous tracker, should upgrade to asynchronous tracker
Google Analytics is used on multiple sub-domains (good) but isn’t implemented for multi-domain tracking so attribution is going to be broken.
found around ~50 internal links using non-descriptive link text like ‘click here’, ‘here’, ..
internal canonical link issues due to inconsistent casing of URLs
The list will keep going if I keep digging but I should get some sleep.

Dan Petrovic, the managing director of DEJAN, is Australia’s best-known name in the field of search engine optimisation. Dan is a web author, innovator and a highly regarded search industry event speaker.
ORCID iD: https://orcid.org/0000-0002-6886-3211

0 Points