WordPress Backup to Dropbox 1.8 has been released and contains a few minor bug fixes and style updates for WordPress 3.8.
Compatible with WP 3.8
The plugin’s core code is separated from the WordPress codebase using the computer science patterns facade and factory thus making it unlikely a WordPress core API change will affect WPB2D.
However, this time I did need to make a few CSS style fixes to match the new WP 3.8 interface, that I am not sold on at the moment.
Fixed menu ordering to better avoid clashes with other plugins
I found a nice little plugin clash that meant that if two plugins landed in the same menu spot only one would display. I have made some changes to try and avoid this.
Fixed PHP notice in Dropbox Facade
Most hosts have notices off, however I treat them as errors and squash them when I can.
Namespace CSS so it’s less likely to clash with other plugins
Sometimes the CSS selectors I use in my admin section are not specific enough to target my plugin only and can cause issues with other plugins. To avoid this I have updated the specificity of all my selectors to reside under a #wpb2d id namespace.
Fix a potential issue where some tables and be missed in backup on resume
This would only affect users whose backups timeout during a database dump in certain circumstances. So 99% of you should not have even noticed this one.
While the WordPress core team have been working hard on WordPress 3.7, I have been working equally as hard on WPB2D 1.7! This rerelease has numerous bug fixes, increased performance and a few UI tweaks.
Database dumps are now tracked so they can be resumed
I have implemented the same resume functionality that is used for the file upload to the database dump. This will make the plugin more robust and capable of resuming the database dump if it times out at this point.
Improved error handling around extension installs
This will help users self diagnose issues related to installing premium extensions.
Updated tests for better code coverage
I have a test suite for the plugin that I had neglected for a while. I have finally had the time to update it and get all the tests passing!
Implemented a factory for better dependency injection
This is more for the nerds out there! WPB2D now has a factory class that allows for dependency injection that leads to better tests.
Updated the Email Extension to use wp_mail
This is the recommended function to use when sending email and it means that other plugins like WP SMTP can override the settings if need be.
Added a time stamp to Zips generated with the Zip Extension
This will allow for multiple zips to be created on the same day.
Improved logging and estimation of percent complete
In order to help me diagnose problems I have introduced better logging and a better attempt at guessing the percent complete.
Add ‘Settings saved’ message and use default WP errors
I have implemented the WordPress standard notifications for errors and success messages.
Fix invalid UTF-8 sequence in argument error
This error relates to json_encode attempting to encode a UTF-8 character and failing. I have implemented a known work around to ensure that the error does not show in the error logs.
For quite some time I’ve had a hunch that using a forum for customer support is a very bad idea.
Two years of customer support has yielded the following forum statistics:
- 543 threads in total
- 2417 replies in total, thats 3.3 a day
- 827 times I have replied to users, thats 1.1 a day.
That’s a fair bit of work for one person! Interestingly, once you drill down into these numbers you soon realise that using a forum for support is an inefficient waste of time.
- 8% of the threads I neglected to answer. Sorry guys!
- A whopping 71% were answered by me only. This coupled with the fact that there were so many repeated cases leads me to believe the user would have been better off sending me a direct email
- That leaves only 29% that might be useful content for a forum.
So, if 71% should have been handled by email and has no place in a forum, what about the remaining 29% of queries?
To dig a little deeper, I manually went through the 159 threads with more than one response and categorised them like so:
- 2% of this subset were specific feature requests
- 4% were other weird stuff of no use to anyone
- 5% included a post from a freelancer I employed for a month that didn’t work out
- 8% contained posts from community members attempting to help which I am super grateful for. Unfortunately, sometimes the help wasn’t quite right
- 8% contained information that was useful to other users. Hurray! Finally content worthy of a forum
- 18% were users adding a new post onto the end of an old one
- 20% where users who had similar issues
- 35% contained a little bit of the last two.
The last three points are interesting because they are where things started to get messy. The fact that 18% of people did not know how to use a forum and just added their issue to an active thread was a pain in the arse. Not only are you trying to help the first person, the next one is confused by your answers that has nothing to do with their questions.
Things got worse with 35% of the threads that just turned into a hodge podge of somewhat related issues with different outcomes that required different answers. Not to mention other users confusing the shit out of each other with half baked attempts to help each other out.
You could argue that the 20% of people that had similar problems gained value from my response, thus being a worthy use of a forum. Even so, I would have still preferred that they created a separate thread so I could have addressed their problems individually.
Or, even better, sent me an email because a lot of issues are user-specific and can’t be answered in a public forum, thus just adding another level of complexity to the transaction.
Conclusion… Forums suck for customer support!
Yep, it’s true, based on this analysis I have concluded that only around 5% of all the support threads created over two years were useful to other users, and therefore worthy of a spot in a forum.
Sadly, the fact that users only helped other users in 2% of threads over two years debunks the premise that your users will bend over backwards to help each other out so you don’t have to.
So, what is the best solution?
I reckon a combination of super friendly, direct email support and a well curated FAQ or knowledge base is the most efficient way to keep your customers happy and keep your workload in check.