I have already mentioned the accidental nuking of the site by tech support, the rebuild decision tied to mental health issues, and my decision to separate everything into two sites plus add Flickr for my photos. But one big question remained…if I was going to build back better, what else would I change in the process?
If I was honest with myself, I had to face the reality that perhaps, shockingly, I didn’t design the site incrementally over time the way I would have liked to design it from day one.
The problem-ridden version I had was a dumpster fire, that was clear. But could this be an opportunity to think outside the box? Could I, in fact, set the box on fire and start with a fresh page?
Yes, yes I could. So where should I start? My big four website choices were relatively straight-forward:
- Front-facing design functions;
- Behind-the-scenes functions; and,
- Behind-the-scenes workflow options.
Theme / layout
For layout and colour, there are a lot of themes out there for websites, often coming down to a series of elements. Some sites go with a top menu and then a big splashy header, often combining that header with a slider to advertise multiple contents or the latest posts. Netflix, Disney+, Amazon Prime all do this with their movie listings for example — slider areas that show videos or pictures that advertise content elsewhere in their offerings. On news sites, there are often similar layouts followed by spotlight boxes, three or four across, that highlight the latest articles or popular posts to click on. High energy, lots of images or videos, driving people to click, click, click!
They may be popular, but that’s not me. I don’t want a big giant flashy site, partly as I live and die by the word, not the graphics that go with it. Heck, other than my featured image, many of my posts don’t even HAVE graphics (gasp! the latest design bloggers would be aghast with my approach!). That’s not accidental or laziness, it’s the fact that I only include what advances the story I’m trying to tell or imparting info I think is relevant. I don’t use “optional flash” to add sizzle or pizzazz. If the words aren’t working, I’m already lost. Nevertheless, even ignoring those big layout issues, I’m still left with three types of theme options.
First and foremost, the latest approach is to sell people on using “page designers”. Secondly, you can choose an theme that has a fair amount of customization options. Or third, you can buy a theme that is EXACTLY what you want, no customization required.
If you think of it like buying Microsoft products, it’s kind of like buying MS Publisher (a completely open page layout program), MS Word (as a good editor with some layout options but generally looks like a page), or a custom background for your document that can’t be edited.
I don’t want or need a big Desktop Publisher / page designer program to do my blog as there’s a lot of extra overhead that comes with it. I confess that I think the common advice that newbies hear i.e., to start with page designers, is the worst advice possible. People get into it having no idea what a post or page even is in WordPress, have no real design background, and suddenly things start going wrong, with no idea why or how to fix them. Eventually, they get frustrated and give up, when if they had started with a simple blog and theme, they could wait and complicate their lives later when they have a working blog to keep them engaged in the meantime. People often buy MS Office for business and have options to use MS Word or Publisher. Which do they use? MS Word. Publisher has its use, and most people don’t want or need the overhead to get something looking the way they want.
I played with a couple of page designer plugins/themes, but honestly, my view is that WordPress is opening up design more and more, and eventually page designers are going to be redundant since WordPress’ block options will be able to do just about all of it anyway. I’ll stick with the core.
I have looked at the third option — prebuilt static themes — and most of them are not quite what I want nor do they come with support or customization. I’m “vanilla” for my design, sure, but I at least want to add some sprinkles or syrup occasionally.
Which leaves me with the MS Word-level of theme (Weaver XTreme Plus) that I have been running for over 10 years in one form or another (premium, free, plus, etc.). As it has so many customization options, I really don’t need anything more powerful. Sometimes I need LESS power. But I like it because it allows me a format that works for me and enough tweaking under the hood to add the sprinkles or syrup occasionally.
But I did my due diligence anyway, played with a couple of new popular themes, tested the layouts, inserted some flashy graphics here and there, tried a header, and in the end, I kept coming back to what I already had for the tool.
Yet, as I mentioned earlier, I DID want to play with branding. So on my PolyWogg site, I went with my main header (a photo where I grew up, with a nice logo overlaid on it) and then added custom headers for my HR Guide, Reviews, Recipes, conference, and Astronomy. I have one waiting for trivia at some point, but I’m not ready for that yet. I’ve also gone with some “splash” pages on PolyWogg with intro pages for each area rather than simple blogs and menus, and ditched the sidebars.
Meanwhile, over on ThePolyBlog, I went with the lilypad header that I like, based on a photo from the Bruce Pit, but no custom headers for the rest of the site, and generally a pure blog-style layout, albeit with reduced content hiding in the sidebars. I have kept them — for now, at least.
Front-facing design options
For the public side of the blog i.e., what the average viewer sees, I have 21 plugins that help with design, layout or functionality. Most are relatively straightforward — adding social media buttons, controlling spam, adding a bunch of layout blocks that aren’t in the core WordPress options, adding navigation menus to go from post to post, providing links to other related posts, or simply displaying the posts in a specific format.
For the post-dumpster-fire build, I tweaked a number of things.
I already mentioned removing sidebars and changing headers, but on PolyWogg, I also have different headers for different areas of the website and with those separate areas, I also installed a plugin that lets me display different menus in each section. So if you’re in the reviews section, you see a narrower reviews-related menu; if you’re in recipes, you see a recipe-related menu. All of that is done through a special plugin (Conditional Menus).
Another part of PolyWogg has my TV Premiere posts, i.e., reviews of individual premieres, and for most of the posts in “reviews”, I created a separate index of them in sortable tables (with TablePress). But the TV Premiere list is different…it’s literally just a list of posts by date. The simplest way to generate that list was a plugin called simply Display Posts along with a related addon called Date View. Between the two, I generate a list of all the posts in that category and place it on a page. It’s a pretty powerful tool, and I may decide in future to expand it to other areas on both PolyWogg and ThePolyBlog. It makes a really simple navigation tool without a lot of overhead.
A friend constantly complains that my website shows posts in feeds as the full post, and she hates it. She feels like they should always be in “excerpt” mode so you can scan more quickly. While I understand her view, it’s a design choice that has two competing views in the design world. One side agrees with her that you should show the excerpt and let people click to see more; on the other side, they argue people want to be able to scan quickly without having to click. The same happens with newsletters or auto-feeds — if you send the entire post by email, then people don’t have a reason to click, and so they don’t go to your site and they don’t generate ad revenue. But I don’t run ads on my sites, not now, not ever. And I have always preferred the long form in feeds and on the site. One fewer click. But in the end, I decided to follow her advice, albeit for a slightly different reason. I use a newsletter (to be mentioned in the next section), and it doesn’t like certain types of blocks. If I use them, then the feed looks terrible. It looks great on screen, but not in email. Soooo, if I only go with excerpts, problem solved. I don’t use those types of blocks in the first 100 words anyway (another plugin to help where it cuts off).
Earlier I mentioned that I have social media icons already, and I have multiple ways to do those. However, while they also include options for printing or PDF, I added two PDF-related plugins as well (PDF & Print, and PDF Embedder) as they work well with my HR guide so people can access it more easily on the PolyWogg site. I have more docs coming, so I added the functionality now.
Before I come to the big front-end change, I’ll mention quickly that I also changed my Forms plugin (to go to WPForms) and I reinstalled my trivia plugin called Quiz Maker.
All of the above changes are, for the most part, relatively cosmetic changes here and there. Tweaks, I would say. The next one is huge.
As I noted earlier, one of my biggest areas and the most work on the original site was my Photo Gallery. I paid for a premium version of NextGen Gallery and I have been using it for sometime. I doubled down on the plugin, I found ways to do things that others hadn’t, and I’ve even thrown myself into the support forum, acting as an informal extra resource for anyone and everyone. Due to COVID, the plugin developer has had trouble providing ongoing support to users of the free plugin, and where I can, I’ve stepped in to offer some suggestions. Sometimes, a week later, the support person will chime in and say, “Yeah, what he said”. Often with more aplomb and gravitas, and with additional suggestions, but a number of people have been able to try what I suggested and fix their problem. Not necessarily with “my” solution, but in testing my solution, it better defined the actual problem. I often joke that I may not be able to fix your problem, but sometimes I can break something else that I can fix.
NGG is the most used plugin in WordPress for photo galleries, it has power out the wazoo and I had almost 10,000 photos uploaded using it as the basis for my galleries prior to the crash. How many have I uploaded since? Zero.
I killed the plugin.
Remember I said I was doing my due diligence to build back better? Well, one of those changes was to go with Flickr as my online repository for my photos and video rather than embedded directly in my website. As I noted in a previous post, I went for “ease of use” over perfection. And NGG was the solution for direct embedding. With all of my photos in Flickr, what I need NOW is a good tool that lets me embed the photos FROM Flickr into the site. Now, WordPress already has a default embed option for individual items, and it works just fine. But if I want to better embed galleries or albums or subsets, then a tool like Photonic Gallery directly connects to my Flickr account and lets me embed. Another tool called Meks Simple Flickr Widget also lets me embed easily too, with a bit more control than the standard WordPress embed. But make no mistake. This is a HUGE change on the front-end and the back-end. It is also the change that reduced my rebuilding workload from a year to probably six months of normal revision.
It’s a total gamechanger for the front-end, the back-end, and everything in between.
On the back-end, I have 18 plugins that I use to manage the site’s functions. The obvious ones that many people use are there — an advanced editor, a broken link checker, tools that maintain wordcounts, security certificate manager, a tool for managing redirections within the site, and a status checker to make sure the Cron tool is running regularly.
Not too long before I lost my site, I had upgraded two back-end features, including a tool that auto-posts to the Buffer website which in turn autoposts to Twitter and Facebook for me, as well as a newsletter plugin called MailPoet that lets me send out posts to people who are subscribed to the site. When I created the initial newsletter, a bunch of people subscribed, but when I was rebuilding the site, I had to “force stop” their subscriptions and invite them to rejoin afterwards (otherwise they would have received 1500 emails saying “Hey, this post is back up”).
But beyond those basics, I made five fairly large changes behind the scenes. Each of them are significant on their own, to be honest, and most of them I had not done previously as they didn’t work well with NextGen galleries and some other tools I had running. With the rebuild, it was time to “fix” the site and do it right.
Up first was simple Google Analytics. Every site owner should get the reports of what Google thinks of their site, but I am not a slave to metrics since that isn’t the kind of site I have (I don’t really care how I rank or the load times as long as they aren’t measured in decades, people come to my site for specific content they can’t get anywhere else, I’m not a random stop on the information superhighway). Yet, even though I’m not a slave, the analytics can help tweak the site for performance and user experience, so I added a good setup to monitor it.
The obvious related area was Search Engine Optimization, and well, as I said, people tend to come to my site because they’re looking for me. They aren’t randomly searching for info on widgets that I happen to sell, I’m not a commerce site. People are either looking for specific posts I’ve written that they have been referred to by others, OR they are looking for me in particular. It’s an unique niche so the “one size fits all” advice that people say “every site should do” doesn’t really apply to my site in the same way. It’s an issue, sure, but it’s not a DEFINING issue. But I can do more to be SEO friendly, if not exactly SEO optimized. I asked in a few online fora about good options, explained that I was really looking for a plugin that does the basics without getting into the weeds, and someone referred me to SEO Press. Everything I need, and easy to run? I’m in.
Security is a necessary staple of any site, and I had a small quandary. I have tried various plugins over the years for firewalls and anti-malware, and I like several of them. But two of them didn’t work well on my servers or with NextGen Gallery. First up was iThemes. I like it, I liked the interface, it seems intuitive to me, and more than a million other users agreed with me. But the way it interacts with the main config file for WordPress for security didn’t seem to work consistently on my previous host (I was overloading their server and they couldn’t tell me why, so I eventually left that company’s hosting plan). When I started with my new host, it did it again after about six weeks, and the guy was able to narrow it down to that plugin and how how/why it was doing it. Simple fix, but not exactly warming to the heart.
I then tried WordFence, which is the big guy on the block, but it didn’t seem to be very intuitive to me at first, and then I had a conflict with an old plugin I was running. I ended up going with All-in-One security, which was working relatively well on the old setup. Except with respect to Next Gen Gallery. There is a partial conflict with their firewall settings and NGG’s update process. Each time NGG would need an update, the plugin would stop working afterwards. But if I deactivated the firewall, ran NGG, and then reactivated AIO, everything would work fine. It just didn’t seem to “complete” the update for NGG when AIO was running. AIO and NGG both tried to narrow it down, and while neither could identify WHAT was causing it, we did manage to agree that temporary deactivation fixed it. But I’m not running NGG any longer, nor the older plugins. So I joined the 4M+ people running WordFence again. I feel it is the most reliable, and it makes more sense to me out of the box with a lot less tweaking than I had to do for AIO.
Another area for the sites is how they send emails from the site. There are default ways to send stuff, and most of the default approaches get flagged as spam or might never send in the first place. Instead, I went with a new upgraded plugin called WP Mail SMTP and linked it to better email sending. I’m not 100% certain it’s a significant update, but it is a more professional way to do things.
Last but not least, I added an optimizing plugin. I know, I know, if you have read my previous blogs, you’ll know that a server-level optimizer was part of what killed my site in the first place. However, this one allows some granularity of control AND it provides a simple caching solution. It does a couple other things too (minify, image compression) that I don’t use, but WP Optimize is compatible with my setup and doesn’t cause gremlins to proliferate. I’ll take it.
Together, those are five pretty significant upgrades to the back-end. I had tools doing some of it before but nowhere near as extensively or even, dare I say it, “properly”.
Behind-the-scenes workflow options
When I had two small websites (for family and friends) plus one big PolyWogg site, it was relatively easy to maintain. I would never confuse any of the three with each other for the back-end as I generally don’t have to do much on the back-end for the other two. But now that I have two BIG sites to manage, and I’m administering both almost daily, I need to make sure in the back-end that I’m working on the right site!
That may seem obvious to the non-user, but it isn’t. The back-end has the same colour, layout, look and feel. Except for the small name in the top left-hand corner, you wouldn’t know you were on ThePolyBlog rather than PolyWogg. And in the first two weeks of the rebuild process, I made two mistakes where I did something confusing the two sites.
The first was simple. Before I nuked the whole account and started over, I moved ALL of my PolyWogg content to ThePolyBlog. ALL of it. Then, when PolyWogg was running again, I moved a copy of it all back. So my content was actually in two places at once, even though ultimately it would be in only one. For example, I put all my reviews on PolyWogg, went through everything, got it up and running, and deleted all the reviews from ThePolyBlog. Good to go. Except I went into PolyWogg the next day, and couldn’t find some of it. WTF? What did I do? Oops, I accidentally changed the category it was in, thinking I was on ThePolyBlog instead of on PolyWogg. I almost deleted all of them, but fortunately I just misfiled it by confusing which site I was on.
The second was stupid but potentially devastating. I run a plugin called TablePress that lets me have sorted tables on various pages. You can even filter them as the user on the front-end. But it is really only used for lists of Reviews, and since that is only for the PolyWogg site, I don’t need the plugin for ThePolyBlog site. So I went into ThePolyBlog, deactivated it, and deleted the plugin plus the data. Next day, I’m on PolyWogg and one of the tables isn’t showing right. Hmm, that’s odd. Go into the admin area, TablePress isn’t even an option. WTF? Ohhhhh. I accidentally deleted it from PolyWogg, instead of ThePolyblog. Okay THAT needs a fix. I could recover the data, no problem, but I needed to solve the confusion problem.
Fortunately, there are two easy ways to do this. Built into WordPress is an option to change the basic colours of the admin layout. Which sounds perfect except I didn’t like any of the other three simple options. Instead, I found a plugin called Admin Color Schemes which adds about 10 more. Voila!
Now my ThePolyBlog admin area is a relatively bland beige because it is the secondary site. If I lose a post, or forget something, no great loss. It’s calm, it’s soothing, it’s daily wear and tear.
The PolyWogg site? I have a very bright blue interface! It screams HEY PAUL PAY ATTENTION THIS IS YOUR POLYWOGG SITE WITH ALL YOUR BIG CONTENT LIKE YOUR HR GUIDE. Nothing calm or soothing about it. It’s not raging red, but it stands out. As soon as that page loads, I know I have to pay attention to what I’m doing.
Problem solved. I think.
Oh sure, I have other tools too, such as an admin menu editor that I used to make both sites look rather similar in the backend for administration (Dashboard / Posts / Pages / Comments / Media / Reusable Blocks/ Menus / Widgets at the top as my most active areas). I also use a nested pages tool to help me drag and drop pages, which is incredibly useful for the PolyWogg site for my HR and astro guides, even my review indices. On both sites, I used WordPress Importer to bring the data back in, and particularly on PolyWogg, I used Tako Movable Comments and Bulk Move to parse comments into 8 different pages rather than all on the one “main” page for my HR guide. For ThePolyBlog, I frequently blog about articles, so I use the Press This plugin, but I don’t need it for PolyWogg. On both sites, I use Yoast Duplicate Post to clone or copy existing posts into a new post, often to duplicate a post layout easily.
Hmm…something’s missing, right?
What could it be?
It would be an obvious conclusion, i.e., better backups. That’s how I got into this problem in the first place, right? The assumption is generally if you have good backups, everything’s right in the world. Except that needs to be unpacked a bit.
Any backup solution actually splits into three elements:
- Creating the backup
- Storing the backup
- Restoring from the backup
If you haven’t read my previous posts, first, what’s wrong with you? How could you read this LONG post without having whetted your appetite with the previous laments? 🙂
Second, you’d know that I actually HAD backups. It was the third part that was the problem.
In fact, I was already using 3 backup options, one that was completely up to date, and two that weren’t. I’ve already covered multiple types of backup tools for a WordPress site, so let’s talk about the three I had in rotation at the time of the nuclear blast.
First and foremost, I had a complete download of EVERYTHING. It’s the simplest form of backup, but is entirely manual to do.
However, it was about four months old, and I had made a lot of changes since the last download, but it was relatively there for structure. I could have tried to restore from it AND then just add in all my content. The problem was that the latest version of my content was a bit suspect. Not completely, but when you merge different sources, you multiply your chances of an ongoing gremlin problem. Which I already had, and was trying to erase. This failure though is on me. When you’re in the process of making updates, and you have another 20 to go, it’s hard to remind yourself to “stop” and make a big manual download when you’re busy. This isn’t news, it’s why systems designs include automated backups. So it doesn’t rely on the dumb user to remember to do it. I could have gone with the existing option, rebuilt the data manually, etc., but with a small corruption problem in the data, it wasn’t clear this would be reliable. Equally, having a complete backup wasn’t the problem. I had that solved already with another tool.
Second, I had a proper backup done automatically by my server about six weeks before. I was running an WordPress-based backup tool, and I disabled it while I was doing some reconfiguration (you don’t want it trying to run while you’re in close to maintenance mode), so again, I had a complete but slightly out of date version. The problem with this one was that it too wouldn’t restore properly.
Third, the core “problem” was again, not the backup. The server runs JetBackup every single day. It had a complete backup of every day going all the way back to November 2019. It COULD restore everything. In theory. And it had a great version from about 2 hours before the nuke.
So again, going back to the top of the section — I had 3 backup options that were creating the backup, check. I had the info stored in three separate places — the host’s server area (separate from mine), within my own area (the WordPress plugin-based backup), AND I had a complete download. Check.
The first two elements wee tickety-boo (creation, storage).
The third? A restore function? Welllll, therein lies the rub. The host was running LiteSpeed Cache at the server level and it was interfering with the restore function. That couldn’t be disabled, so that’s a problem. But I had an “out”. I could download my backups, up to date as of 8:00 a.m. the morning of the nuke, and take my whole site somewhere else. Photos, WP, data, everything. All good.
Except we had done a full restore. Which completed without errors. And yet there were gremlins in the new version. If I went to “new host provider X” and installed those versions, chances are that some of the gremlins would remain. AND I had already paid for the website for the next three years. Which meant I would be paying again somewhere else with no guarantees of a successful restore, I would have to “pay” to try it out, and I would likely have to repair everything manually as I suspect part of the gremlin lay in the creation initially.
I needed a fourth backup option
I mentioned above that generally you need an option that generates and stores the backup and permits restoration. All good. But all three of my options above were generally tied to the server that my site was running on. If it malfunctions, all three of my results were generally at risk. It’s not exactly technically accurate, but I tend to think of it more as an unmitigated risk. I wanted a fourth option, and by happenstance, it links to another tool I had in mind.
I already said that I was going from 1 active site + 2 inactive sites, to 2 active sites and 2 inactive sites. Or put more simply, doubling my workload (1 active to 2 active). There’s an app for that though.
Well, technically, there are plugins or websites for that. What they all basically do is tie all your websites into a single interface. They have different ways to do that, some require a separate website to be created, etc., but you end up with one ring to rule them all, errr, one site to administer them all. Sorry, my precious, I got distracted.
In the end, I went with ManageWP. What I essentially can do, for free, is link all four websites into the ManageWP site, and it will let me go to that site in order to THEN go to any of the four sites for individual management. Or if I want to upgrade a common plugin on all of them, I can do it from the main site. It’s a bit like running MultiSite, without the sharing of plugins and themes. Each site runs its own install, I just get to share admin in the background.
Know what else I can do? Add a regular backup that stores that backup on a TOTALLY separate server and doesn’t require me to do downloads manually. It costs me about a $1 a month per site to do it, but for about $50 a year, I have a fourth backup. One that, in theory, I should be able to re-install on any server anywhere with no issues in the future, and isn’t quite as dependent on the original server. More of a “pull copy” than a “push copy”.
I can’t rely on it as my SOLE tool, but it is a nice addition to the ones I already had. So I’m ditching the server-based one, since it had some challenges running properly anyway with a cron issue, I still have the daily backups, AND I will set a schedule for regular full downloads, something that will be simpler now that I don’t have 10K photos installed with it too.
Wait, didn’t you say it would take you six months?
I did say that. I estimated a year if I was rebuilding EVERYTHING the way it had been, including the photo galleries. Working at a normal website management pace. When I ditched the gallery tool to upload into Flickr, I reduced that estimate to about six months. It seemed ambitious, even though a few people thought I was being dramatic. Nope, I knew it was going to be a lot of work.
Yet I managed to do it all in six weeks. How? By killing myself.
Starting on March 6th, when I fully committed to the rebuild, I worked about 4h a night for Monday to Friday every night and week. EVERY SINGLE NIGHT. Sometimes longer. It was not uncommon for me to start working around 7:30 and at 2:30 to be shutting down so I could sleep. Sometimes I would work at lunch too. Or from 5-6 in the afternoon before supper. Some nights I wouldn’t start until 9:30 or 10, or I would watch a bit of TV while doing the grunt work.
And make no mistake, it was straight grunt work. No fun, no designing, no creativity, it was either reconfiguring plugins until my eyes crossed, or opening, editing, and resaving posts or pages until everything started to blur together. At a minimum, it was 4h a day through the week for six weeks.
On weekends? I generally doubled- and tripled-down. For each Saturday and Sunday in those six weeks, I generally worked a minimum of 10h a day on the rebuild. I’d take a break, hang out with Jacob, do some more, eat, watch some TV, and then be back at it for the night shift. Between the regular week 4h/day x 5d, and the weekend of 10h/d x 2d, it was about 40h a week I was spending on the website. On top of working my normal full week for work.
I’ve been doing almost nothing else for the last six weeks in my spare time. My estimate is that I spent somewhere between 240 and 260h to build it all back the way I want it. If I had put a more normal 10h per week into the rebuild on top of my normal workweek, it would have been the full six months.
But I just wanted to do my due diligence, check a few things as I went, but then to just blast through. It took about 2.5 to 3w of that time to get through most of the structural changes, testing new plugins too, and to get PolyWogg relatively where it needed to be. I then spent 3w getting all the content back up to where it needed to be, and another half week or so putting the final touches on some of the elements. It was a ton of work, and almost none of it was fun. It was just a grueling slog.
How now, brown cow?
About the 5.5w mark, I had to take a break one day and just BLOG about something. I needed to feel a sense of progress. That I could actually create NEW content again. So I wrote a few posts and it gave me enough of a boost to get over the final hump.
There are a few niggling things I want to tweak, or fix. There are about 100 posts where I am missing a photo that will need to be repasted from Flickr instead of from an upload to the gallery built into the website, but generally speaking, I’m done. The sites are running almost at full capacity.
And yet, I’m essentially starting from scratch. I sent out my invites to all the previous subscribers, and only 2 have resubscribed. I even sent all the invites MANUALLY so it wouldn’t go to their spam filter.
I’m also posting a bit each day on this journey, summarizing what I went through, part catharsis, part transparency. And because I have those great trackers, I know that almost nobody is reading any of it. My traffic is way way WAY down from low to mid triple digits all the way to single digits. I get almost no comments, not here or on FB. I’m typing into the abyss.
I’ve been here before. Hopefully it will grow again as I write about more interesting topics that don’t sound like I’m writing computer manuals in four parts. Or, in the end, if nobody reads it, then my site is a glorified digital diary.
But it’s back, and I hope, better than it was before. Time will tell, I suppose.