Build from scratch vs. CMS, what would be the main onpage wins?

[BCN]

Hi,

I've normally built my affiliate sites in WP.

Now I'm launching a new batch, and these are a bit more work and have more data and API integrations, so I've opted for Flask + Postgres (SQL database). Themes are regular HTML themes based on bootstrap 3 and 4 ...

This gives me a lot more opportunities when it comes to setting it up properly from scratch, without the constraints of the typical post <-> tags and post <-> categories constraint in WP.

When making a site from scratch, let's say you could make it any structure or with focus on any onpage factor. Where would be the biggest wins be? I've added options for more split testing, and also structured data.

 

[Ryuzaki]

without the constraints of the typical post <-> tags and post <-> categories constraint in WP.

I'm not sure how'd I'd answer your main question yet, or if I think there is an answer (Wordpress handles all on-page stuff just fine in my experience).

But rather than completely reinventing the wheel (I've been tempted to make my own Laravel based CMS) you know you can query Wordpress based on custom fields or basically anything else? You could do this with normal tags too, if you don't allow that taxonomy to be indexed or visible anywhere. You could put whatever you want in there and create any loops you want.

 

[CCarter]

Where would be the biggest wins be?

It REALLY depends on your exit plan. If you plan on selling the site in 2-7 years it make sense to go with Wordpress or another known CMS cause the new website owner will be able to get up and running a lot faster. This isn't to say that a custom setup will not sell - they will, but a wordpress setup will be a lot more appealing since so many people can jump in the driver seat.

So consider your end goal when thinking about a CMS setup.

As well some of these "re-invent the wheel" scenarios, actually most of them, are forms of procrastination, learn to identify whether you are pretending to get to the money or actually on the road to the money. This may not necessarily mean you, but could fit for someone else as well that's reading and has that itch that when scratched doesn't add any dollars to the bank account.

 

[BCN]

Thanks for the input, I started this in WP, but it seems a bit out of the scope for WP at least for my skills. I think it will be quicker to do it custom, but I could be wrong.

These will be built on aged domains and sold when they rank and bring some income most likely.

The flow is basically. Pull data from multiple feeds (atom/XML), we are whitelisted to grab and publish data from some government feeds, scraped data from competitors, generate pages, geocode via google api, mix in content from writers along with tables and graphs on data. All this is done in about a day, now I just need to render the frontend, so I just grabbed some html themes from wrap bootstrap.

I guess it could post all this data to WP somehow, not sure.

As well some of these "re-invent the wheel" scenarios, actually most of them, are forms of procrastination, learn to identify whether you are pretending to get to the money or actually on the road to the money. This may not necessarily mean you, but could fit for someone else as well that's reading and has that itch that when scratched doesn't add any dollars to the bank account.

This is definitely a good point.

 

[turbin3]

You might look into WP All Import. It's easy to generate complex pages with it. It uses CSV, XML, or JSON files that you import.

I don't think it has capability for making external API calls. Depending on your use case, that may or may not be a big issue. If it's any indicator, I've used it before to generate thousands of pages based on geographic areas, and building complex page types that had tables, graphs, and incorporated different types of media (images, videos, and podcasts IIRC). I was pleasantly surprised with how well it worked.

Based on the use case you've mentioned, sounds like you might just need a few additional steps like:

• Create a few PHP scripts for API calls.
• Create post-processing scripts if needed.
• Set up cron jobs for recurring feeds.
• Output the data as CSV, XML, or JSON.

If you can't build those scripts, most of those should be super easy (and cheap) for any halfway decent developer. Usually stuff like that is just doing these sorts of things:

• Requesting a URL & downloading a file
• Looping over contents & manipulating
• Outputting them to a desired format

Now you have your local source. Setup your import template configs in WPAI, and you can also set those on a schedule if you want. You can bulk generate pages. You can interface with ACF, Yoast, and all sorts of custom fields.

The point I'm getting at is this. If you can get most of your data types into one of those 3 formats, combined into a single file, WPAI can do most of the page generation for you.

Now, that may not be a "perfect" solution. Think about the alternative though. Building a CMS, from the ground up. Gotta account for security, performance, scalability, robustness... It's at this point you become devops. Devops in the marketing realm tends to be a negative-ROI use of time and money, unless maybe it's for an actual high value product.

 

[BCN]

Thanks, I will look into that.

One thing I find "confusing" in WP, is that I'm not very good with PHP. Python I can do pretty well, and I'm used to working with Python + Postgres/SQL. I imagine I can split the data processing and harvesting (keep it in python), and use WP All Import, and get away with minimal coding in WP itself.

 

[backwoods]

As well some of these "re-invent the wheel" scenarios, actually most of them, are forms of procrastination...

/thread

I can think of a few people here that worry WAY too much about what CMS they use, page speed, plugins, design, etc., and fail to look at the big picture and move forward and make any $$.

 

[BCN]

We launch new sites monthly, so it's not about that.

The idea with the thread was more to see if there was any wins that can be done for onpage if you're not restricted by WP or any other CMS for that matter. I remember when we used to do masse pages, you ended up having tags/categories/pages/subpages to get deep silos, and then I think it's a fair question to ask if bending WP into what you need, or building something custom is better use of time.

WP works great for us for the typical content sites, but i.e. earlier we've done crypto/forex sites with a guided "Find your broker", loan comparison, cloaking etc. where it's not always the best option to do WP.

We also have some amazon sites that are made in WP, using the Amazon API to pull the data. It's a custom setup, and it works OK in WP, but it's really just some content with extra meta data (i.e. Manufacturer, attributes and custom fields on the post level).

This is not a typical build, but relies a lot on data gathered from APIs, cleaned and used to do calculations.

 

[BCN]

OK finished this now. It got a bit of a different twist in the end, to be able to run many sites from the same server, all served from a webapp frontend (simple bootstrap sites), and an admin panel.

I've used this to test a lot of marginal sites, mass pages etc. Works pretty well. So just putting it on different servers, assigning all domains to it via Cloudflare, adding domain to DB and adding content and kws from scrapes. It also has some fun features for doorway pages.

For the other sites, I've decided to use Python for the data flow, getting xml and json data in, scraped data, etc, then clean it up and put it into WP. Like suggested above.

The reason for this is really for resale, and because with WP we can anyways use the API with a JS framework and still get a 100% custom site should we need it. And WP is tested and works well for this type of content sites.