Experimenting with Programmatic SEO | Hacker News Reader

Experimenting with Programmatic SEO

allisonseboldt.com

17 pointsasebold9 comments

Great... a ton of automatically generated nonsense content that is "human readable" so you can sell stuff. Just what the internet needs, more low quality, bullshit content. And the author seems so proud of this.

I tend to agree, but there are circumstances where templated content is useful. Businesses with lots of regional locations or which sell thousands of similar products or components, for example.

In those circumstances, you can hire a cheap writer to churn out hundreds of crappy pages based on product data sheets or you can hire a developer to make a template system that does more or less the same thing.

Yes, this is a good point. It’s not like written articles always ensure higher quality.

This experiment isn’t about creating “Nonsense content”. I agree the internet definitely doesn’t need more misinformation or fluff. And that isn’t my goal with this experiment.

I don’t think content is automatically low quality with automated SEO. And if it is, it won’t work because users won’t find it helpful. The goal, like all good content, is to help people. But I understand if you’re skeptical.

> I don’t think content is automatically low quality with automated SEO. And if it is, it won’t work because users won’t find it helpful. The goal, like all good content, is to help people.

I don't think anyone's suggesting that. After all, most content produced by humans is low quality as well.

Generating good content generally requires someone to be a subject matter expert and to present the issue at hand in sufficient depth. Most content answers questions, but good content answers the right questions.

Looking at your site, I believe the main issue is that so much content is repeated between each individual plant. There's also just not that much content to begin with. A human could have copy-pasted these articles and swapped out the pictures within a few minutes.

Yeah I agree with that. I talked about it at the bottom of the article in “the sourcing problem” section. It’s something I’m aware of and hope to correct.

This is common in certain verticals, think e-commerce organizations with relatively niche products like replacement auto parts. They (the big fish being Autozone and Advance) go through the effort of having separate pages for each model year and the part name-- even when parts are interchangeable across multiple years. There are tools sold on a commission basis like http://www.youramigo.com that automatically create these pages, but they never seemed to get much traction.

The largest issue is Google's crawl budget. With limited domain authority, it will be hard to get xx,xxx+ pages indexed without a meaningful number of quality backlinks. Low MSV queries are even tougher if you don't rank #1, too, so using GSC to A/b test titles/descriptions CTRs will be super helpful. On-site interlinking will also be another easy way to juice visibility.

This is exactly what I see. Run and e-commerce site with 175k different product pages. Each have a ton of differentiation and keywords. Huge investments is page metadata and sitemap haven't resulted in Google ever indexing more than half the pages.

Do you have internal links to those pages directly from other pages linked from your homepage? Like a rotating popular products section? We added this recently and we went from having 5% indexed after 6 months to 90% 6 weeks later.

Reply on news.ycombinator.com