You want more traffic. You want it to be organic. You want it to be targeted. And you want to use a strategy that will scale to bring you more. Me too. This post details my recent success at that exact goal.
I read Patrick McKenzie’s blog post on Scalable Content Generation [SCG] an embarrassing number of times without ever understanding how to make it work for me. A couple months ago, what he was saying really hit home with me and I developed an SCG experiment.
There were two key insights from that post that propelled me forward. Only top ranking pages for a keyword get traffic and only long tail terms are easy to rank near the top for without a lot of work and calendar time waiting periods.
I had previously outsourced some keyword research (key component here is difficulty of ranking) and I identified a few potential keyword phrases that I could pair with other keywords to build a scalable content strategy around. The problem with most of them was that I didn’t have a library of content to push out 100s of pages. I also didn’t want to spend the cash to get 100s of custom articles written. It’s not really scalable content generation if I need to pay $50+ per page I can create.
Patrick uses bingo cards for this purpose and the reason struck me: he already has them, and more represent an asset. Yes, this really never occurred to me before. This meant I needed to look for keyword terms that match assets I already have and would like to grow. Now only a couple keyword terms fit the bill and I grabbed the most promising one, “cold calling scripts.”
I created a template with an image, call to action, 300+ words per page, and sharing buttons. I wrote a script to generate a list of unique industry names from some sample lists I found with some searches poste. I planned to create pages for every combination of “cold calling scripts <industry>” I could possibly target.
My template included a slot for some text that was unique to the industry I was targeting. I outsourced a paragraph of text on that subject to TextBroker at about $1 per paragraph. This was an affordable way to create the pages. Each template also had a space for scripting tips or a breakdown of a good or bad call script. I created a Jekyll plugin that would randomly assign one of a small library of the scripts I formatted for this task to each page. I plan to add to this library as I go. More on this later.
Using this methodology I generated an initial run of 83 pages each detailing a different industry. Jekyll generated the whole thing including categories/tags and index pages. Jekyll is quite suited to this task.
After I published the pages and linked to them from the footer of my site, I sat back and waited for Google to re-index. I think it took about 10 days before I started seeing organic search results that sent people to these pages. The keyword research paid off. The vast majority of my new pages rank of the first SERP and most rank in the top 3. After another 5 days this experiment had paid for itself in book sales.
Orange is December and Blue is January.
My organic traffic actually increased by 169.84%. My time on page, bounce rate, and pages per visit are essentially static, but I have seen an increase in return visitors. This is all the result of some keyword research, content generation, and content repurposing stitched together with various scripts and Jekyll.
My conversion rates on those pages are worse than the rest of the site. The site average is 14% and those pages run closer to 11%. I have more exits from those landing pages than others, and some of them have a dismal bounce rate. This isn’t really unexpected. Those pages generally have more visitors that have never heard of my book or me and the traffic is less targeted than referrals I might get from my blog or guest posts.
That said, the TextBroker content is lower quality than I would like it to be. Some of that is my fault for failing to give clear instructions, being impatient with results, and being cheap. I also have a fairly small set of script breakdowns and tips that are randomly placed on the pages. There are some significant page quality issues I need to tackle.
The nice thing about tackling those issues now is that I can measure which pages get traffic and how they convert. I can prioritize which pages I improve and even A/B test improvements to the overall template. None of this would be possible without validating the idea or without the traffic the experiment brought. If I had started with a really small set of pages (instead of 83) I might have picked ones that were harder to rank for or didn’t see substantial traffic in a month. It wouldn’t be a very good experiment.
I’m currently improving the template, working with better writers, editing some existing content, and adding more script breakdowns to improve the pages. I think these types of improvements are a worthy investment in light of Panda and Penguin. Patrick’s Bingo Card pages have no more unique content, but continue to rank well. Those updates are focused on down-ranking pages people don’t find useful and those pages that are intentionally duplicative. I want these pages to be valuable to people. I’m not interested in a demand media business.
I added 179 more pages on 26 January, but there is nothing to report yet. While I work to improve quality, I plan to experiment with some new ideas and keywords.