One Month In
July marked the first FULL month of running IntroCave and I feel like I was able to make a few improvements last month. I'm still working on two major projects that I want to wrap before starting to bring in new content.
03:40
"What gets measured gets managed." One of the first things I added to the site was a couple of simple timestamps. When you submit your customizations and I fire it off to our render server, I start a stopwatch. When that video preview comes back, I start the stopwatch. Across all previews for the month of July, our average render time was three minutes and forty seconds. I could get more granular in the future and break that down to true render time and queue time, but for now that's the number to beat. This leads into the first major ongoing project: as I rebuild our render servers, I'll have a benchmark to know if I'm actually getting better or not! I'm zeroing in on some pretty major architectural changes that require re-processing all of our current intro templates, and (unfortunately) it doesn't make much sense to add more content until this is complete.
Making it easier to browse
I don't have a good quote for this one, but dead ends suck. If you click on a video intro that you think might be perfect for your youtube channel or stream, what happens if it's just not quite right? Until a week or so ago, you probably hit the back button, clicked on the IntroCave logo to go back to the home page, or just bounced entirely. Similarly, we have a few templates with slight variations (most commonly one with a logo versus one with a text title). Unless they happened to be listed right next to each other in search results, how would you ever know that there were variations available? I think the solution to both of these problems is the new "Similar Intro Videos" feature underneath each intro.
Similar videos are computed by comparing keywords they share in common, which means variations should share all but one or two keywords and rank highly against each other. Even if there are no variants, maybe there's another "dark glitch logo" intro video that shares a bunch of keywords with the current video and might be perfect! The experiment seems to have borne fruit, as page views are up since adding similar videos. Not all of our intro videos have good keyword coverage and descriptions right now, so I've also added some simpler keyword lists just to add a little more content. This brings me to the second major content task: I'm about 30% of the way through rewriting every title, description, and keyword, and screenshot for every intro template on the site.
Writing this content should be fairly quick, but because I also need to go through every template for the render server changes, I end up spending a lot longer on each video than if I were just rewriting copy.