The following post is meant to explain in detail the thought process and methodology used to build our Millennial Content Analyzer Tool. If you are looking for more detail in to what the tool does, how it does it, and why we think it features a relevant and fair scoring system, please read on.
Part 1: Defining What Content is Interesting to Millennials
In order to develop a tool that looks for content of interest to young adults on a church website, one obviously first needs some idea of what topics are of interest to millennials in the context of Christianity. Fortunately, a core building block of our millennial outreach initiative was to conduct a research project to determine the interests of millennials and young adults. This project, our Millennial Topic Survey Report, conducted statistically valid research and presented us with data on over 100 topics identified to be of interest to young adults.
Not All Created Equal
The Topical Survey Report not only helped determine what topics are of interest to young adults, but also produced an average interest rating for each topic investigated. This average interest rating is a numerical representation (out of 100) of how interesting each topic was found to be relative to its peers. As such, while two topics may have value to millennials and both should be considered when detected on a church website, certain topics may be of significantly higher value than others. Because of this, our Millennial Content Analyzer Tool uses each and every topic’s average interest rating from the Topic Survey Report to proportionally scale the impact that topic contributes to the overall score. Put more simply, when the tool detects content from a topic with a high average interest rating, more score is added than when it detects content from a topic with a lower average interest rating.
Part 2: Searching for Content
In order to actually locate the content on your church’s website, we execute what is known as a crawl. This process begins at your root home page, and once that page has been analyzed, then follows each internal (meaning, part of the same site) link detected to its destination. It then analyzes that next page, and follows any new links found. This process repeats itself until there are no unexplored links remaining, or we reach a cap of 300 pages (the majority of sites aren’t this big).
Keywords are the Treasure
Our Topical Survey Report showed us what topics to look for, but not how to find them. To locate these topics we developed a list of keywords associated with each topic, that when found, would suggest that said topic was being explored by the website. These keywords were carefully selected to be sure they would rarely be used out of context of the topic in question.
If you’re interested in what keywords we were specifically able to locate on your site, as well as the master list of keywords we searched for, you can click the “Download Keyword Data” button on the results screen of the tool. This will serve you a text file containing the information.
When analyzing each individual page being crawled, we look in three different places for keywords:
- Page Headings
- Page Meta Description
- Page Text
Additionally, much like how different topics have differing associated levels of interest, these three locations are themselves not equals. If a keyword is found in a page heading or the page meta description, it is quite likely that at least part of the page is focused on that topic. Conversely, if the keyword is only found in page text, it is entirely possible that the topic is only being mentioned in passing, and is not the focus of the content.
Because of the above, we weight occurrences of keywords located in page headings and meta descriptions to have twice the value of those found in page text. Further, if a keyword is detected in all three locations, we can be very confident that the topic in question is being dealt with in detail. In such instances, we apply an additional scoring bonus.
Part 3: How Much Content is Enough?
Perhaps our greatest challenge in developing the Millennial Content Analyzer Tool was determining how to present the results. How much millennial related content is good for a church website to have? How much is great? We didn’t know the answers to these questions.
We decided the most meaningful way we could express such an otherwise subjective result would be in relative terms. Thus, we elected to express the result of a crawl by comparing it to all of the other church websites that have been crawled. For this scoring model to be valid however, we first had to establish benchmark over a large sample size – which we did by running the tool on a large number of church websites all over the web. We found a huge amount of millennial content on some, and very little to none on many others. By the end of this process we had the data we needed.
Shoot for the Best
Looking at the bigger picture, it has been our feeling that many churches aren’t yet effectively able to reach young adults online – which is the driving force behind why our millennial outreach initiative exists. Because of this, we decided it was important to display the results of the tool in two forms:
- How your church website compares to those that are doing an outstanding job – the top 5% of results.
- How your church website compares to the average score derived from every single church website we’ve crawled.
We believe that doing this paints the most complete picture possible – and for churches that have an average or above average amount of content relative to their peers, might inspire them to shoot for the top five percent!