A Closer Look At EveryZing
For those of you in need of a turn-key, hosted solution to gain natural search engine visibility for your video and audio content, the company EveryZing says they’re the ticket. EveryZing’s solution consists of a suite of products: ezSEARCH, ezSEO, and RAMP. The products are designed to provide users with a powerful multimedia search tool, […]
For those of you in need of a turn-key, hosted solution to gain natural search engine visibility for your video and audio content, the company EveryZing says they’re the ticket.
EveryZing’s solution consists of a suite of products: ezSEARCH, ezSEO, and RAMP. The products are designed to provide users with a powerful multimedia search tool, publishers with the metrics and editorial controls, and search engines with crawlable text-based fodder they reward with higher rankings.
At the core of EveryZing’s product is a speech-to-text technology developed by BBN Technologies with $50 million in government funding. BBN spun off this Internet-focused product into a company called PodZinger, which was recently renamed to EveryZing.
BBN is a pretty whiz-bang group of engineers who have been tackling complicated natural language processing issues since 1969, primarily for government applications. They are currently working on machine translation technologies across a broad set of languages.
With this base technology, EveryZing can transcribe text from speech across hundreds of documents per second, translating each one into searchable text. Once transcribed, millions of video and audio files can be quickly searched and specific spoken content can be found in seconds.
However, what impresses me most about this solution is how smart EveryZing is at leveraging the translated content into a search engine friendly format.
For example, multimedia.boston.com, a subdomain of the website for the Boston Globe, is powered by EveryZing’s technology and contains an extensive collection of local news videos and full-length radio programs, the latter provided by WEEI, a local AM sports radio station.
Let’s take a look at how effective EveryZing was at gaining search engine viability for their client’s video and audio assets. I will evaluate them on the three pillars of natural search optimization: indexing, content optimization, and links.
To provide continuity to the audit, I will focus on the same search topic throughout the review. Now that spring is here and baseball has returned to our daily lives, I have selected the popular bruising slugger from the Boston Red Sox, Manny Ramirez.
Until your content is found and indexed by the search engine crawlers, not much else matters to natural search. So, it’s crucial that your video and audio assets are well organized on your site and “live” at a unique and persistent URL address.
EveryZing tackles this challenge by creating static landing pages that house a dynamic collection of audio and video clips that share content about each major topic. For example, here’s Manny Ramirez’s page. This URL is free of parameters that are sometimes tagged onto the back of a dynamic URL. Search engines have traditionally struggled with parameters, especially in situations where the URL has three or more.
The EveryZing search engine can also quickly scan all multimedia assets for nearly any term. For instance, a query for “SEO” brings up 12 relevant audio files on the topic. The URL of this dynamic result is http://multimedia.boston.com/search?q=seo&s=. Yes, this URL contains parameters (q=seo and s=), but just two.
While this search capability is very cool, the returned pages will often not be indexed since search engines have difficulty typing queries into search boxes—something necessary to generate the page. So, for this page to be found, someone will need to post a crawlable link to that specific URL. EveryZing overcomes this challenge by automatically generating static landing pages for every topic with high demand, then provides links to those new pages. (We will cover this in more detail in the “Links” section of the review).
The ezSEO generated pages are impressive. They combine a collection of articles, videos, and audio files and rank them based on the publishing date or their relevance to the queried topic. These results are housed in a template that is made to order for natural search optimization. The template generates meta data for each page that combines the targeted keyword into a pre-set template. To demonstrate their approach, I’ve made the keyword bold in each string; the rest of the meta data is from the pre-set template.
- Page title: “Manny Ramirez Audio and Video brought to you by Boston.com”
- Meta description: “Watch and Listen to Manny Ramirez videos and audio clips on Boston.com”
- Meta keywords: “Manny Ramirez video, Manny Ramirez videos, Manny Ramirez video clips, Manny Ramirez audio, Manny Ramirez multimedia”
- Body content: This text is encased in a search engine friendly H1 tags. The headline for this page is “MANNY RAMIREZ AUDIO & VIDEO”
The title and meta data was used by Google to compile this search result for the query “manny ramirez video”.
Next, EveryZing’s ezSEARCH engine grabs all audio and video files that include “Manny Ramirez” in the file’s meta data or in the spoken content of the piece.
But what if the file is of Manny speaking? Well, unless Manny starts speaking in the third person which, sadly he sometimes does, the ezSEARCH engine must rely on the meta data added to each file by its producer. EveryZing’s product does not perform facial recognition so human editors are still needed to ensure the best results.
The constructed landing page created for Boston.com has simple elegance. Each listing is presented as a thumbnail image next to the file’s title and relevant content snippet.
This example demonstrates that while the speech-to-text translation is often not perfect, it does a fairly good job of understanding the general idea of what’s being said.
The real beauty of this design is that it relies on a search engine friendly use of CSS, which allows the search engines to crawl the exact audio snippet that contains the targeted keywords. This approach increases the number of mentions for the targeted keyword, a concept called “keyword density.” As a positive side effect, the use of off-page CSS to control layout reduces the HTML code “footprint,” improving page performance and the likelihood that search engines will index and rank the page for the desired keywords.
Search engines have more trust in content receiving internal and external links. To fulfill this need, EveryZing places links to general multimedia categories on the right side of each page. Additionally, pages are created with links to popular topics.
These lists of topic links provides convenient navigation for humans, as well as optimized anchor text for the search engines. The keyword-rich tag cloud at the top of the page highlights the most popular queries of the moment.
While internal links like these will help disseminate PageRank and guide the crawlers deeper into the subdomain, relevant external links will provide significant validation to the search engines. Because each topic maintains a persistent and unique URL, topic landing pages should gain organic external links as visitors find the collection of fresh media valuable.
The final analysis
EveryZing offers a turn-key solution for converting a publisher’s video and audio files into searchable text that is smartly integrated into classically optimized landing pages. For large publishers and aggregators of video and audio content, EveryZing is just what the SEO doctor ordered.
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.