ellipsis flag icon-blogicon-check icon-comments icon-email icon-error icon-facebook icon-follow-comment icon-googleicon-hamburger icon-imedia-blog icon-imediaicon-instagramicon-left-arrow icon-linked-in icon-linked icon-linkedin icon-multi-page-view icon-person icon-print icon-right-arrow icon-save icon-searchicon-share-arrow icon-single-page-view icon-tag icon-twitter icon-unfollow icon-upload icon-valid icon-video-play icon-views icon-website icon-youtubelogo-imedia-white logo-imedia logo-mediaWhite review-star thumbs_down thumbs_up

Video search 101: Marketing and optimization

Abe Snyder
Video search 101: Marketing and optimization Abe Snyder

In November '08 alone, more than 12.5 billion videos were viewed, up from 11.4 billion in July '08. In November, 146 million people accounted for these views, up from 142.5 million in July '08. At these rates, 50 percent of the U.S. population will be watching online videos by the end of April '09.

David L. Smith is CEO of Mediasmith in San Francisco.

So how can marketers reach this growing audience with relevancy and appropriate messages?

First of all, it's important to understand that this is a new marketplace that needs new technologies. The legacy search technologies are based on text and HTML, and cannot readily discern what is in a video.

Two video practices that can be used to facilitate this targeting are:

  • Video search engine optimization (VSEO): The optimization of any video asset to lift organic search listing results in order to increase views and create a "forced virality" of your content

  • Video search engine marketing (VSEM): Targeting video content, on any site with indexed video, by the selection of paid keywords pertaining to what you are marketing, with video ads and/or accompanying banners

Next page >>

How does VSEO work, and why is it important?
Every marketer should use VSEO to ensure content can be found and watched.

VSEO is directly comparable to standard site SEO. Both are done to improve the probability, page share, and relevancy of your content. Page share is often overlooked with standard search, but video search should be a major part of any VSEO effort. Video can live on multiple indexed pages, where each one can appear as a separate result, increasing your results page share.

How does one practice VSEO? Once again, it can be compared to standard SEO, with a few major differences.

Here are the key elements VSEO, which should be practiced with any and all videos you distribute on the web to gain visibility and relevancy:

  • Meta-tags: Like a web page, these are keywords associated with your video, built into the coding of both the video and on the page where it lives.

  • Context: The contents of any page surrounding a video will be indexed to help find relevancy to some broader keywords.

  • Unique URL: Every video should have its own URL to ensure all contextual and metadata indexing is as relevant as possible.

  • Format: Knowing your content and potential viewers can determine the most appropriate format. For some viewers, HD is very important. For others, fast downloads is the priority.

  • Seeding: The more places a video is available for viewing, the higher your page share with any video search engine results can be.

Deciding what you want out of your video will help focus your VSEO efforts appropriately. The two options are promotional video, where the video is the message and is geared toward distribution and seeding efforts, and stationary video, where the video is used to draw traffic to a specific site, creating inventory that can be monetized.

  • Promotional: Meta tags and format are crucial for indexing your video, as you must rely on only what is packaged with the video and not necessarily the surrounding content as it is shared.

  • Stationary: Format, context and unique pages are all important, as these will allow publishers to most effectively monetize content for VSEM purposes.

What opportunities do VSEO efforts create?
The monetization opportunities for VSEO lay with both advertisers and publishers. For advertisers, it's the promotional videos. For publishers, it's mostly with the stationary videos, though the distribution of content can be monetized for publishers as well.

Advertiser monetization of VSEO practices falls into a totally new category. Though it could rely on some paid media to help seed and push the content, the core of this is video distribution. Systems to monetize VSEO are based on number of videos distributed, number of sites distributed to, and how deep each individual video should be tracked. The overall effort still starts with VSEO because once a video is distributed, it won't be picked up without giving it the best chance through VSEO.

Publisher monetization, around stationary content video, is best thought of as creating relevant, targeted video inventory. The targeted inventory could then be sold in the form of pre-, mid- or post-roll, or through the banners accompanying a video. The value of this inventory is increased when it can be targeted by keyword rather than by channel or bucket. Publishers can take steps to either make the video sharable, and potentially further increase value to advertisers if a pre-roll or other unit is embedded with a video, or to prevent the video from being shared in order to retain the ad inventory for content on your page.

Video distribution through companies like TubeMogul have made the process of uploading to multiple video share/hosting sites a one step process (after the initial accounts on each site are set up). They also feature a large number of tracking options that will let you follow your videos wherever they end up.


How can advertisers best use the indexed videos for marketing campaigns?
The answer? Video search engine marketing.

How does VSEM work?
Targeting pre-qualified users through keywords is the basis for all search marketing. This has been accomplished through various techniques, mainly with algorithms designed to rank the relevancy and quality of content, though other techniques do exist. The techniques for searching video content have developed out of the traditional search models, followed some other trends in search, and are now evolving with the development of new search technologies.

The first generation of video search engines (universal search) relied on the text content surrounding videos and other traditional search techniques, like scanning metadata. Though this method can still be found in some engines, it is no longer in use with the major video search providers. Search engines started to move away from this method because of the inherent risks of relying on metadata provided by the video's publisher, a lack of relevant contextual surroundings, and a lack in precision. At this point, all search engines -- video and standard -- realize that the best results come from the most accurate results, so how can they get better video search results? 

Enter human-powered video search, another technique directly bowered from traditional search methods. In this system, videos are watched and indexed by a staff of people. The indexing is based on the actual content of the video (human image recognition), and then transferred to metadata. From there it operates like traditional search methods. This technique is still one of the most accurate but has a limit of scale compared to some of the automated engines. So how can we have the accuracy of human-powered search with the scale of automated indexing?

Video search engines like Blinkx, Google Video, Singing Fish (for AOL video properties), CastTV, and VideoSurf are now all using methods that take the scalability of automated search and blend them with the accuracy of human-powered search. The models used by current providers are all very similar, with slight variations in the process. The development of tools to crawl page content, the video (automated image recognition) and/or audio has been the first step to take VSEM to a critical mass. As these systems develop and we see the accuracy continue to increase, we will see the demand for VSEM capabilities increase -- not only with the video search engines but all sites, from sharing/hosting sites to news sites through white label video search providers like Pixsy.

The technologies supporting VSEM are evolving slowly, and we're sure to see applicable developments in the near future with a reliable AI system still a bit farther down the road. For example, current audio technology can convert voice to captions which can be indexed, yet anyone who has used one of the many voicemail-to-text offerings can attest to some of the interesting interpretations these provide. On the video side, there is little clarity other than that current AI can identify if there is a lot of moving flesh color, which is best used for trying to avoid inappropriate content not contextually identifying videos. These current technologies rely on the original algorithms that came from SEO and SEM to find relevant metadata and page content.

Is there any VSEM inventory to buy? How and what exactly do we buy?
Currently, the VSEM market is in its infancy with most vendors and publishers still getting their general video serving systems updated to handle the growing demand. Video search engines and some video sharing/hosting sites are currently the major sources to find VSEM inventory. Outside of YouTube, there are few sites, even among the sharing/hosting sites, currently offering true VSEM to its advertisers. The two means of purchasing VSEM on video sites are through targeted keywords and vertical keyword buckets. The main difference is that the vertical keyword buckets are usually predetermined and are more of a best-fit model, where targeted keywords are a specific list created by the advertiser.

In both current VSEM models, there are challenges one faces when putting together a campaign. These include a lack of readily available inventory (depending on your product and keywords), large minimum buys where inventory is available and a lack of understanding from the media community.

The lack of inventory tends to affect the B2B audience far more than consumer, because of the type of content desired to be associated with. Some of the major business publishers like Forbes and Fortune are taking large steps to create video libraries around a wide variety of topics. Along with smaller business publishers, such as Small Business TV, an online library of useful business information on every topic a business owner or decision maker may need is being created. It is these efforts, along with the eventual white labeling of video search platforms for destination and news sites that will help open up VSEM to every marketing segment.

When faced with new media most advertisers will test before investing. The minimum buys seen with some VSEM offerings make this difficult. Without proving the effectiveness, it is hard for most advertisers to spend so freely, especially considering the current economic turn.

Finally, the media planners of the world are still learning what this all means and how it would fit into their media mix. As case studies develop and the use of VSEM as a practice develops, this will become less of a barrier.

Now, quickly, if you're a vendor or publisher, you may ask, "Why should we invest in VSEO/VSEM?"

The answer is simple.

Sites sell remnant inventory through video ad networks at greatly discounted rates. Investing in video search engine technology can give your videos new life. For instance, a tech advertiser may have a product that answers a specific concern in the business space. If they are able to target all the video content specifically covering the issue -- no matter what channel of a site it originally lived in -- it becomes valuable to that advertiser. This will likely necessitate innovative pricing structures, like CPC for accompanying banners because the inventory will literally be created.

What will we learn from using these practices?
VSEO and VSEM will both provide a vast amount of information to optimize and learn from. The media purchased through VSEM offers traditional types of metrics like clicks, click-through rates, cost-per-click, and any post-click tracking desired. VSEO enables us to see many new and interesting data points. The two main areas where statistics will come from are through distribution of the video(s) and the actual viewing of the videos.

Videos that have proper tagging and are distributed through servers designed for videos will allow full tracking of all videos. These numbers include number of views and completion rates (25 percent, 50 percent, 75 percent...) and how many times the video was shared and to where (e.g., Facebook, through e-mail). This data will show the path of the video as it distributes across the web and how the viewing varies between points. Does the sharing end at one site but have the most completed views there? Or vice versa? What does this say about the content in the video; how does it change your approach to future videos?

What's next?
There are continual developments in the video space. The next iteration of video search engines or the next VSEO techniques will have nowhere to go but up. The technologies are still in their infancy. White-labeled audio recognition technology and questionable video recognition mean that companies are still heavily relying on meta data and page context for keyword targeting. Will this be perfected in the next year? Unlikely, but as techniques develop, the marketplace will be there growing with it.

As the online space continues to develop and connections become faster and faster, content will continue to shift into video, creating an expansive and ever growing platform to be a part of. As publishers and marketers grow, these vast libraries of video content and the content spreads to more locations. VSEO and VSEM practices will monetize this medium as much, if not more, than traditional search.

Abe Snyder is senior media planner at Mediasmith, where David L. Smith is CEO.


to leave comments.