Vivvix measures Mobile Web Video advertising activity appearing on approximately 1,900 Mobile Web sites in streaming video content. Our data provide strategic intelligence about the brands advertising in Mobile Web Video including their site placements, flighting & scheduling patterns, message length, and message content.
Vivvix's monitoring of Online Video encompasses ad streams up to four minutes in length served from pre-, mid-, and post-roll ad positions.
Ad Collection and Classification
Vivvix uses both a spider/bot technology and human panel to collect and process Mobile Web Video ads. The spider is compatible with all media players and streaming video ad formats. Outlined below are the collection methods for each.
Spider/Bot Collection
Each site is monitored at least one time per day and a sample of video ads is collected to represent complete activity on the site. A spider starts at a specific entry URL and navigates its way through the site. As it visits pages, the spider performs several processes in real time.
1. Identifies Video Content Originating from Ad Servers
The spider analyzes the web page markup and identifies video streams that are queued up in an ad server waiting to be delivered. It collects and processes this material in sequential order. In addition, the spider searches for other video streams on the sites, launches them and looks for ads.
2. Creates a Unique Signature
The audio/video is analyzed, and a unique signature is generated for matching against a Master Library of known signatures. The signature generation and matching process is sensitive enough to detect very minor variations in the source content.
3. Matches Against Known Signatures
As the spider creates signatures, it concurrently scans the Master Library for matching signatures of previously captured ads or non-ad content. If a match is found and confirmed as an ad, the spider will write information to an Occurrence File indicating the web site, the time of collection, and a reference code that uniquely identifies the video segment. The reference is used to retrieve stored classification details about the video segment, including the name of the advertised product and a product category assignment.
4. Detects New Video Ads
If a matching signature is not found in the Master Library, the spider has potentially found a new advertisement. The signature and the full audio/video are saved and routed to an editor for visual examination. When it is identified as a valid commercial or a non-ad, the signature is sent to the Master Library for permanent storage and future use in the matching process. For a new ad, the editor classifies it into our proprietary Vivvix Central database of advertised products. A record is entered into the Occurrence File indicating the web site, the time of collection and the associated Vivvix Central details.
Panel Collection
Panel collection is driven by live human interest and behavior of 1.2 million panelists, comprised of 850 thousand desktop and 375 thousand mobile panelists. While the panel is representative of a cross section of consumers throughout the United States, the panel is fully anonymized, and as such personally identifiable information is not collected, such as demographic and geo-location details. Ads are passively collected as panelists organically browse the internet, using a similar imaging matching algorithm deployed in the spider collection for the purpose of classification.
Classification
• Product Classification – The service uses the Vivvix Central classification and brand/product naming conventions.
• Video Site Classification – Video sites are associated to the site category that best reflects the site’s primary business focus. Examples include; Business Finance & Investing, News & Current
Events, Sports, etc. Video sites that are major portals are assigned to the Video Portals category.
Placement Detail
Vivvix collection of Online Video activity captures a variety of details pertaining to the ad placement.
Outlined below are key dimensions available in Vivvix’s reporting taxonomy.
Ad Length
For each collected unit, Vivvix reports the ad length, distinguishing between :05s, :15s, :30s, :60s, etc.
Transaction Type
When available, Vivvix collects information about the type of buy (direct vs. indirect), identifying the intermediaries participating in the deployment chain and defining the transaction type (programmatic, programmatic RTB, etc.). Each valid ad occurrence is used to identify ads served from a defined list of ad networks. When a match is found, the ad network identification is added to the record in the occurrence file. (Note: our reporting rules do not permit (1) site level detail within ad networks, or (2) spend estimates for ad network activity).
Please note that while Vivvix collects pre-, mid-, and post-role positions, it is not currently possible to identify ad position in Vivvix reports.
Ad Impression Estimation
Vivvix measures ad impressions using a combination of engagement and experience metrics for each site. Engagement metrics are provided by SimilarWeb for site visitation, average pages per visit, and average duration per visit. Impressions are available from October 2019 onward. Experience metrics measure ad exposure for each site and are derived using industry benchmarks (Google ad rules for banner and video placement), observational data obtained from spider collection, and primary research replicating the site’s ad-to-content experience.
For Mobile Web Video, on a site-by-site basis, Vivvix uses engagement and experience metrics to calculate ad impressions by taking the product of Mobile Visitation and Average Duration per Site Visit to derive Total Time Spent on the Site, which is then adjusted by a ratio reflecting the average number of ad streams served per minute of site duration to estimate total impressions delivered. Vivvix then analyzes the sample of a site’s spidered ad occurrences to determine the distribution of ad impressions for all brands participating on that site. Each spidered ad occurrence that was collected across all pages and the full period of monthly coverage is assigned a weighted ad impression for that site. This weighted value becomes the projected impression estimate for that unique ad occurrence and the brand associated with that occurrence.
Ad Spend Estimation
Monthly ad expenditures are available from October 2019 onward and are calculated by applying a cost-per-thousand (CPM) estimate against ad impression values:
Expenditure = (CPM) * (Impressions / 1000)
Site specific CPMs are aggregated from multiple resources, including published rate cards, publisher survey submissions, third-party analysis, and ad tech intermediaries and aggregators, where available. Rates are analyzed to determine the relative value of inventory served on a specific site. Vivvix’s CPMs generally reflect a more premium, direct / reserved buy for a site’s inventory.
In instances where a site-specific CPM is not available, Vivvix uses the site category average. In rare instances where a site category average is not available, Vivvix will use the average for the media type.
CPMs are not adjusted for length of the ad unit.
Preliminary Expenditure Estimation
Prior to the release of final monthly ad expenditures, preliminary estimates are updated daily and based on a weighted value informed by the most current month of final spending estimates. Weighted values are defined at the property level. In instances where the most current month of final spending is not available at the property level, the weighted average for the total media channel will be used. Brands reported in preliminary daily expenditure estimates reflect confirmed sighting of ads served on the site.