Commentary

Set-Top-Box Data: Next Steps

by Charlene Weisler , Op-Ed Contributor, February 4, 2009

Once a hypothetical exercise, set-top-box data as a measurement tool is quickly becoming a reality. At this time there are several well-positioned companies jockeying for dominance with their varying methodologies, data footprints and sources. But in this seemingly chaotic competitive land rush, there are those who are actively seeking conciliation and structure. MPG's Set Top Box Data luncheon on Jan. 7, organized by MPG's Mitch Oscar, was a very positive step in this direction.

Attending were agency researchers, content company researchers, data processors and providers, developers and programmers, all of whom engaged in lively discussion of possible next steps for data usage and acceptance.

As an advocate for the use of set-top-box data as a measurement tool, I think that now is the time to examine the next steps needed to standardize, analyze and ultimately monetize this breakthrough data source. Here are a few suggested next steps:

Set up a non-affiliated, non-partisan advisory council either through an established industry or respected accreditation organization. This council, composed of programmers, agencies, processors and suppliers, would help build consensus and guide the formation of a new measurement currency that works for all interested parties.

Standardization of data processing rules. Can we agree on certain data attributes as the foundation for which other data points are based? For example: standard latency periods.

Standardization of the metrics and nomenclature. Do we conform to the current currency of ratings, shares, HUTS and PUTS? Or do we find new, relevant measurements that all processors can agree upon?

Address and overcome data deficiencies. Yes, there are bonafide issues with set-top-box data. For example: There is no agreed-upon national footprint, and efforts to weight the data to make it more national are not universally accepted.

Agree to dispense with "red herring issues" that only serve to confuse. Some stated problems are not really problems. For example: Lack of demographics is not a deficiency with the data, since datapoints can be matched to actual spending and lifestyle information via companies such as Acxiom. Isn't that what advertisers and programmers really want? Not all women 18-49 are alike or equally valuable to advertisers.

Aggressively market data advantages. Set-top-box data remains the only source of second-by-second measurement. That, along with DVR usage, extensive and stable out-of home information, the ability to integrate seamlessly with online usage, and the ability to parse out actual ad performance, makes the data unique and valuable to a range of customers and is currently unavailable within the current currency. These are market positions worth repeating.

Cooperation and partnerships can help form a foundation for data acceptance and provide the missing pieces of this media measurement puzzle.

Work with Accreditation services to help vet the process and speed acceptance of the data as a possible industry currency.

10 comments about "Set-Top-Box Data: Next Steps".

Check to receive email when comments are posted.

Brian Cauley from media iq, February 4, 2009 at 4:48 p.m.
Well said, Charlene. I especially agree with your points about the data advantages.
Reply

John Grono from GAP Research, February 4, 2009 at 9:36 p.m.

Charlene I applaud the post and the initiative shown under Mitch's leadership. I truly believe that STB data is essential for future audience measurement. But I am far from convinced that STB data alone will provide this - I think we will still need panels of some sort. That is, we will need a hybrid solution, just like online is looking at with regard to server-centric data and panel data.

But I do have a specific question (remember I am from Down-Under). How does Axciom match actual spending and lifestyle with STB data? Surely they don't have actual spending for all homes with an STB and for all product categories and brands! Are we talking some sort of fusion process between a sample of spending/consumption/lifestyle and STB data?

I would also like to caution analysis of data at the secon-by-second level. There is considerable 'drift' in STB clocks, and of more importance their is latency delays in broadcasts across platforms. Here in Australia on fibre-optic cable that latency is around 8 seconds (the mode). 8 seconds out in a 15-second ad is an eternity ! I would recommend some sort of content matching rather than time-based matching if this is an ongoing objective.

John Grono, GAP Research, Sydney Australia

Charlene Weisler from Writer, Media Consultant: WeislerMedia.blogspot.com, February 5, 2009 at 10:02 a.m.

Hi Brian,
Thank you. I often find that the unique advantages of the data get lost in general discussion.

Charlene Weisler from Writer, Media Consultant: WeislerMedia.blogspot.com, February 5, 2009 at 10:14 a.m.

Hi John,
Certainly panels can provide data nuance that we currently cannot get directly from box. But there are initiatives underway that will be able to match data from various other sources (including Axciom and other credit and consumer services) that may eventually supercede the need for the currently accepted panel datapoints of gender, age, education etc.

I believe that agency researcher, planners and buyers should be consulted about what datapoints and methodologies best serve their needs. Viewership data is a currency and they are the ultimate end users/clients.

I agree too that second by second is a challenge as long as we can't agree on latency standards. Content matching is the ideal but I can think of no easy or cost effective application for that at this time.

However I know of engineering companies who can take the feed off a satellite and stream it over the internet. I wonder if that might be an efficient first step to content match?

Rebecca Rachmany from AdsVantage, February 5, 2009 at 10:27 a.m.

Clearly, click data provides tremendous insight regarding second-by-second behavior and response to television content and advertising. As suggested, combining it with existing demographic and behavioral data can provide a very broad range of measures.

While we've all become accustomed to a "currency standard", looking at Web analytics gives us some different perspective on standard metrics. Some companies measure impressions (eyeballs viewing an ad), while the majority measure Click Per Action (CPA). Even within CPA, companies measure clickthroughs, time on a web site, conversion, size of purchase, etc.

While there is always a need for an industry standard, it's quite easy to envision a world where different organizations use different metrics. Some companies may be interested in whether a customer viewed an advertisement from start to finish, while for others, measuring how many people called a toll-free number within half an hour of viewing is the only relevant parameter.

I've elaborated on this, and described some of the challenges in looking at STB clicks in a full blog response at http://ads-vantage.com/Blog/February-2009/STB-Clicks--The-Good,-The-Bad-and-The-Ugly.aspx

Charlene Weisler from Writer, Media Consultant: WeislerMedia.blogspot.com, February 5, 2009 at 10:57 a.m.

Hi Rebecca,
I just read your blog post and I agree whole heartedly on your points.

My emphasis on standardization of metrics is strictly for posting purposes. Currently it may be more valuable for one advertiser to seek a length of tune (see more than 10 seconds of the ad) but that will be part of the vetting process for networks under consideration. The contract of performance will be based on audience grps for posting.

I think that our two perspectives are really short term (standard metrics) vs medium to longer term (individual metrics based on specific advertiser needs and wants).

And frankly I think many agencies are realizing that the current currency does not work for them any longer. But market uncertainty may require baby steps towards metric individualization.

Rebecca Rachmany from AdsVantage, February 8, 2009 at 6:07 a.m.

Hi Charlene,

Thanks for the response. I agree that the change in metrics will come slowly, not because of the lack of need, but because any new system takes time to be accepted. We are right now at a turning point where companies willing to take risks with ratings stand to gain. The current system puts everyone on an even playing field, in that the same standard metrics can be purchased by everyone. Using STB data may change that balance, in that the STB clicks are in the hands of the service providers, not the advertisers or agencies.

Rebecca Rachmany, AdsVantage

John Grono from GAP Research, February 10, 2009 at 5:03 p.m.

Thanks Charlene. Regarding the data matching this of course can be a two-way street. The panel data can be fused onto the STB data, and the STB data can be fused onto the panel data. Not all targets are consumption based - many are psychographic or attitudinal. I see it as essential that STB data be a donor as well as a receptor.

We have been working with fused data here in Australia (and New Zealand) since the mid 90s. We haven't been fusing STB data (though that project is recently underway) but have been fusing programme based 'currency' TV ratings (and radio) on to a large consumer-based panel. The most difficult hurdles so far have not been technical, but have been getting advertiser support financially. Interestingly, component analysis has shown that the good old stalwarts of age, gender, education, occupation, lifecycle etc always come out way up the top of the list - meaning the age/gender surrogates we have used for decades are extremely meaningful. They also have another benefit - they are easily and accurately applied across all media measurement.

Cheers.

Charlene Weisler from Writer, Media Consultant: WeislerMedia.blogspot.com, February 15, 2009 at 2:24 p.m.

Hi John,
How are advertisers using the fused data? Is is a sales metric that they can post against?

Charlene Weisler from Writer, Media Consultant: WeislerMedia.blogspot.com, February 15, 2009 at 2:28 p.m.

Hi Rebecca,
I may be naive but I would think that advertisers would embrace set top box data since it is based on actual behavior of the full footprint (whatever footprint that is) and not from a small sample that sometimes requires weighting and ascription.

And what if the stb data was accredited and processed by a third party processor?

One continues to hope....