Commentary

CBS, Nielsen Hit Stalemate

1. Douglas Ferguson from College of Charleston, January 8, 2019 at 11:01 a.m.

   I wonder which dinosaur will prevail. 

   Reply
Ed Papazian from Media Dynamics Inc, January 8, 2019 at 12:39 p.m.
It's amazing that this mainly pricing-based squabble is being played up so much as to suggest the possible end of Nielsen. What rubbish!. As for the decision not to use Nielsen in-home ratings for CNBC's morning business news reports---which have a  handful of out-of-home viewers---that has nothing to do with the current situation. Same goes for stations using Arbitron ratings, locally, instead of Nielsen. Why not? At the time---35-40 years ago--- Arbitron, not Nielsen, was the leader in local market rating measurement and many agencies, including BBDO, subscribed to and used Arbitron local market TV ratings. As for who subscribes to what, it's a common practice for TV rating wannabies to give agencies what amounts to free---or virtually free---subscriptions to their service and then use this to claim that they have broad based acceptance--when, in fact, they don't. Finally,  in the case of "C-flight", the" other data" that is used is set-top-box ratings which produce a profiling index for each show that is used in conjunction with Nielsen viewer ratings to make "advanced" targeting buys for a relatively small number of brands---but the primary quantitative benchmark in virtually all of these  deals remains the Nielsen ratings. It's not that improvements don't need to be made, nor that Nielsen is infallable, but gee wiz, why all the drama?
Reply
Frank Anthony from Self, January 8, 2019 at 12:58 p.m.
Radha, who now heads up research, is former Nielsen so figures she can hold their feet to the fire on pricing and make a good impression with the suits at CBS for at least trying.  I don't blame her for trying.  If she can hold out long enough to the point NLSN stock takes a hit, the headliners at Nielsen may cave a bit.  
Reply
Evan Brown from Atlatl Media, January 8, 2019 at 6:14 p.m.
Why does everyone seem to think you cannot exist without Nielsen?  Why do we continue to accept the flawed notion that crappy ratings are better than no ratings at all?  Why do advertisers contine to fit their "square" market definitions into the "round holes" that are Nielsen DMAs. 

We can absolutely do better, and in this day and age where engagement, conversion, and other z-axis metrics have come front and center, the notion of "viewership", impressions, and age-gender demographics are becoming less and less relevant, we have to. 

True, the media auditing firms would be run out of business because they are all connected at the hip to Nielsen, but these organizations stifle creativity and get in the way of sound strategies because of outdated ratings guidelines.  I know of one firm that insists that in a market of 450 samples, they can project ratings to the nearest 100th of a rating point. That's 1/100th of a percent...1/10,000th of a target audience. 

Plus, retail businesses have been held captive to the Nielsen DMA description which may or may not accurately reflect their footprints. Nielsen's tentacles go deep into chain retail marketing strategies, and they really shouldn't

Plus, there IS another service out there, you know.  comScore has become a viable solution especially as it expands into new cable households.  comScore brings greater flexibility and near real-time numbers even in the smallest markets.

Maybe the is the kick in the ass marketers need to finally release themselves from defining their business to an arbitrary market description based solely on TV viewership and more on definitions unique to each individual advertiser. 

Let's face it.  Nielsen is soooooo last millennium. 
Reply
John Grono from GAP Research, January 8, 2019 at 6:52 p.m.
And Evan, you are sooooo misguided.   In my youthful zeal I used to think like that as well.

One example, 'real-time' numbers were availble in South American markets in the 1990.   It was a disaster.   Companies like comscore are good at measuring traffic, and through the great work of Josh Chasin are making a good fist of trying to report on audience.   I hope you understand the intricacies of the difference between traffic and audience.

I do get concerned when I hear that businesses convey ratings precision to the n-th decimal point so I agree with you there.   It is as bas as the vanity of reporting digital traffic to the person or to a specific individual.

In fact, the challenge is NOT accuracy.   All marketers really need viable, timely estimates of who was likely to see their ad.   Was it 1.857m ... or was it 1.889m.   The difference simply doesn't matter to a marketer.

The first challenge is measuring all media with similar degrees of precision and to comparable definitions (geography, demography and time).   For example, TV measures by 'average minute audience'.   Digital video is measured (in the main) on stream starts (and generally) at a global level, and generally over some long but unspecified time-frame.   That is OK - as long as you don't use it for comparative or planning purposes.   All the savvy media planners and strategists I know (i.e. hundreds) are very wary of "the big number".

The second challenge in then integrating each medium into a single holistic picture of users media consumption (and profiles etc.) on a like-for-like basis, but most importantly de-duplicating for cross-media consumption.   Sadly there is little point in tackling the second challenge until the first is addressed.
Reply
Evan Brown from Atlatl Media, January 8, 2019 at 11:12 p.m.
Mr. Grono,

We most likely approach these issues from two entirely different perspectives. That is why you may think I am so "misguided." I am not. I just come from a position of planning and buying media for over 35 years.

You are correct in that there are acceptable margins of error however you do not address inconsistency at all, and that is the real primary challenge.  I do not ask for perfection. However, I do demand "uniform imperfection." Nielsen’s inaccuracy is not the problem. It is wild inconsistency is.  

Wheel of Fortune and Jeopardy have been in syndication for decades, and in nearly all markets, they have run at the same time of day and week for decades. They have loyal audiences that have shared common demographics for over 30 years. Few people start watching them. Few people stop. The ratings should show steady consistency.

However, if you pull a 20-year ratings history you might see that Wheel may get a 2.0 in Nov. 2009, but then jump to a 5.6 in the Nov. 2010 book, then fall back to a 1.8 in Nov. 2011, especially in smaller markets. This should not happen.  If I can’t trust these ratings, how can I trust the rest of them?

Next: Berkshire County, MA is in the Albany-Schenectady-Troy DMA because most broadcast viewing comes out of that market. However, people in Berkshire County have no connection beyond TV viewing to Albany New York.  They are more closely aligned with Springfield, MA, which is half as distant from most areas of Berkshire than Troy NY, the city closest to it. 

People in Berkshire County were thrust into a DMA over half a century ago solely because the mountain range between Berkshire and Springfield blocked the broadcast signals of most Springfield TV stations when the DMAs were conceived. 

Local spot TV is rated by the average quarter-hour, NOT by the minute. In the digital world we discuss viewability, human impressions, transparency, etc. Meanwhile, Nielsen still has not figured out how to account for the segment of viewers who get up to pee during a commercial break.  Of course, neither does comScore, but a single sample of a Nielsen survey with a full bladder is a lot more impactful than a single sample of comScore’s survey.

Even more frustrating is that unlike comScore who has a single methodology, ensuring consistency in measurement techniques from New York to Glendive, MT, Nielsen has four different methodologies, depending on market size.

Nielsen is far too stuck in its legacy and existing hierarchal structure to make the "frame-up" changes it needs to make.  CBS is right in holding off, and I seriously hope they do not renew. Only then may Nielsen realize that no band-aid is big enough anymore.
Reply
John Grono from GAP Research, January 9, 2019 at 5:41 a.m.
Evan you are correct about small local markets.

In my 42 years experience of research and media, the biggest limiting factor is sample size - and that is always limited by budget.   [Full disclosure: I worked for Nielsen in retail and media for 20 years up to 1997].

Research is like a three-legged stool.   The three legs are accuracy, speed and low cost.   You can however only choose two of them.   The US DMA situation is a text-book example.   Personally I think that 210 DMAs is WAYYYYY too many.   But that is a market imperative that doesn't have the funds to back up its needs.

Here in Australia we have 26 markets to cover our 25m population.   The five large metro markets release daily data on panels ranging from n=650 to n=1,475 metered homes (total n= 5,250).

The 'aggregate' / state markets also release daily data but on less demos.   The small sub-markets currency is 4-weekly average data.   The state panels range n=180 (pop. 559k) to n=812 (pop. 1.866m).   The sub-markets range from n=102 households (pop. 193k) to n=339 (pop. 1.027m).   The all-up regional total is n=3,198.   The 'national' total is well north of n=8,300 homes which is around n=20,000 people every day.

All these markets are People Meter markets and have been since the early '90s.   It was a cost imposition back in the 1990s but one that all markets understood they had to underwrite and the system works well.   There is the usual 'noise' with small programmes in fringe time-slots in small markets but the data are fit-for-purpose.

We have an issue with measuring small regional radio markets where populations get down to the low tens of thousands.   Radio is still measured using diaries (electronic measurement is good at measuring how people listen but poor on measuring how much people listen so it understats audience and overstates reach), but in those small regional towns with one public and one commercial broadcaster we can't get enough diary sample.   We have recently developed a telephone survey that is providing acceptable listening metrics ('most listened') that is helping kick along local advertising and some national ad dollars as well.   And by the way it recently won the best paper at the ASI conference in Athens.

I think the above shows that a small country like Australia can produce audience measurement systems using a properly funded and collaborative approach.

Cheers and best wishes.
Reply
Ed Papazian from Media Dynamics Inc, January 9, 2019 at 6:38 a.m.
Regarding local market TV ratings, John is quite right in pointing out the element of cost---who pays for the much larger panels and superior research standards that are needed for a small market---or any market? It has always been true that national TV ratings were considered to be far more important than their local counterparts not only regarding sample--or panel size--but the quality of the methodology. This remains so today. Still, it is perfectly valid to note the inconsistencies in survey methods and other ills that plague local stations trying to make sense of the ratings that they are supplied with. Perhaps John's other point ---too many markets---must also be considered. If small market stations as well as many in larger markets can't afford the price of "better" rating surveys---as is obviously the case---and no one else cares enough to step in and foot the bill, perhaps Nielsen should stop measuring markets below the top 100 and work a deal with comScore that let's the latter supply set usage ratings with a larger, more stable panel base in markets 100+. The national TV agencies would adjust to this by developing viewer-per-set ratios for the smaller markets  while local advertisers in these markets, who aren't making comparisons with national TV buys anyway, would have more stable numbers to deal with---even if they aren't viewer-based.
Reply
Evan Brown from Atlatl Media, January 9, 2019 at 9:49 a.m.
John,

The 2017 Australia population estimate is 24.6 million persons (World Bank estimate). Divide that by 26 markets and you have an average of roughly 950,000 persons per market.  US 2017 population is 325.7 million (US Census Bureau).  Divide that by 210 markets and yuou have an average of 1,550,000 persons per market.   The averages however suggest that if anything Australia has too many markets...not the US.

The problem isn't the number, it's how they were drawn up.  

The US market definitions came about by Nielsen largely by whether or not a TV station was located in a particular city or not. Zanesville, OH has one single TV station (WZTR). The Zanesville, OH DMA is almost completely surrounded by the Columbus, OH DMA.

In other cases, especially out west, hyphenated markets were created to bring multple network affiliates into one DMA.  So we have Lincoln-Hastings-Kearney, NE.  Three small cities that are hundreds of miles apart cast into a single DMA.  Yet Dayton and Cincinnati are two separate DMAs in which the downtowns are less than 50 miles apart. Meanwhile, Akron OH (Summit County) has a larger than Rochester New York (Monroe County), but for some reason doesn't deserve to be in its own DMA and is lumped together with Cleveland, even though Akron has more TV stations than Zanesville.

The US was Nielsen's guinnea pig, but rather than improve it after the flaws were apparent, they left it alone.  Sixty years later, the market definitions are still screwed up. 

Ed,

You are correct in asking about the cost. panels cost money. So the question becomes how much of this can be automated to save money. What do we gain? What do we give up?  In Rochester, NY, roughly 66% of all DMA households are monotired by comScore. The measurement through set top boxes is completely automated. The cost for my agency to buy comScore is about 60% that of the cost of Nielsen.

With a sample of over 200,000 households vs. Nielsen's roughly 500 sample, does it really matter that we don't know who in a particular household is watching a particular program? With a sample size that is 400x larger, I'm betting that the comScore data that is available is far more stable than the data from Nielsen, even if Nielsen knows the samples by name and comScore doesn't. Do you want deeper levels of extremely unstable data, or shallower levels of highly stable data?
Reply
Ed Papazian from Media Dynamics Inc, January 9, 2019 at 1:09 p.m.
Evan, it does matter that we know who is watching not just that the set is on. On average a typical household resident is present and "viewing" only half of the time whan a set is tuned in.Using set usage as a surrogate for viewing creates a huge error margin for advertisers targeting a particular kind of individual in the home. My point, however, is that it is probably cost prohibitive in smaller markets to utilize samples of peoplemeter homes on a scale  that would provide ratingdata on a fairly reliable basis. So one has to create solutions. For example, if a small market was rated only by comScore using STB metrics which only tell us that one or more sets in a home were tuned in to a given telecast, and we were satisfied that the comScore panel represented a fair approximation of all homes in the market, it would be mandatory for a national agency buying spot time for a national client, to adjust the comScore ratings using viewer-per-set ratios for similar shows obtained via Nielsen. If not, then the use of household ratings in this market, compared to viewer ratings elsewhere would significantly infklate the GRP attainment in the smaller market as household ratings are usually twice those noted for people. This, in turn, would cause the agency to reduce its buy in the smaller market, in the mistaken belief that it was overdelivering on audience---when it wasn't.
Reply
Ed Papazian from Media Dynamics Inc, January 9, 2019 at 1:18 p.m.
Evan, regarding how TV markets were defined, the process was not as haphazzard as you suggest. For example,  the reason that Akron does not "win" its own market and is credited to Cleveland is that the latter's stations, in aggregate, dominate total viewing time in Akron. The reason why Cincinnati isn't credited with Dayton is the reverse. In Dayton, the Dayton stations garner most---though not all--of the viewing. Also, many markets with only one or two stations, individually, were, indeed cobbled together, in part because of the desire to have at least one outlet for each network available locally. In retrospect, this proved to be a good thing for the stations involved for had each been defined as a separate market, we would now have something like 300 markets, not 200, and many more of these would be of the smaller variety, meaning that they would have little chance for national spot business.
Reply
Evan Brown from Atlatl Media, January 9, 2019 at 2:29 p.m.
Ed,  The DMAs may have worked 30 years ago, but they are less effective and less relevant now, especially with the advent of digital advertising where geographic customization is standard operating procedure, as well as the fact that 25 years ago 80% of average national advertising budgets were devoted to television whereas now only about 45% are. 

Also, at a time when we were only counting eyeballs Nielsen's other procedures may have been appropriate as well. However now, the planning approach is to count entire consumers.  To do that requires us to think outside the bounds of current metrics and demographic descriptions.

Regarding knowing who the sample is, I was being sort of tongue-in-cheeek when I asked if it mattered if we knew the viewer's name. I have made comparisons of comScore to Nielsen ratings of various programs.  Nielsen's target may be A18-49 and comScore's may be HH's with A18-49,  but the ratings provided by each tend to trend the same.  The difference is that there are not huge swings in comScore's like ther are in Nielsen's.  So while comScore may not know the name of the individual viewer, because of the volume of samples, their ability to project the actual demographic within a household is just about as accurate as Nielsen's ability to project a demographic.  

Additionally, with that large sample size it is far easier to append things like Polk automotive data, Mastercard transaction data, etc. 

I buy political advertising and comScore can tell me what programming is most likely to be viewed by likely Republican, Democrat, and Independent voters.  In a recent campaign, we were outspent 3-1, but my ability to target using this knowledge (coupled by a tight negotiating strategy) enabled me to focus more attention to swing voters, and move the election our way.

So while you may poo poo STB data as not being as scientific as survey data, the reality is that since the volume of STB samples exponentialy dwarfs those of surveys, the accuracy of prediciting consumer behavior beyond the HH level with STB data, especially when appended to other volume-based data sets such as auto registration, credit card, property tax, etc., is very good. It is so good that I can differentiate behaviors between one target set in Akron and a similar set in Cleveland. Since I can do that, I don't need DMAs any more. If I don't need DMAs, then why buy a ratings service that only reports in DMAs?

The way I see it from the street perspective, while the marketing world has evolved dramatically over the past 20 years, Nielsen has gallantly and doggedly fought to stay the same.  They have succeeded...and they are now obsolete. 
Reply
John Grono from GAP Research replied, January 9, 2019 at 2:53 p.m.
Thanks Evan.

I neglected to say that I didn't like that we have 'sub-markets' but they are there for 'geographic necessity' and historical reasons.

Take the state of Queensland.   It is around 2 million square kilometres, around 1,500 km wide and 2,000 from top to bottom, and has a 7,000 km coastline.  It has a population of around 5 million people and a population    The geographic population density is around 2.8 people (one average family) per square kilometre).   The largest property is Cubbie Station at 96,000 square kilometres (around 5% of the entire state).

Queensland has six sub-markets.   Those sub-markets primarily exist to allow local station sales teams to sell directly to local businesses - things like Tom's Tyres of Talebudgera.

Agencies and brand marketers in the main do a 'state buy' and use the n=812 daily People Meter panel because the differences between sub-market ratings at a Total People level are generally not large, and by buying at a state level you can get deeper demographic data.    If the buy has to be geographic, say, the three southern sub-markets, we simply instruct the network to only air the ad in those sub-markets.

In essence, national buys and media agencies use the five metropolitan markets (around two-thirds of the population) and the five state regional markets.

And as Ed explains with the Rochester example, comscore is based on a sub-set of the geographic market (homes with STBs - which tend to have very different viewing patterns) but has large sample sizes of RPD with unknown/synthesised demographics.   The trade-off is data stability vs. partial coverage/bias.
Ed Papazian from Media Dynamics Inc, January 9, 2019 at 3:20 p.m.
Evan, the typical network TV brand still allocates anywhere from 65-75% of its branding budget to TV, not 45%. In contrast, and contrary to the stats frequently seen in the trade press which are vastly inflated for digital media by the inclusion of direct advertising  and other types of spending, the digital media percentage  is closer to 10-15%, again for the branding portion of the campaign, not the sales promotion/DR effort.

As for the issue of people ratings vs set usage ratings, the problem is very simple. The typical advertiser---not all, but a high proportion----values younger ( 18-49 ) and upscale consumers much higher as marketing prospects than older folks and people with lesser educations and/or incomes. According to set usage ratings, younger and affluent households are, by far, TV's heaviest users---ergo, one might mistakenly assume that younger and upscale adults watch more TV than anyone else. But they don't. In fact and by an overwhelming amount, older adults and lowbrows are TV's biggest fans. Why the discrepancy? Because younger/affluent homes  have more residents---about 3.5-4.0 per home and more receivers, while older homes have only 1.5 residents and fewer receivers. Therefore the sets in younger /affluent homes may be tuned in much more frequently, but the individual adults living in such homes are actually lighter than average viewers. Often, it's someone else who is "watching".

It may well be that comScore's STB ratings "trend" in the same direction as 18-49 ratings---in terms of who is winning and losing, for example. That's to be expected. However, using STB data in the way it is likely to be utilized in buying time-- as GRP "currency"-  in terms of program selection as it relates to targeting and cost efficience is highly suspect. We learned this lesson way back in the early 1960s, when mnay TV sets featured several people watching together. Now, approximately 70% of all set usage involves only one viewer so the distinction between tuning and viewing may be even larger and more worrysome.
Reply
Evan Brown from Atlatl Media, January 10, 2019 at 11:40 a.m.
Ed,

When did this conversation switch to netwrt TV.  We've been discussing it from a local perspective.  IN that case the top two TV advertsing categories, automotive and QSR have severely cut back dollars.  I TV station manager told me last fall that their 2018 McDonalds buy was 10% of what it was two years ago and they hardly get any more tier 2 automotive.  Local TV station revenues across the country are hurting badly in some markets.  Why do you think there is so much consolidation?

As far as "currency" is concerned I recall the disclaimers in the Nielsen bbooks warning that their ratings were to be used directionally for spot programming and only considered gospel for the survey in total.  Those disclaimers for some reason are no longer there, but as I recall, they warned against using Nielsen as a tool to post.  And rightfully so.  At the 95% confidence level, if you calculate the margin of errors, which

Nielsen ALSO took the margin of error tables out of the local books and put on a web site, even the Super Bowl rating in a local market had a +/-10-12% margin of error against A18-49.  When a spot garnered a 1 rating, it could be as low as 0.1 and as high as 2.  That's almost as bad as saying this dime could be worth a dime, but it also could be worth just a penny or as much as a quarter.  Some currency.
Reply
Evan Brown from Atlatl Media, January 10, 2019 at 11:58 a.m.
Ultimately, if CBS does pull out the impact on networks could be that Nielsen will have to chaging their business models.  If CBS can rely on national agencies footing the bill for ratings, then other networks may jump on the wagon, and that would be a significant revenue hit for Nielsen.

Nielsen make a lot more money from networks and local affiliates than from agencies. In my local market, it costs about $4,000 to pick up the local Nielsen book for a year, while the local TV stations are charged nearly seven figures. 

Network costs for national agencies vary based on network billing, but only when an agency pushes 9 figure bi9lling does Nielsen's netowrk fees come close to seven figures.  Where are they going to make up that shortfall?  Certainly not from the agencies or the advertisers. 
Reply
Ed Papazian from Media Dynamics Inc, January 10, 2019 at 12:23 p.m.
Evan, regarding the standard error calculations there is no way to predict a survey's accuracy, as a layman understands the term, by knowing its sample --or panel---size. All that these tables tell you is the likliehood of getting the same answer---not the riight answer---if the same study, with all of its assumptions and flaws were repeated. As for spot TV ad revenues, of course certain categiries are down as you noted, but overall, ad spending, while hardly robust in nature, is more or less flat---if political dollars are left out of the equation. I do agree that spot TV is in worse shape than national TV regarding ad support and the failure of many stations to develop a meaningful digital presence as well as sales teams that understand how to market such opportinities for more selective targeting to advertisers. Certainly, this is one reason why station ad revenues are in a sluggish holding patern. Another reason is the rise of spot cable, which now garners about 20-30% of local TV ad dollars. 
Reply
Ed Papazian from Media Dynamics Inc, January 10, 2019 at 1:06 p.m.
In the good old days many advertisers not only sponsored their own TV shows but they were intimately involved in their production and content---so they subscribed to Nielsen to keep an independent track on how "their shows" were doing. All of that fell by the wayside long ago and now there are virtually no advertisers who subscribe to Nielsen---a few in-house operations, excepted. If Nielsen were to change its business plan, due to defections by major TV entities such as CBS. NBCU, Disney, etc. it would  try to charge the agencies far more than they now pay---and the bulk of such burdonsome fees would be passed on to their clients---who would, in most cases, refuse to pay. As the agencies would resst getting stuck with the bill the result would be a major reduction in Nielsen's service---probably abandoning all but the largest markets as well as major cutbacks in its national operations---all of which would be to the detriment of the medium. Under this scenario, we would be confronted with a situation where each seller picks whatever data source serves it the best---a wild west situation where comparability as well as objectivity is almost totally lacking.

It may well be that an alternate "big data" source might step in to fill the void if Nielsen bowed out, but it, too, would face issues. Like getting a proper fix on viewing not just set usage and, more important from a business viewpoint, how to operate profitably. 

At this point, I don't see any of this happening and CBS will probably work things out with Nielsen, but the alternatives, if an impass were to develop are certainly intriguing.
Reply
Evan Brown from Atlatl Media, January 10, 2019 at 1:49 p.m.
Ed, don't flatter Nielsen that much.

If Nielsen raised agency fees, agencies would just simply defect to comScore. Nielsen is already about 60-70% more expensive.  There is no way agencies could absorb even 5x inflation, and the rates agencies pay would most likley need to increase 20x or more for Nielsen to recoup lost revenues from TV stations and networks.  The $100 million Nielsen contract with CBS is concievably about 5 to ten times greater than the combined contracts with Initiative, UM, OMD, and DAN, combined. Zenith dumped Nielsen three years ago. 

You are probably right that CBS will fall back in line.  SBG did.  However, the scars, and they are very deep, will remain for a long time. SBG now has both comScore and Nielsen, and comScore has become their default currency, like it or not.

The two largest local TV groups, SBG and Nexstar, representing about 22% of local TV stations in the US, have shifted priority to comScore. Spectrum Cable (formerly Time Warner) has agreed to allow comScore to measure their STBs, increasing STB monitoring to about 60 million HH's or so. Other significant broadcast groups are cosidering the shifts.  In NY, outside of NYC, only about 40% of the broadcast TV stations still subscribe to Nielsen, while about 80% subscribe to comScore. 

Friedman also neglected to point out that CBS already uses comScore, and that FOX is also using comScore to leverage its contract with Nielsen. This Forbes article is a little less biased toward Nielsen and clearly points out that this is not a 100% Nielsen issue.  Forbes indicates that comScore is the biggest threat to Nielsen's hold on ratings in decades. I do remember when Arbitron tried to go against Nielsen.  Seems to me they had a lot shallower pockets and their offering was weak.  https://www.forbes.com/sites/bradadgate/2019/01/07/can-cbs-go-without-nielsen/#7b692f984b57

Your "fall-on-a-sword" defense of Nielsen is admirable, but from my perspective, comScore is winning most of the battles, even in the cases where the Nielsen contracts are renewed. 

Good debate, sir.  We'll see what happens. 
Reply

CommentRead more comments >

Next story loading