Back then, it was because Nielsen had transitioned to using "peoplemeters" from old-fashioned TV set meters and paper diaries, and the methodological shift wreaked havoc on agencies trying to calculate audience estimates to negotiate media buys for the new TV season.

In its aftermath, then CBS research chief Dave Poltrack estimated the Big 3 networks (yes, there were only three back then) lost an estimated $40 million in network upfront ad sales, which would be about $125 if adjusted for inflation today.

This year, the disruption is due to Nielsen's shift to a hybrid Big Data + panel measurement system that was actually introduced a year ago, but which according to a whistleblower statement released by the Media Rating Council last week, has been producing "seemingly unusual" audience data for at least a year.

This column is about why the MRC chose to go public with the problems now, what Nielsen is doing to fix them, and how people on both the buy and the sell side of the marketplace are preparing for what the next big Nielsen upfront tizzy.

The reason for the MRC's disclosure timing was to help Nielsen customers prepare for uncertainty and to put some heat on Nielsen to fix the problems ASAP, because Nielsen's Big Data + panel measurement service remains accredited by the MRC, although that accreditation is under review.

To its credit, Nielsen has already begun implementing some of the necessary fixes, including integrating independent, third-party universe estimates from a credible source, the Advertising Research Foundation's DASH data.

The other big changes -- mainly modeling and weighting its data to make sure it is representative of underrepresented audience segment, such as Hispanic, Spanish-dominant viewers -- is ongoing.

But the biggest problem has been generating so-called "impact" data that buyers and sellers can use to effectively factor realistic estimates for what audiences will be when Nielsen fixes the problems.

At presstime, Nielsen has committed to providing 10 months of impact data by the end of March and 12 months of data by mid-April, which buyers and sellers would still be tight for them even if Nielsen makes those deadlines.

"If we get all 12 months of data delivered at the 11th hour -- at the end of March or the first week of two of April -- I don't think people are going to be in a place where they will be ready to go," one Nielsen customer told me, adding that researchers will still need to look at the data, evaluate it and see how it impacts various other models used to forecast the marketplace and allocate advertising budgets, including marketing-mix models, etc.

"Can all those models wait until the beginning of May?," the Nielsen customer said, "I don't know. It will be really tight."

Traditionally, the major networks unveil their schedules in the second week of May, although this year, CBS is unveiling its new season lineup on April 15.

And while those upfront events are just the beginning of the media-buying season, the latency of Nielsen's impact data, the need to evaluate it, develop "conversion factors," generate new audience forecast estimates, and flow them through the array of models that people on both sides of the marketplace use to make their positions means the 2026-27 upfront media-buying season will be in a tizzy, to say the very least.

There are some other ancillary problems related to the timing of the MRC's whistle, the steps Nielsen still needs to take, the continuing role the MRC accreditation plays, and the longer term implications for the advertising marketplace post-recalibration of Nielsen's Big Data + panel measurement service.

To some, this is just one phase in a much longer-term marketplace correction, some of which reflects real changes in audience behavior, and some of which reflects changes in how audiences are calculated.

Anyone who tracks Nielsen data -- including its monthly "The Gauge" reports -- knows that linear TV viewing has been taking a hit as Americans increasingly shift to online streaming and connected TV (CTV)-connected devices. But some believe a significant factor has been Nielsen under-reporting the representation of key audience segments.

That's something they say has been going on for a long time, but really manifested during the COVID-19 shutdown, when Nielsen's field forces were extra strained, which contributed to an erosion in the representation of households with younger heads of household -- especially those 35 and younger -- and which likely overstated streaming-first households and understated multichannel households.

In other words, there already was a base of under-representation of audience segments favoring linear TV that needed to be corrected for as the marketplace was preparing for Nielsen's shift to its Big Data + panel system, which in part was designed to correct for just those sorts of understatements.

The problem is that Nielsen's recalibration adjusting for them might appear as an audience "spike" in mix and allocation models, when it really was just a restatement of actual viewing behavior that has been going on all along.

As for the MRC's continuing accreditation of the service while Nielsen works through the fixes and recalibration, some people in the marketplace will see that "stamp of approval" as indicating the service is stable and producing viable audience estimates.

"It's accredited so that must mean it's good," quipped one Nielsen customer.

For its part, the MRC says its review of Nielsen's service is ongoing, implying that accreditation could change at any time. This wouldn't be good for Nielsen, since it also is competing with other "alternative currencies," three of which -- Comscore, iSpot and VideoAmp -- currently are "certified" by a networks-controlled JIC (joint industry committee), and Nielsen's and Comscore's only point of differentiation is that they currently are the only TV audience measurement services actually accredited by the MRC, which is the industry's official accrediting body.

I have no idea what word the late Mike Moore would use to characterize this year's upfront if he had been preparing for it today, but I think it would be more choice than a "tizzy."

Lastly, to keep all this in perspective, I think it's important to remind everyone that -- as longtime Nielsen and NBCUniversal exec Kelly Abcarian pointed out out during last year's CIMM East conference in New York City -- "every number" we use in the advertising marketplace today is "modeled," and not derived from empirical audience measurement.

And however they are derived, audience estimates are just that -- estimates. And it is up to the people participating in a marketplace to decide whether they are "currency" that can be used as the basis of a transaction, and if so, how much they are worth.

Or to paraphrase Woody Allen, planners and buyers and sellers still need the eggs. Can Nielsen put them back together again?